Reg3β removes aged neutrophils after myocardial infarction

myocardial infarction , innate immune response, neutrophils, resolution of 30 inflammation 31 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted February 20, 2026. ; https://doi.org/10.64898/2026.02.19.706868doi: bioRxiv preprint 2

32 Recruitment and resolution of immune cell accumulation after myocardial infarction (MI) is 33 critical for effective wound healing and prevent ion of maladaptive remodeling . Numerous 34 signals are known to recruit neutrophils, critical drivers of inflammation, but knowledge about 35 signals containing and resolving their accumulation is limited. We discovered that 36 Regenerating-islet derived protein 3 beta (REG3β) limits the persistence of neutrophils after 37 MI, thereby promoting resolution of inflammation. REG3β selectively binds to aged and 38 hyperactive neutrophils and induces rapid cell death , enabling clearance via macrophage -39 mediated efferocytosis. Selective binding of REG3β is achieved by interaction with 40 paucimannosylated proteins that translocate from azurophile granules to the plasma 41 membrane in an activation- and age-dependent manner. Endocytotic uptake and accumulation 42 of REG3β in lysosomes initiates programmed cell death of neutrophils via lyosomal membrane 43 permeabilization and release of cathepsins. Our work establishes REG3β as a local immune 44 checkpoint essential for neutrophil resolution and cardiac repair. 45 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted February 20, 2026. ; https://doi.org/10.64898/2026.02.19.706868doi: bioRxiv preprint 3 MAIN 46 The myocardium undergoes a series of wound healing processes in response to damage, 47 which eventually result in formation of scars and cardiac remodeling. Immediately following 48 destruction of myocardial tissue different waves of immune cells arrive, a mandatory 49 prerequisite for cardiac repair1. The first type of immune cells that infiltrate the infarcted heart 50 in higher numbers are circulating neutrophil granulocytes2,3. Neutrophils are recruited to sites 51 of injury for removal of cellular debris but also for release of tissue-degrading enzymes, 52 cytokines, and reactive oxygen species (ROS), which are parts of the inflammatory 53 response4,5. On the other hand, n eutrophils possess anti-inflammatory and proangiogenic 54 properties that facilitate cardiac repair 6,7. Despite these beneficial effects, accumulation or 55 belated removal of neutrophils have deleterious consequences, often leading to cardiac 56 rupture8. Likewise, elevated neutrophil numbers in the blood are associated with worsened 57 prognosis and increased mortality after MI9. The seemingly conflicting functions of neutrophils 58 when coping with consequences of MI may be explained by differential activities and dynamics 59 of different neutrophil subsets. For example, aged neutrophils represent an exceedingly active 60 subset, promoting inflammatory conditions . Tight control of infiltration, maintenance , and 61 persistence of different subsets of neutrophils appears to be crucial for effective wound healing 62 after MI and for avoiding maladaptive remodeling of the heart10. 63 The mechanisms regulating recruitment of neutrophils in response to MI are complex but well 64 characterized11,12. In contrast, less is known about local circuits controlling maintenance and 65 removal of different subpopulations of neutrophils after MI. It is assumed that clearance of 66 neutrophils from infarcted hearts is primarily initiated by apoptosis of neutrophils, followed by 67 macrophage-mediated efferocytosis 13. Activation of the extrinsic apoptotic pathway in 68 neutrophils has been claimed to rely on cell death receptors, namely Fas protein (also called 69 CD95 or APO -1) and receptors for TNFα 14. However, experimental evidence about the 70 importance and specificity of individual ligands, critical for controlling the life time of different 71 types of neutrophils in the infarcted heart, is mostly missing. In addition to apoptosis, other 72 types of cell death have been described to contribute to the removal of neutrophils in different 73 tissues, such as necroptosis, pyroptosis, necrosis, and NETosis14. 