Extracellular adenosine signaling reverses the age-driven decline in the ability of neutrophils to kill S. pneumoniae

The elderly are susceptible to serious infections by Streptococcus pneumoniae (pneumococcus), which calls for a better understanding of the pathways driving the decline in host defense in aging. We previously found that extracellular adenosine (EAD) shaped polymorphonuclear cell (PMN) responses, which are crucial for controlling infection. EAD is produced by CD39 and CD73, and signals via A1, A2A, A2B and A3 receptors. The objective of this study was to explore the age-driven changes in the EAD pathway and its impact on PMN function. We found in comparison to young mice, PMNs from old mice expressed significantly less CD73, but similar levels of CD39 and adenosine receptors. PMNs from old mice failed to efficiently kill pneumococci ex vivo; however, supplementation with adenosine rescued this defect. Importantly, transfer of PMNs expressing CD73 from young mice reversed the susceptibility of old mice to pneumococcal infection. To identify which adenosine receptor(s) is involved, we used specific agonists and inhibitors. We found that A1 receptor signaling was crucial for PMN function as inhibition or genetic ablation of A1 impaired the ability of PMNs from young mice to kill pneumococci. Importantly, activation of A1 receptors rescued the age-associated defect in PMN function. In exploring mechanisms, we found that PMNs from old mice failed to efficiently kill engulfed pneumococci and that A1 receptor controlled intracellular killing. In summary, targeting the EAD pathway reverses the age-driven decline in PMN antimicrobial function, which has serious implications in combating infections.


Summary 24
The elderly are susceptible to serious infections by Streptococcus pneumoniae 25 (pneumococcus), which calls for a better understanding of the pathways driving the 26 decline in host defense in aging. We previously found that extracellular adenosine (EAD) 27 shaped polymorphonuclear cell (PMN) responses, which are crucial for controlling 28 infection. EAD is produced by CD39 and CD73, and signals via A1, A2A, A2B and A3 29 receptors. The objective of this study was to explore the age-driven changes in the EAD 30 pathway and its impact on PMN function. We found in comparison to young mice, PMNs 31 from old mice expressed significantly less CD73, but similar levels of CD39 and 32 adenosine receptors. PMNs from old mice failed to efficiently kill pneumococci ex vivo; 33 however, supplementation with adenosine rescued this defect. Importantly, transfer of 34 PMNs expressing CD73 from young mice reversed the susceptibility of old mice to 35 pneumococcal infection. To identify which adenosine receptor(s) is involved, we used 36 specific agonists and inhibitors. We found that A1 receptor signaling was crucial for 37 Introduction surface of bone-marrow derived PMNs using flowcytometry. We found that compared to 117 young controls, PMNs from old mice expressed significantly less CD73, the EAD-118 producing enzyme, but significantly higher levels of the EAD-degrading enzyme ADA, 119 while the expression of CD39 was unchanged (Fig 2A). These findings demonstrate that 120 expression of EAD enzymes on PMNs are significantly altered with age. 121 122 CD73 expression on circulating PMNs decreases with age 123 Next, we wanted to confirm the importance of our findings in vivo. We compared 124 the expression of the EAD-pathway components on circulating PMNs in young and old 125 mice. We found that similar to what we observed in bone-marrow derived cells, 126 circulating PMNs in old mice expressed significantly lower levels of CD73 on their 127 surface ( Fig 2B). We did not find any differences in the expression of the rest of the EAD 128 pathway enzymes (CD39, ADA) on circulating PMNs ( Fig 2B). Further, circulating 129 7 different organs at 18 hours post infection. We found that transfer of WT PMNs from 140 young mice significantly reduced pulmonary bacterial burdens in old mice (Fig 3A). WT 141 PMNs further reduced the systemic spread of pneumococci resulting in a 50-fold 142 reduction in blood and brain bacterial loads when compared to no transfer controls ( Fig  143   3B and C). Importantly, transfer of WT PMNs from young mice ameliorated clinical 144 signs of the disease in aged hosts (Fig 3D). In contrast, transfer of PMNs from CD73 -/-145 young mice had no effect on reducing bacterial burdens or improving the disease score in 146 old recipients (Fig 3). These findings demonstrate that PMNs from young mice are 147 sufficient to boost resistance of old hosts to S. pneumoniae infection and that this 148 protection is dependent on CD73 expression by PMNs. 149 150 Supplementation with adenosine reverses the age-driven decline in PMN 151

