The C-type Lectin Receptor Dectin-2 is a receptor for Aspergillus fumigatus 1 galactomannan 2

: 24 Aspergillus fumigatus is a ubiquitous environmental mold that causes significant 25 mortality particularly amongst immunocompromised patients. The detection of the 26 Aspergillus- derived carbohydrate galactomannan in patient sera and bronchoalveolar 27 lavage fluid is the major biomarker used to detect A. fumigatus infection in clinical 28 medicine. Despite the clinical relevance of this carbohydrate, we lack a fundamental 29 understanding of how galactomannan is recognized by the immune system and its 30 consequences. Galactomannan is composed of a linear mannan backbone with 31 galactofuranose sidechains and is found both attached to the cell surface of Aspergillus 32 and as a soluble carbohydrate in the extracellular milieu. In this study, we utilized 33 fungal-like particles composed of highly purified Aspergillus galactomannan to identify a 34 C-type lectin host receptor for this fungal carbohydrate. We identified a novel and 35 specific interaction between Aspergillus galactomannan and the C-type lectin receptor 36 Dectin-2. We demonstrate that galactomannan bound to Dectin-2 and induced Dectin-2 37 dependent signaling including activation of spleen tyrosine kinase and potent TNF a 38 production. Deficiency of Dectin-2 increased immune cell recruitment to the lungs but 39 was dispensable for survival in a mouse model of pulmonary aspergillosis. Our results 40 identify a novel interaction between galactomannan and Dectin-2 and demonstrate that 41 Dectin-2 is a receptor for galactomannan which leads to a pro-inflammatory immune 42 response in the lung. Galactomannan is a carbohydrate that is part of the cell surface of Aspergillus but is also released during infection and is found in patient lungs their bloodstreams. The significance of our research is that we have identified a mammalian immune cell receptor that recognizes, binds, and signals in response to galactomannan. These results enhance our understanding of how this carbohydrate interacts with the immune system at the site of infection and will lead to broader understanding of how release of galactomannan by Aspergillus effects the immune response in infected patients.


Recognition of these carbohydrates by immune cells by carbohydrate lectin receptos 48
can lead to clearance of the infection or, in some cases, benefit the fungus by 49 dampening the host response. Galactomannan is a carbohydrate that is part of the cell 50 surface of Aspergillus but is also released during infection and is found in patient lungs 51 as well as their bloodstreams. The significance of our research is that we have identified 52 a mammalian immune cell receptor that recognizes, binds, and signals in response to 53 galactomannan. These results enhance our understanding of how this carbohydrate 54 interacts with the immune system at the site of infection and will lead to broader 55 understanding of how release of galactomannan by Aspergillus effects the immune 56 response in infected patients. 57 INTRODUCTION 9 carbohydrate-coated FLPs mimic cell-associated or particulate carbohydrates. These b-172 1,3 glucan FLPs can activate Dectin-1 signaling and elicit the formation of a phagocytic 173 synapse (20,34,35). Generation of galactomannan FLPs was performed using 174 methods previously described for the creation of b-1,3 glucan and mannan FLPs (34) 175 (Supplemental Methods). All generated FLPs underwent rigorous quality control using 176 flow cytometry to demonstrate the stable attachment of carbohydrates to the FLP 177 surface ( Figure S1). As an added control, multiple independent purifications of the 178 galactomannan carbohydrate were used for FLP creation for all experiments and 179 yielded similar results. We stimulated the reporter cell library with unmodified beads, b-180 1,3 glucan FLP, mannan FLP, and galactomannan FLP. Activation of CLR signaling 181 resulted in expression of a LacZ reporter and was quantified using a CPRG assay to 182 measure b-galactosidase activity. As expected, b-1,3 glucan FLPs activated Dectin-1 183 expressing reporter cells and S. cerevisiae mannan FLPs stimulated Dectin-2 184 expressing reporter cells. The galactomannan FLPs also stimulated Dectin-2 ( Figure  185 1B) and Dectin-2/Dectin-3 co-expressing cells (data not shown), suggesting that not 186 only did galactomannan bind to Dectin-2, but also triggered signaling by the CLR. None 187 of our FLPs stimulated Mincle or Dectin-3 signaling alone, however these cells were 188 activated by incubation with trehalose-6,6-dibehenate (TDB) ( Figure S2) demonstrating 189 that the receptors are functional, and that absence of observed stimulation is due to lack 190 of CLR activation by the galactomannan FLPs. These data suggested that Dectin-2, but 191 not Dectin-3 or Mincle, is a receptor for Aspergillus-derived galactomannan. 192

