key: cord-0268266-1fjybxtx
authors: Cipolla, Ellyse M.; Yue, Molin; Nickolich, Kara L.; Huckestein, Brydie R.; Antos, Danielle E.; Chen, Wei; Alcorn, John F.
title: Heterotypic Influenza Infections Mitigate Susceptibility to Secondary Bacterial Infection
date: 2022-04-12
journal: bioRxiv
DOI: 10.1101/2022.04.12.488032
sha: 449e2256835278139b3c9be155f209448fd217c7
doc_id: 268266
cord_uid: 1fjybxtx

Influenza associated bacterial super-infections have devastating impacts on the lung and can result in increased risk of mortality. New strains of influenza circulate throughout the population yearly promoting the establishment of immune memory. Nearly all individuals have some degree of influenza memory prior to adulthood. Due to this we sought to understand the role of immune memory during bacterial super-infections. An influenza heterotypic immunity model was established using influenza A/PR/8/34 and A/X31. We report here that influenza experienced mice are more resistant to secondary bacterial infection with methicillin-resistant Staphylococcus aureus as determined by wasting, bacterial burden, pulmonary inflammation, and lung leak, despite significant ongoing lung remodeling. Multidimensional flow cytometry and lung transcriptomics revealed significant alterations in the lung environment in influenza-experienced mice compared with naïve animals. These include changes in the lung monocyte and T cell compartments, characterized by increased expansion of influenza tetramer specific CD8+ T cells. The protection that was seen in memory experienced mouse model is associated with the reduction in inflammatory mechanisms making the lung less susceptible to damage and subsequent bacterial colonization. These findings provide insight into how influenza heterotypic immunity re-shapes the lung environment and the immune response to a re-challenge event, which is highly relevant to the context of human infection.

To determine the role that immune memory plays in influenza associated super-infections we 247 primed wild-type (WT) C57BL/6 mice with influenza A/X31 H3N2 (X-31) on day zero followed 248 by re-challenge with a heterotypic influenza strain, A/PR/8/34 H1N1 (PR8), and inoculation with 249 methicillin-resistant Staphylococcus aureus USA300 (MRSA) six days after PR8 infection (Fig.  250 1a). Super-infected mice that had pre-existing heterotypic influenza immunity were better 251 protected upon bacterial infection, with decreased weight loss and bacterial burden than their 252 acute infection counterparts ( Fig. 1b and c) . When lung infiltrating cells were assessed, super-253 infected memory mice had lower levels of total infiltrating cells, and upon further assessment it 254 was determined that monocytes made up more of the total cell distribution (Fig. 1d) . Previous 255 research on influenza associated bacterial super-infections have elucidated several potential 256 mechanisms to understand the driving forces behind these types of infections with one of the 257 prevailing ideas being that the anti-viral mechanisms in the lung environment dampen the anti-258 bacterial response (7, 9, 35, 36) . We observed that memory super-infected mice at the time of 259 MRSA challenge had non-detectable levels of PR8 gene expression in their lungs indicating that 260 the anti-viral memory mechanisms controlled the infection, potentially ensuring a more effective 261 anti-bacterial response to secondary bacterial infection (Fig. 1e) . Finally, we observed a 262 reduction in levels of Type I, II, and III interferons (Ifn), which are important signaling 263 molecules during viral infection and have been shown to interfere with antibacterial response and 264 survival outcome in secondary bacterial infections (17, (37) (38) (39) (Fig. 1f) . These data suggest that 265 mice with influenza memory have an advantage over their previously naïve acute infection 266

Lung injury is present in acute and memory influenza challenged mice. 270

Since viral burden was effectively controlled in heterotypic influenza challenged mice, we 271 investigated if lung injury was altered in our model. Memory and acute super-infected mice were 272 equally susceptible to lung damage following PR8 and MRSA challenge; however, the memory 273 mice displayed areas of epithelial remodeling and metaplasia ( Fig. 2a and b) . Interestingly, when 274

we looked at other measures to assess tissue integrity, we found that memory mice had 275 attenuated levels of IgM and protein in their BALF when compared to their acute counterparts 276 (Fig. 2c ). The respiratory system has several mechanisms to defend itself against pathogens and 277

to ensure a balance in the lung environment, because of these protective factors we also 278 examined the expression of genes associated with lung function, protection, and remodeling. 279