antimicrobial function 152
Our above findings suggested that the ability of PMNs to produce EAD declines 153 with age. As circulating and bone marrow-derived PMNs exhibited mostly similar 154 phenotypes (Fig 1 and 2), for feasibility, we proceeded with experiments using PMNs 155 isolated from the bone marrow that are routinely used for ex vivo assays (Bou Ghanem et 156 al., 2015; Siwapornchai et al., 2020; Standish & Weiser, 2009). To directly determine if young wildtype mice produced 3-fold more EAD as compared to PMNs from either old 163 or CD73 -/mice ( Fig 4A). Upon pneumococcal infection, we observed a 6-fold increase 164 in EAD produced by PMNs from young mice, but no significant increase in PMN 165 cultures from old mice (Fig 4B). To rule out EAD release from dying cells as a potential 166 source for the observed differences between PMNs from old and young mice, we 167 compared apoptosis and necrosis using Annexin V/PI staining as previously described 168 (Siwapornchai et al., 2020). We found no difference in the number of apoptotic/necrotic 169 PMNs between the different mice groups (not shown), indicating that the differences in 170 EAD production is not due to differences in cell viability. 171 To test if supplementing with EAD rescues the antimicrobial function of PMNs 172 from old mice, we added EAD to the opsonophagocytic reactions. Strikingly, we found 173 that supplementation with 1µM EAD fully restored the ability of PMNs from old mice to 174 kill bacteria to a comparable level observed with young controls (Fig 4C). Taken 175 together, these findings suggest that the age-driven impairment in pneumococcal killing 176 is largely due to a decrease in EAD production, and that this impairment can be fully 177 reversed in vitro by supplementation with EAD. 178 highly expressed on PMNs as compared to the other receptors, while A3 expression was 186 the lowest (Supplemental Fig 1A). As the A1 receptor antibody is polyclonal, we further 187 confirmed A1 receptor expression on PMNs by western blots comparing wildtype (WT) 188 and A1 receptor knock-out (A1R -/-) mice (Supplemental Fig 1B). We did not observe any 189 differences in the expression of any of the adenosine receptors on bone-marrow derived 190 or circulating PMNs with age (Fig 2A and B). 191 To test which of the adenosine receptors were important for PMN antimicrobial 192 function, we treated PMNs from young mice with specific inhibitors for each of the 193 different receptors and compared their ability to kill bacteria ex vivo. We found that 194 blocking A1 receptor signaling significantly blunted the ability of PMNs from young 195 mice to kill S. pneumoniae (Fig 5A), while inhibition of the other three receptors did not 196 significantly impair bacterial killing by PMNs. Importantly, none of the adenosine 197 receptor inhibitors had a direct effect on bacterial viability (Supplemental Fig 2), 198 indicating that effect on bacterial killing was specifically due to PMN anti-bacterial 199 function. We further confirmed the role of A1 receptor in the anti-bacterial function of 200 PMNs by comparing the ability of WT, A1R -/+ and A1R -/-PMNs to kill bacteria ex vivo. 201 We found that the ability of A1R -/+ PMNs to kill pneumococci was severely impaired as 202 compared to WT controls, while A1R -/-PMNs completely failed to kill S. pneumoniae 203 ( Fig 5B). In fact, bacterial numbers increased in the presence of A1R -/-PMNs ( Fig 5B). 204 Taken together, these findings demonstrate that A1 receptor signaling is required for the 205 anti-pneumococcal activity of PMNs. 206 207 Activation of A1 receptor signaling rescues the ability of PMNs from old mice to kill 208

S. pneumoniae 209
Next we wanted to test if A1 receptor signaling plays a role in PMN function in 210 old mice. To test this, we treated PMNs from old mice with 2-Chloro-N6-Cyclopentyl 211 adenosine, an A1 receptor specific agonist, whose specificity we had confirmed before 212 (Bhalla et al., 2020). We then measured the ability of PMNs to kill bacteria ex vivo. We 213 found that in comparison to vehicle control, treatment with the A1 receptor agonist 214 boosted the ability of PMNs from old mice to kill S. pneumoniae more than two-fold ( Bacterial uptake was previously demonstrated to be important for killing by 235 PMNs (Standish & Weiser, 2009), therefore, we wanted to test if phagocytosis of 236 pneumococci was impaired with aging. To do so, we established a flowcytometry-based 237 bacterial uptake assay with GFP-expressing bacteria and inside-out staining (Smirnov,238 Solga, Lannigan, & Criss, 2015). We first infected PMNs with GFP-expressing S. 239 pneumoniae and differentiated between associated vs. engulfed bacteria by staining the 240 cells with anti S. pneumoniae antibodies (PE-labeled). We found that of all bacteria that 241 associated with PMNs within 15 minutes, 40% of them were engulfed (GFP + / PE -) 242 (Supplemental Fig 4A and B). We confirmed the validity of the assay using Cytochalasin 243 D which impairs phagocytosis (Standish & Weiser, 2009) and found that in its presence, 244 the majority of PMN-associated bacteria (~90%) remained extracellular (Supplemental 245 Fig 4A and B). When we compared hosts across age, we found no significant differences 246 in bacterial association or uptake between PMNs from young or old mice (Supplemental 247 Fig 4B and C). However, when we compared the amounts of engulfed bacteria that had 248 survived using a gentamicin-protection assays we had previously established 249 (Siwapornchai et al., 2020), we found that there were four-fold more viable bacteria in 250 PMNs from old mice as compared to young counterparts ( Fig 6A). These findings 251 demonstrate that while phagocytosis of S. pneumoniae does not change with age, the 252 ability of PMNs to kill the engulfed bacteria is significantly diminished in PMNs from old mice compared to young mice. 254