Soluble Dectin-2 binds to galactomannan FLPs 194
After identifying Dectin-2 as a receptor for galactomannan using the CLR reporter cell 195 screen, we next assessed whether Dectin-2 binds directly to galactomannan FLP. To 196 evaluate this, we used soluble murine Dectin-2-human IgG1 Fc fusion protein and a 197 control human IgG1 Fc protein (lacking Dectin-2). Flow cytometry using an anti-human 198 IgG1 Fc antibody and an AF488 labeled secondary antibody was used to detect 199 adherence of Dectin-2-Fc to the surface of FLP (Figure 2A). Dectin-2-Fc was 200 specifically bound to both Aspergillus galactomannan FLPs and Saccharomyces 201 mannan FLPs but did not bind to either unmodified or b-1,3 glucan FLPs. As expected, 202 human IgG1 Fc protein did not bind to any FLP. As an additional control to demonstrate 203 that binding to the galactomannan FLPs was Dectin-2 specific, we incubated the Dectin-204 2-Fc protein with anti-Dectin-2 neutralizing antibody or isotype control for 30 minutes 205 prior to incubation with galactomannan FLPs. The Dectin-2 neutralizing antibody, but 206 not the isotype control, potently blocked binding of Dectin-2-Fc to galactomannan FLP 207 demonstrating a specific interaction ( Figure 2B). 208 209

Galactomannan FLPs induce Dectin-2 dependent Syk activation 210
The CLR reporter assay and binding studies demonstrated that Dectin-2 is capable of 211 binding Aspergillus galactomannan. Next, we investigated whether this binding results 212 in downstream signaling. Activation of Dectin-2 through ligand binding triggers 213 phosphorylation of the ITAM motif of FcRg and subsequent signaling through Syk. FcRg 214 is also required for localization of Dectin-2 to the cell surface (36). Macrophages 215 express Dectin-2, however immortalized C57BL/6 murine macrophages have low levels 216 of Dectin-2 surface expression at baseline (data not shown). Therefore, we transduced 217 macrophages with lentivirus containing the murine Dectin-2 receptor (Supplemental 218 Methods). As macrophages produce abundant FCRg, we did not observe any difference 219 in Dectin-2 surface expression between macrophages transduced with only Dectin-2 or 220 those transduced with both Dectin-2 and FCRg. To determine if galactomannan is 221 sufficient to trigger Dectin-2 dependent Syk activation, we stimulated macrophages with 222 galactomannan FLPs for 1 hour as well as unmodified FLPs, b-1,3 glucan FLPs, 223 mannan FLPs, and C. albicans. We used C. albicans as a positive control in these 224 experiments since it is a strong activator of Syk phosphorylation in wild-type 225 macrophages. Lysates from stimulated cells were immunoblotted for phosphorylated 226 and total Syk. As expected, we observed phosphorylation of Syk when wildtype 227 macrophages were stimulated by Candida albicans and b-1,3 glucan FLPs (positive 228 controls for Syk stimulation) ( Figure 3A). There was increased phosphorylation of Syk in 229 response to both galactomannan and mannan in Dectin-2 expressing macrophages 230 demonstrating that Aspergillus galactomannan is sufficient to drive Dectin-2 dependent 231 Syk phosphorylation ( Figure 3A). 232 233