74 We previously described that inactivation of Regenerating-Islet derived protein 3 beta (REG3β) 75 attenuates recruitment of monocytes to the infarcted heart, presumably prolonging persistence 76 of neutrophils. Genetic inactivation of Reg3β frequently causes cardiac rupture and decreases 77 cardiac function in surviving animals after MI8. REG3β belongs to the C-type lectin superfamily 78 and is secreted by cardiomyocytes in the border zone of infarcts. The receptors for REG3β or 79 other REGs are essentially unknown. Claims have been made for different membrane proteins 80 to act as receptors for individual REGs, but conclusive evidence is missing and no direct link 81 to intracellular processes downstream of REGs exists15. In this study, we discovered that 82 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted February 20, 2026. ; https://doi.org/10.64898/2026.02.19.706868doi: bioRxiv preprint 4 REG3β has direct cytotoxic activity on neutrophils in the infarcted heart, primarily targeting 83 aged neutrophils, which is instrumental for spatial and temporal control of neutrophil presence. 84 REG3β exerts its function by binding to distinct paucimannosidic carbohydrate motifs on the 85 cellular surface of aged and hyperactive neutrophils . Bound R EG3β enters neutrophils via 86 endocytosis and accumulates in lysosomes, causing lysosomal membrane permeabilization. 87 Subsequent release of cathepsins and exposure of p hosphatidylserine enable clearance of 88 neutrophils via macrophage-mediated efferocytosis. 89 90 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted February 20, 2026. ; https://doi.org/10.64898/2026.02.19.706868doi: bioRxiv preprint 5

91 Reg3b limits persistence and spatial expansion of neutrophils in the heart after 92 MI 93 Neutrophils were profiled and quantified in infarcted hearts of wildtype ( WT) and Reg3b 94 deficient (Reg3b-/-) mice by multicolor flow cytometry at different time points after permanent 95 ligation of the left anterior descending (LAD) coronary artery . We observed comparable 96 numbers of CD45hi/CD11bhi/Ly6Ghi cardiac tissue neutrophils in WT and Reg3b-/- mice at day 97 1 and day 2, suggesting that recruitment of neutrophils does not depend on Reg3b (Fig. 1a). 98 In contrast, infarcted hearts of Reg3b-/- mice contained significantly more neutrophils at day 3 99 and day 4 compared to control mice. Neutrophils substantially declined in WT mice between 100 days 3 and 4, whereas the reduction of neutrophils in infarcted hearts of Reg3b-/- mice was 101 moderate (Fig. 1a). No differences in the concentration of neutrophils were detected in the 102 blood of WT and Reg3b-/- mice at day 4 after MI, indicating that the reason for neutrophil 103 accumulation in Reg3b-/- hearts lays within and not outside the organ (Fig. 1 b). Additional 104 live/dead cell stain analysis identified increased ratios of viable 7-Aminoactinomycin D (7AAD)-105 /Annexin V (AnnV)- neutrophils in Reg3b-/- mice hearts, accompanied by a decrease of dead 106 7AAD-/AnnV+ neutrophils (Fig. 1c). 107 To analyze whether REG3β not only regulates the temporal persistence of neutrophils after MI 108 but also their s patial distribution we performed immunofluorescence staining. Labeling of 109 LY6G+ neutrophils uncovered a pronounced expansion of neutrophils from the infarct (IZ) to 110 the left ventricular remote zone (RZ) in Reg3b-/- hearts at day 3 and 4 after MI (Fig. 1d–f; Suppl. 111 Fig. 1 ). Fractionation of infarcted hearts into IZ and RZ followed by western blot analysis 112 detected increased concentration of neutrophil-derived proteases including Myeloperoxidase 113 (MPO), Neutrophil elastase (ELANE ), and Matrix metalloproteinase-9 (MMP-9) in the RZ at 114 day 4 after infarct, corroborating enhanced spatial expansion of neutrophils in Reg3b-/- hearts 115 (Suppl. Fig. 2a, b). 