Memory experienced mice displayed elevated expression levels of important molecules involved 280 in the mucociliary escalator: mucin 5b (muc5b) and mucin 5ac (muc5ac) (Fig. 2d) . Further, 281 secretoglobin family 1a member 1 (scgb1a1), a key club cell protein, and forkhead box j1 282 (foxj1), a key transcription factor in ciliated epithelial cells, were elevated in memory super-283 infection versus acute infection (Fig. 2e) . Interestingly, we did see reductions in gene expression 284 of epithelial remodeling markers in the memory mice most notably in collagen 1 a 1 (col1a1) and 285 αsmooth muscle actin (acta2), but saw no change in tight junction protein (tjp1) (fig 2f) . These 286 data indicate that heterotypic memory to influenza drives changes in the lung environment that 287 differ from their previously naïve acute counterparts and may hinder opportunistic bacterial 288

To study the broad-spectrum changes in the lung environment of influenza memory super-293 infected mice and their previously naïve acute counterparts we compared gene expression 294 changes in influenza memory versus acute infected mice via bulk-RNA sequencing analysis. We 295 performed a deconvolution analysis on our day seven bulk-RNA sequencing dataset using a 296 single-cell reference dataset of combined PBS and 48-hour influenza infected mice (40). 297

Deconvolution analysis revealed changes in the proportions of genes associated with 298 granulocytes, mononuclear phagocytic cells, epithelial cells, and α-sma between memory and 299 acute infected groups (Fig. 3a) . Influenza memory experienced mice had a transcriptomic 300 signature driven more by epithelial cells and less by granulocytes compared with acute infection 301 controls. Sequencing analysis showed clustering of samples within groups (Fig. 3b) . When up or 302 down-regulated genes were assessed in memory versus acute infection we found that memory 303 mice displayed a reduction in genes typically associated with inflammation and control of 304 inflammatory mechanisms such as interferon gamma (Ifn-γ), granzyme B (Gzmb), and 305 interleukin (IL)-10 ( Fig. 3c ). Next, we ran gene ontology analysis on our sample groups and 306

found that the pathways activated in memory are associated with the epithelium and mucociliary 307 mechanisms, whereas those that are suppressed deal with response to virus, and immune effector 308 processes as well as the inflammatory response (Fig. 3d) . Gene ontology biological pathway 309 analyses revealed that the top enriched pathways dealt with leukocyte migration, response to 310 interferon-beta, response to chemokine, as well as chemotaxis (Fig. 3e ). Together these data 311 demonstrate that the controlled lung environment of influenza memory experienced mice is 312 mechanisms potentially accounting for a better immune response to a secondary bacterial 314 infection. 315

populations play a vital role in clearance of influenza infections, therefore, we looked at how the 321 T cell compartment changes in influenza memory versus acute super-infected mice. Using 322 clustering (flowSOM) and dimensionality reduction (UMAP) methods, we tracked marker 323 expression changes in T cell populations in mouse lungs. From our gating strategy we found that 324 ( Supplementary Fig 1) , in an acute super-infection there are a higher proportion of CD4 + FoxP3 + 325 regulatory T cells, this significant change was seen in both percentage of parent gate as well as 326 absolute cell counts (Fig. 4a-c) . The shift towards a lower proportion of regulatory T cells in the 327 memory lung environment could indicate an earlier controlled inflammatory response towards 328 viral infection allowing for an unimpeded switch towards immune mechanisms generated in 329 response to bacterial infections. Further, we observed an increase in the proportion of CD8 + NP + 330 tetramer-positive T cells, an immunodominant epitope, in memory experienced mice when 331 compared to the acute infected mice. This was again confirmed by significant differences in cell 332

proportions and absolute cell counts consistent with increased memory expansion (Fig. 4a, b, d) . 333

We next looked at the expression of cytokines associated with T cell migration, proliferation, and 334 regulation. We found that protein levels of pro-inflammatory cytokines IL-12p40 and IL-12p70 335 saw a trend toward higher levels of IL-17a and IL-22 via gene expression, although this was not 340 significant, however, we did see a significant increase in expression of the Type 17 immune 341 promoting cytokines IL-23 and IL-1β (Fig. 5e ). These data demonstrate that the T cell 342 compartment in memory experienced mice is characterized by an influx of tetramer positive 343 CD8 + T cells that are better able to control influenza infection, thus creating an environment that 344 is potentially oriented towards improved bacterial immunity. 345 346 Heterotypic influenza memory alters the innate immune cell compartment during bacterial 347

Finally, we explored the impact of heterotypic influenza memory on innate immune cell 349 populations, which are imperative in defense against bacterial infections. We used flowSOM and 350