A. fumigatus and galactomannan FLPs 235
Having demonstrated that galactomannan is sufficient to activate Dectin-2 dependent 236 Syk signaling, we next interrogated the downstream effects of this activation. 237 Specifically, we sought to determine if galactomannan stimulation of Dectin-2 is both 238 sufficient and necessary for cytokine production. To determine if galactomannan 239 stimulation of Dectin-2 is sufficient to trigger cytokine production, Dectin-2 expressing 240 murine macrophages were stimulated overnight with unmodified FLPs, b-1,3 glucan 241 FLPs, A. fumigatus galactomannan FLPs, or S. cerevisiae mannan FLPs (positive 242 control) and TNFa production was measured by ELISA. Galactomannan and mannan 243 FLPs enhanced TNFa production in Dectin-2 expressing macrophages ( Figure 3A). To 244 demonstrate that this effect was specific to Dectin-2, we pre-incubated the 245 macrophages with a Dectin-2 neutralizing antibody or an isotype control for 30 minutes 246 prior to stimulation with FLPs. TNFa production was blocked by the Dectin-2 specific 247 antibody, but not by the isotype control, demonstrating a specific role for Dectin-2 in 248 mediating the TNFa production in response to galactomannan ( Figure 3B). 249 To examine if Dectin-2 is required for primary BMDMs to respond to 250 galactomannan FLPs, primary BMDM from wildtype or Dectin-2 deficient mice were 251 stimulated by galactomannan FLPs, unmodified FLPs (negative control), or LPS 252 (positive control). There was near complete reduction of TNFa production in response 253 to galactomannan FLPs ( Figure 3C), indicating that galactomannan triggers the 254 secretion of pro-inflammatory cytokines in BMDMs in a Dectin-2 dependent manner. 255 256

Deletion of Dectin-2 results in increased recruitment of immune cells, but does 257 not alter fungal burden 258
We sought to determine if the immune cell influx in response to infection was altered in 259 the absence of Dectin-2. Since treatment with steroids affects immune cell function, we 260 chose to perform this experiment in an immunocompetent, rather than a steroid-induced 261 immunosuppression mouse model. Immunocompetent mice were administered either A. presumably due to an exuberant inflammatory response. Infected Dectin-2 -/mice 273 recruited more CD45 + cells to the lungs than did infected wild-type mice. Additionally, 274 Dectin-2 -/infected mouse lungs had a significant increase in innate cells ( Figure 4B, 275 Supplemental Figure S3), as well as neutrophils compared to wild-type ( Figure 4C). 276 Notably, the proportion of the cellular infiltrate comprised of innate cells ( Figure 4D) and 277 neutrophils ( Figure 4E) was not significantly different between infected wild-type and 278 Dectin-2 -/mice suggesting that there is increased immune cell recruitment to the lungs 279 during infection in Dectin-2 -/mice, but no change in the overall distribution of innate cell 280 and neutrophil populations. Our results suggest that Dectin-2 is involved in the 281

coordination of inflammatory responses in this airway model of infection. 282
To determine whether the increased influx of CD45 + cells could be the result of 283 difference in fungal burden between wild-type and Dectin-2 deficient mice, we quantified 284 fungal burden in the lungs of mice infected with 4 x 10 7 conidia at 48 hours post-285 infection using qPCR to detect and quantify Aspergillus DNA(37). Aspergillus fumigatus 286 14 DNA was detected in all samples from infected mice, but not in PBS controls. There 287 was no difference in Aspergillus DNA between wild-type and Dectin-2 deficient mice 288 ( Figure 4F). CFU assays were also performed as a confirmatory test to ensure that 289 there was no difference in live Aspergillus fungal burden, since fungal DNA testing 290 cannot distinguish between live versus dead organisms. Lungs from both wild-type and 291 Dectin-2 deficient mice still contained live Aspergillus fumigatus at 48 hours, and there 292 was no difference in CFU between strains ( Figure 4G). Taken together these results 293 suggest that the enhanced immune cell infiltration in Dectin-2 deficient mice is not due 294 to a difference in fungal burden. 295 296