116 We reasoned that the enhanced levels of neutrophil -derived proteases might be responsible 117 for the increased incidence of cardiac rupture in Reg3b-/- mice, which occurs at the intersection 118 between injured and healthy tissue 8,16 (Suppl. Fig. 2 c). To validate this hypothesis , we 119 transiently depleted neutrophils by injections of anti-LY6G lytic antibodies into Reg3b-/- mice at 120 day 2, 3, 4, and 5 after MI (Fig. 1g). Anti-LY6G injections efficiently prevented cardiac rupture 121 in Reg3b-/- mice after MI and increased survival to WT levels (Fig. 1h). Moreover, magnetic 122 resonance imaging-based analysis of heart function revealed that antibody-mediated depletion 123 of neutrophils decreased endsystolic volume and increased ejection fraction of Reg3b-/- mice 124 14 days after MI (Fig. 1i, j ). The normalization of heart function by injection of anti -LY6G 125 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted February 20, 2026. ; https://doi.org/10.64898/2026.02.19.706868doi: bioRxiv preprint 6 antibodies demonstrate s that accumulation of neutrophils is responsible for the impaired 126 function of Reg3b-/- hearts after MI. 127 REG3β directly interacts with neutrophils and initiates cell death 128 REG3β may affect maintenance of neutrophils either indirectly or by direct interactions. Three-129 dimensional visualization of LY6G+ neutrophils and REG3β protein in mouse hearts after MI 130 via light sheet fluorescence microscopy demonstrated a strong overlap of signals, particularly 131 at the intersection between IZ and RZ (Fig. 2a). Colocalization analysis based on 3D light sheet 132 fluorescence data for LY6G and R EG3β in whole infarcted hearts yielded a Pearson’s 133 correlation coefficient of 0.55 (Fig. 2b ). Localization of R EG3β and the human orthologue 134 REG3A on neutrophils (LY6G+ neutrophils in mice and CD66b + neutrophils in humans) was 135 also confirmed by high-resolution fluorescence imaging of infarcted mouse and heart samples 136 from human patients diagnosed with MI (Fig. 2c, d; Suppl. Fig. 3a, b). Flow cytometry analysis 137 uncovered an increase of R EG3β-binding (REG3βpos) neutrophils in the heart after MI, 138 whereas the numbers of REG3βpos neutrophils in the bone marrow and blood were much lower 139 and not affected by MI ( Fig. 2e, f ). Notably, we observed comparable ratios of R EG3βpos 140 neutrophils in cardiac tissue samples in humans and mice (approximately 11% in human and 141 15% in mice) (Fig. 2d, e). 142 We previously described that the main source of R EG3β in the infarcted heart are 143 cardiomyocytes8,17. Nevertheless, we wanted to make sure that the R EG3β signal on 144 neutrophils is indeed derived from secreted R EG3β and not from REG3β translated within 145 neutrophils. Bulk RNA -sequencing of sorted R EG3βneg and REG3βpos neutrophils from 146 infarcted hearts did not detect any transcripts of Reg3b in either population, corroborating 147 previous data (Fig. 2g). Interestingly, the ratio of neutrophils with intracellular localization of 148 REG3β was substantially higher than the ratio of neutrophils with REG3β at the cell surface, 149 suggesting that bound REG3β is rapidly taken up by neutrophils (Suppl. Fig. 4a, b). Even more 150 importantly, REG3βpos neutrophils, either intracellular or membrane bound, showed reduced 151 viability (Fig. 2h, I; Suppl. Fig. 4c). 152 The observation that R EG3βpos neutrophils showed reduced viability prompted us to test 153 potential direct cytotoxic effects of REG3β on neutrophils. We isolated mouse neutrophils from 154 peritoneal cavities after induction of peritonitis with casein, treated them with different 155 concentrations of REG3β, and monitored cell death with the IncuCyte Cytotox Green Assay in 156 real time. We observed a rapid increase of dead Cytotox+ neutrophils already after 60 minutes. 