UMAPs to visualize changes in the myeloid compartment when mice have an established 351 memory compartment. We found that neutrophils make up the largest proportion of myeloid 352 cells in both treatment groups ( Fig. 6a and b ). Using our myeloid gating strategy (Supplementary 353 Fig 2) , we also found that the total number of natural killer (NK) cells, monocytes, macrophages, 354 and eosinophils were significantly reduced in the influenza memory mice as compared to their 355 acute counterparts (Fig. 6c) . Interestingly, no change was seen in the total number of neutrophils 356 between both groups (Fig. 6c ). Next, we assessed changes in common cytokines and markers 357 associated with myeloid cells and their function via relative gene expression as well as protein 358 expression. We first looked at two mediators of the inflammatory response, the alarmins IL-1α 359 and IL-33. Both of these cytokines were found to be significantly elevated in the lungs of 360 memory mice via protein expression or relative gene expression (Fig. 6d ). When we looked at 361 markers that were significantly downregulated in the influenza memory mice we observed a 362 decrease in protein expression of monocyte chemoattractant protein-1 (mcp-1), macrophage 363 inflammatory protein-1 beta (mip-1β), macrophage inflammatory protein-1 alpha (mip-1α), and 364 eotaxin (Fig. 6e) , which suggests a less inflamed environment. Lastly, we found that arginase-1 365 (arg1), cathepsin g (ctsg), nitric oxide synthase 2 (nos2), and amphiregulin (areg) expression 366

were also downregulated in memory mice (Fig. 6f) . The promotion of an inflammatory balance 367 in the lung environment through the stages of infection from influenza to secondary bacterial 368 infection in the heterotypic memory mice may create an environment that is suitable for bacterial 369 clearance. 370 371 372 applicable across many respiratory viral diseases including the recent Covid-19 pandemic (41). 375

Influenza associated secondary bacterial infections are multi-faceted infections that employ 376 multiple immune cell players of both the adaptive and innate immune systems to cover the broad 377 spectrum of the host response from viral infection to bacterial infection. It is widely established 378 that during an acute secondary bacterial infection the anti-viral response hinders the anti-379 bacterial mechanisms that are required to protect against opportunistic pathogens such as MRSA 380 (7, 9, 12, 18-20, 36, 39, 42-46) . Secondary bacterial infection also results in heightened 381 immunopathology driven by an overactive immune landscape ultimately creating an ideal 382 environment for opportunistic pathogens (13). The results presented herein expand on the 383 prevailing views in the field generated in previously naive animals, and suggest that memory to 384 influenza infections creates a balanced lung environment at the time of secondary bacterial 385 infection susceptibility resulting in a more effective anti-bacterial response. 386

The epithelium represents the first line of defense against respiratory pathogens and assault from 388 environmental stressors. Influenza directly targets the epithelial barrier leading to extensive 389 damage and impairment of the mucocilliary escalator resulting in an environment that heavily 390 favors bacterial adherence and colonization (42, 47) . In the present study we found that influenza 391 memory experienced mice have an altered lung epithelial environment as compared to their 392 previously naïve acute infection counterparts. Interestingly, our data shows that the influenza 393 memory lung still undergoes significant damage due to influenza infection. However, we 394 observed preservation of epithelial gene expression of structural and functional markers in 395 (also known as club cell secretory protein), which has been shown to promote alveolar 397 macrophage survival and response to inflammation (48). Deconvolution analysis also suggested 398 that memory mice expressed a higher proportion of the lung gene signature associated with the 399 epithelium. Further, we observed upregulation of markers associated with cilium and ciliary 400 movement, as well as mucus production. Taken together these data suggest that influenza 401 memory doesn't prevent lung injury, rather the epithelium is re-programmed in a manner that 402 favors the enhancement of bacterial clearance. Barrier integrity is also targeted by aberrant 403 immune mechanisms including enhanced cytokine production in response to influenza that in 404 cases of severe viral infections can cause severe edema (49). Lung leak was also limited in the 405 context of heterotypic influenza memory. Aberrant immune activation, such as cytokine storm, is 406 key to influenza related lung injury. Our data suggests that there is a reduction in production of 407 cytokines that are detrimental and associated with immunopathology, in particular mcp-1, mip-408 1α, cxcl10, and IL-6 (49-52). aureus as well as MRSA due to the role it plays in neutrophil recruitment to the infected lung 433 (55). IL-1β production has been shown to be impaired by preceding influenza infection during 434 secondary bacterial super-infections. This impairment results in a detrimental outcome for the 435 host because of its ability to influence and activate the Type 17 pathway, an important pathway 436 during antibacterial response against numerous extracellular pathogens including S. aureus (56). 437