Dectin-2 is dispensable for survival in a murine model of infection 297
Our data suggest that Dectin-2 is a receptor for A. fumigatus galactomannan, and that  Although BMDMs from Dectin-2 deficient mice had decreased cytokine 318 production in response to galactomannan FLPs, there was no difference in overall 319 survival of immunosuppressed Dectin-2 deficient mice compared with wild-type mice 320 when challenged intranasally with A. fumigatus. There is significant redundancy in CLRs 321 signaling, therefore it is possible that the lack of Dectin-2 is compensated for through 322 Dectin-1 or other CLR-ligand interactions in the context of the whole organism. 323 However, publications have suggested a critical role for Dectin-2 in the immune 324 response to A. fumigatus (38,39). A recent case report identified a patient without 325 previously recognized immunodeficiency that developed invasive aspergillosis and was 326 found to have a mutation in Dectin-2 that led to decreased responsiveness to 327 Aspergillus (38). Previous work has also demonstrated a role for Dectin-2 in mediating 328 plasmacytoid dendritic cell (pDC) responses to A. fumigatus (39). These observations, 329 coupled with this study, suggest that while Dectin-2 may not be required for survival in a 330 murine model, it may contribute to human disease through modulating interactions with 331 discrete immune cell populations. 332 Interestingly, immunocompetent mice that lacked Dectin-2 have increased 333 recruitment of immune cells into the lung tissue upon infection with Aspergillus, which is 334 in sharp contrast to galactosaminogalactan (GAG) from A. fumigatus which led to a 335 reduction of neutrophil infiltration (40). The difference in inflammatory cell infiltrates was 336 not due to difference in the fungal burden within the lungs, suggesting that it is a direct 337 result of Dectin-2 deficiency. This result is interesting, as our data also demonstrated 338 that galactomannan-induced Dectin-2 signaling in macrophages leads to pro-339 inflammatory TNFa release. We anticipated that the absence of Dectin-2 would result in 340 decreased pulmonary inflammation, but surprisingly observed increased immune cell 341 recruitment. Our results suggest that Dectin-2 may normally temper inflammatory cell 342 recruitment to the lungs, or that Dectin-2 signaling normally leads to increased 343 inflammatory cell turnover. There are multiple potential hypotheses to explain these 344 findings which will provide fertile ground for future studies. In the absence of Dectin-2, 345 expression of Dectin-1 and/or other C-type Lectin Receptors may be upregulated and 346 thus these receptors may play a more substantial role in driving the immune response. 347 Additionally, Dectin-2 is expressed not just by macrophages, but also by many immune 348 cells and other lung resident cells, including epithelial cells. Thus, a cell type other than 349 macrophages, may be responsible for mediating the difference in immune cell influx that 350 we observed. For instance, studies by Loures,et al (39) demonstrated that 351 plasmacytoid dendritic cell production of type I IFNs in response to Aspergillus 352 fumigatus is dependent upon Dectin-2 mediated recognition of Aspergillus fumigatus. 353 While we did not observe a difference in pDC recruitment to the lungs in wild-type 354 compared to Dectin-2 deficient mice, these cells are likely impaired in their ability to 355 produce type I IFNs during infection. Interestingly, the increased immune cell influx 356 observed in Dectin-2 deficient mice did not affect fungal burden suggesting, that the 357 increased immune cells infiltrate did not result in more rapid clearance of infection. 358 Whether the increased inflammatory influx seen in Dectin-2 deficient mice could have 359 detrimental effects on the host, such as increasing lung damage during infection is 360

unknown. 361
Our current work addresses the mechanism by which cell-wall associated