A1 receptor signaling controls intracellular killing of engulfed S. pneumoniae 256
Since intracellular killing of bacteria was impaired in aging, we wanted to 257 determine whether this was controlled by A1 receptor signaling. To test that, we treated 258 PMNs from young mice with the A1 receptor inhibitor and compared intracellular 259 survival of S. pneumoniae. We found that upon inhibition of A1 receptor signaling, the 260 number of viable intracellular bacteria in young PMNs almost doubled ( Fig 6B). On the 261 other hand, treatment of PMNs from old mice with the A1 receptor agonist significantly 262 reduced the amounts of intracellular bacteria that survived by half ( Fig 6C). The observed 263 differences in bacterial viability were not due to any effect on bacterial uptake by PMNs, 264 as inhibition or activation of A1 receptor had no effect on bacterial phagocytosis by 265 PMNs from young and old mice, respectively (Supplemental Fig 4B and C). Taken 266 together, these data demonstrate that A1 receptor signaling significantly contributes to the 267 ability of young PMNs to kill engulfed pneumococci and that this process is impaired 268 with aging. Importantly, activating A1 receptor signaling restores the ability of PMNs 269 from old mice to kill engulfed S. pneumoniae. significantly declined with aging, while levels of CD39 were unchanged. We also found 285 that surface expression of ADA, the enzyme that breaks down adenosine, was actually 286 higher in bone-marrow derived PMNs in old mice as compared to young controls. The 287 activity and levels of ADA were previously shown to decline in peripheral blood 288 lymphocytes of elderly donors (Crosti et al., 1987) and in senescent CD8 + T cells (Parish 289 et al., 2010). To our knowledge, this is the first study to demonstrate that aging is 290 accompanied by changes in EAD-pathway enzymes on PMNs. 291 The changes in EAD-pathway enzymes in aged mice were associated with lower 292 amounts of adenosine present in the extracellular milieu of PMN cultures from old mice 293 as compared to young counterparts. EAD can be either released into the extracellular 294 environment from intracellular compartments via equilibrative nucleoside transporters 295 (Boswell-Casteel & Hays, 2017), or can be produced as a breakdown product of 296 extracellular ATP, a process that requires CD39 and CD73 (Eltzschig,Macmanus,& 297 Colgan, 2008). At baseline the amount of EAD produced by PMNs from old mice was 298 comparable to that made by CD73 -/-PMNs, suggesting the amounts released were CD73-independent. However, PMNs from young mice, which expressed higher levels of CD73, 300 produced significantly more EAD even at baseline. Upon infection, we further observed a 301 significant increase in EAD production, but only by PMNs from young mice. 302 Pneumolysin, a toxin produced by S. pneumoniae (Marriott, Mitchell, & Dockrell, 2008), 303 triggers the release of extracellular ATP from infected cells (Domon et al., 2016;Hoegen 304 et al., 2011). Therefore, it is likely that in PMNs from young mice, the increase in EAD 305 production seen upon infection is due to the conversion of ATP into EAD by the 306 sequential action of CD39 and CD73, a process that is impaired in PMNs from old mice 307 due to blunted expression of CD73. 308 We previously found that CD73 expression and autocrine EAD production by 309 PMNs is crucial for the antimicrobial phenotype of PMNs during S. pneumoniae Adenosine binds its receptors with different affinities ranging from an EC50 <0.5µM for 316 the high affinity receptors A1 and A3, to an EC50 >0.6µM for the intermediate affinity 317 receptor A2A and an EC50 between 16-64µM for the low affinity A2B receptor (Hasko et 318 al., 2008). Here, we found that supplementation of PMNs with 1µM of adenosine was 319 sufficient to restore the antimicrobial function of PMNs from old mice suggesting that the 320 higher affinity receptors were involved. In fact, using a combination of pharmacological of PMNs to kill S. pneumoniae. Importantly, triggering this receptor reversed the age 323 driven defects in the ability of PMNs to kill these bacteria. This is in line with our 324 previous findings demonstrating that activating A1 receptor signaling in vivo boosts the 325 resistance of old mice to pneumococcal pneumonia by enhancing pulmonary bacterial 326 clearance and reducing systemic spread of the infection (Bhalla et al., 2020). 327 In exploring mechanisms of why PMN function was impaired with aging we 328 examined ROS production, which we and others have previously reported to be 329 controlled by adenosine ( Bogaert, 2013), we found that ROS production was not altered with aging. Similarly, we 333 found that release of antimicrobial MPO and CRAMP as well as phagocytosis were not 334 impaired in response to S. pneumoniae infection. Previously A1 receptor was shown to 335 enhance phagocytosis by PMNs (Salmon & Cronstein, 1990), which we did not observe 336 in our study. However, the previous work was performed using erythrocyte coated beads 337 and stimulation with FMLP or PMA Salmon & Cronstein, 1990), 338 while here we are performing our studies with live bacteria that can actively modulate 339 PMN responses. Rather, we found that PMNs from old mice failed to efficiently kill 340 engulfed bacteria. The fate of intracellular S. pneumoniae and the mechanisms of 341 clearance of these engulfed bacteria have not been fully elucidated. We identify here A1 342 receptor signaling as the host pathway controlling intracellular killing of pneumococci. 343 Importantly, activation of this receptor enhanced the ability of PMNs from old mice to 344 kill intracellular bacteria.
In conclusion, we demonstrate here for the first time, that immunosenescence of 346 PMNs is shaped by the EAD pathway. We find that the age-driven impairment in 347 bacterial killing by these innate immune cells is largely due to a decrease in EAD 348 production and signaling. Importantly, supplementing with EAD and triggering A1 349 receptor signaling fully reverses the age-driven decline in PMN antimicrobial function.  Todd-Hewitt broth supplemented with 0.5% yeast extract and oxyrase untill mid-369 exponential phase. Aliquots were frozen at -80˚C in the growth media with 20% (v/v) 370 glycerol. Prior to use, aliquots were thawed on ice, washed and diluted in PBS. Bacterial 371 concentrations were confirmed by serial dilution and dribble plating on Tryptic Soy Agar 372 plates supplemented with 5% sheep blood agar. 373