The IL-12 family has also been implicated in its role in protection against bacterial infections. In 438 our model, we found that three members of the IL-12 family: IL-12p40, IL-12p70 and IL-23 439 were increased. IL-12p70 and IL-23 have subsequently been shown to play a role in bacterial 440

super-infections in human cells due to their alteration by influenza resulting in impairment of the 441 cytokines, IL-17a and IL-22 in our influenza memory versus acute infected mice, we did see 447

increased IL-1β, IL-23 and decreased Ifn-γ, suggesting promotion of Type 17 immunity. 448

Together these data indicate that the heterotypic memory mice display increases in pathways 449 related to bacterial defense and further studies on these pathways should be done to elucidate 450 direct mechanisms. 451

The lung is a dynamic organ that requires a state of homeostasis characterized by a balance in 453 pro-inflammatory and anti-inflammatory factors (59). Influenza infections result in a highly 454 inflamed lung environment that is characterized by an influx in immune cells and their 455 mediators. To return to homeostasis, the lung initiates anti-inflammatory mechanisms. One of the 456 key primary anti-inflammatory mediators of the immune system is Foxp3 + regulatory T cells. 457

These cells expand in response to an overly inflamed mucosal environment to dampen the 458 immune response and have been shown to acquire memory recall for targeted pathogen 459 responses during re-challenge events (60, 61) which are imperative to bring the inflamed 460 environment back to homeostasis. In our study we saw a reduction in the proportion and counts 461 of FoxP3 + regulatory T cells in heterotypic memory experienced mice, as well as a reduction in 462 the production of the anti-inflammatory cytokine IL-10. IL-10 has further been implicated in 463 impairment of immune defense against influenza associated secondary bacterial infections with 464 environment of the influenza memory mice is targeted against the virus early on in re-challenge, 466 which is reflected in the lower number of regulatory T cells present at time of bacterial infection. 467

Our data further supports this notion with the increase seen in tetramer specific T cells in 468 memory mice and reduction in innate immune cells that are elevated in acute infections and are 469 associated with aberrant pro-inflammatory responses. 470

In order to promote an environment that is beneficial for bacterial clearance and limits 472 immunopathology caused by aberrant immune activation, a balancing act must be struck between 473

anti-viral and anti-bacterial host defense. As shown herein, this balance can be accomplished in Greenplate, E. J. Wherry, and U. C. P. Unit ‡. 2021. mRNA vaccines induce durable 585 immune memory to SARS-CoV-2 and variants of concern. Science 374: abm0829. sections. Lung injury scoring was carried out on three areas of the lung: parenchyma, artery, and 750 an IgM ELISA (acute n=11, memory n=12). BAL protein was quantified from BAL collected at 752 day of harvest (PBS n=3, acute n= 10, memory n= 12). D-F) RNA was extracted from mouse 753 lungs and made into cDNA which we then used to probe for various genetic targets (muc5b, 754 muc5ac, scgb1a1: acute n=8, memory n=8; foxj1: PBS n=6,acute n=7, memory n=5; 755 col1a1:acute n=6, memory n=7; acta2: acute n=12, memory n=8; tjp1: acute n=11, memory 

Global Influenza Programme: Burden of disease

IFN-gamma-inducible protein 10 (IP-10; CXCL10)-deficient mice reveal a role for IP-10

in effector T cell generation and trafficking

Type I IFNs mediate development of postinfluenza 664 bacterial pneumonia in mice

Murine Type III interferons are functionally 667 redundant and correlate with bacterial burden during influenza/bacterial super-infection

Novel protective mechanism for interleukin-33 at the mucosal barrier during 671 influenza-associated bacterial superinfection

Influenza A exacerbates Staphylococcus aureus pneumonia by 674 attenuating IL-1β production in mice

Approximation and Projection for Dimension Reduction

FlowSOM: Using self-organizing maps for visualization 707 and interpretation of cytometry data

Critical role of IL-17RA in immunopathology of 710 influenza infection

Differential requirement for c-Jun N-terminal kinase 1 in lung 713 inflammation and host defense

Moderated estimation of fold change and 715 dispersion for RNA-seq data with DESeq2

clusterProfiler: an R package for 717 comparing biological themes among gene clusters

Determining cell type abundance and expression from bulk tissues with digital 721 cytometry

Robust enumeration of cell subsets from tissue 724 expression profiles

A)