Flowcytometry 374
Whole blood was collected from mice using cardiac puncture using EDTA as an anti- PMNs were isolated from the bone marrow through density gradient centrifugation, using 392 Histopaque 1119 and Histopaque 1077 as previously described (Swamydas & Lionakis,393 2013). PMNs were isolated from the circulation through density gradient centrifugation, 394 using a Percoll gradient as previously described (Siwapornchai et al., 2020). The isolated 395 PMNs were resuspended in Hanks' Balanced Salt Solution (HBSS)/0.1% gelatin without 396 Ca 2+ and Mg 2+ , and used in subsequent assays. Purity was measured by flow cytometry 397 using CD11b and Ly6G and 85-90% of enriched cells were positive Ly6G/CD11b + . 398

Adoptive Transfer of PMNs 399
Bone marrow PMNs were isolated from uninfected mice as described above and 2.5x10 6 400 cells were adoptively transferred as previously described (Siwapornchai et al., 2020). 401 Control groups received PBS. One hour following transfer, mice were challenged intra-402 tracheally with 5x10 5 CFU of S. pneumoniae. Twenty-four hours post infection, mice 403 were scored for clinical signs of the disease ranging from healthy (0) to severely sick (21) 404 as previously described (Bhalla et al., 2020). Mice were euthanized and the lungs, brain, 405 and blood were collected and plated on blood agar plates for CFU. 406

Opsonophagocytic (OPH) Killing Assay 407
The ability of PMNs to kill S. pneumoniae ex vivo was measured using a well-established 408 Opsonophagocytic (OPH) killing assay as previously described (Bou Ghanem et al., sequences were amplified using primers hlpA-up-F and hlpA-link-R (Table 1). The sfgfp 486 gene was amplified from plasmid pTHSSd_34 using primers gfp-link-F and gfp-R-spec 487 (Table 1). pTHSSd_34 was a gift from Christopher Voigt (Addgene plasmid # 59960). 488 The aad gene encoding spectinomycin resistance was amplified from pMagellan6 using 489 the primers spec-F-gfp and spec-R-hlpA (Table 1). The region downstream of hlpA was 490 amplified with primers hlpA-down-F-spec and hlpA-down-R (Table 1). The four linear 491 pieces of DNA were mixed in equimolar amounts, and splicing by overlap extension PCR 492 was used to assemble one linear piece of DNA, which was amplified using the primers 493 hlpA-up-F and hlpA-down-R (Table 1). The resulting linear DNA was transformed into 494   where each condition was tested in triplicate (n=3 technical replicates) per experiment. 582 Asterisks indicate significant differences calculated by one-way ANOVA followed by 583 Tukey's test. 584