157 Potent cytotoxic effects were recorded at 10 ng ml -1 and 100ng ml -1 REG3β, whereas 158 cytotoxicity declined at 1000ng ml -1, probably due to enhanced aggregation of R EG3β (Fig. 159 2j). Lactate dehydrogenase (LDH) release assays confirmed cytotoxic effects of recombinant 160 REG3β protein on activated neutrophils. In stark contrast , bone marrow- and blood-derived 161 neutrophils, which are in a more quiescent and non-inflammatory state, were not responding 162 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted February 20, 2026. ; https://doi.org/10.64898/2026.02.19.706868doi: bioRxiv preprint 7 to REG3β (Fig. 2k)18. Importantly, interferon gamma (IFNγ), which was used as a positive 163 control did not show any selectivity in respect to activated or non-inflammatory states but 164 exerted cytotoxic effects on all types of neutrophils (Fig. 2k). 165 REG3β binds to hyperactive and aged neutrophils 166 Our data indicated that REG3β did not bind to all but a subset of neutrophils. To characterize 167 REG3βpos neutrophils more closely, we performed RNA sequence analysis of R EG3βneg and 168 REG3βpos neutrophils from infarcted hearts. Subsequent principal component analysis 169 separated REG3βneg and REG3βpos neutrophils into two distinct subsets (Fig. 3a). Compared 170 to R EG3βneg neutrophils 1124 genes were up - and 285 genes were downregulated in 171 REG3βpos neutrophils (Fig. 3b). 172 Selective p athway enrichment analysis using KOBAS (KEGG Orthology -Based Annotation 173 System) demonstrated increased expression of genes related to cell death in R EG3βpos 174 neutrophils. Furthermore, we detected increased expression of markers for nearly all critical 175 neutrophil effector functions, including cell adhesion, chemotaxis, activatio n, phagocytosis, 176 ROS metabolism, metallopeptidase and cytokine activity ( Fig. 3c). Flow cytometric analysis 177 corroborated enhanced presence of neutrophil activation markers such as CD54, TLR2, CD63, 178 and TREM2 in R EG3βpos compared to pan and R EG3βneg neutrophils (Fig. 3d–g). Levels of 179 the ROS producing enzyme NOX2 and the metallopeptidases MMP -9 and ADAM9 were 180 substantially elevated in R EG3βpos neutrophils ( Fig. 3 h–j). We also measured increased 181 production of ROS with DHR123 and increased phagocytotic activity by incorporation of 182 pHrodo and cardiomyocyte-derived cardiac troponin (Fig. 3 k–m). The concentration of 183 REG3βpos neutrophils was much lower in the bone marrow and blood compared to the infarcted 184 heart but, if present, showed the same hyperactivated state, indicated by increased geometric 185 mean fluorescent intensities (gMFI) of CD54, CD63, NOX2 and DHR123 compared with 186 REG3βneg neutrophils (Suppl. Fig. 5a–d). 187 To position REG3βpos neutrophils in the continuum of neutrophil activation and maturation, we 188 subjected REG3βneg and REG3βpos neutrophils sorted from infarcted hearts to scRNA-seq19, 189 20. Analysis of 5180 REG3βpos and 10814 R EG3βneg neutrophils with a median of 1248 190 expressed genes per cell identified five different subsets of neutrophils in the infarcted mouse 191 heart (Fig. 3n; Suppl. Fig. 6 a). Next, we calculated bone marrow proximity (BMP) and aging 192 scores of each cluster according to studies analyzing neutrophil development and aging 21. 193 Cluster 2 showed the highest BMP score with gene expression characteristic for young 194 neutrophils (Sell, Retnlg, Lcn2), whereas clusters 1, 3, and 5 h ad higher aging scores and 195 expressed genes typical for aged neutrophils (Cxcr4, Siglecf, Ncf1) ( Supp. Fig. 6 b–e). 196 Pseudotime analysis, defining cluster 2 as root, placed clusters 1, 3, and 5 at the end of the 197 neutrophil maturation trajectory ( Suppl. Fig. 6 f). We also identified a unique differentiated 198 subset of neutrophils, cluster 4 , characterized by increased expression of interferon (IFN) 199 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted February 20, 2026. ; https://doi.org/10.64898/2026.02.19.706868doi: bioRxiv preprint 8 stimulated genes (Irf7, Ifit1, Ifi204), whose presence in the heart was independent of MI and 200 did not follow the neutrophil differentiation trajectory (Suppl. Fig. 6g, h)19,22. 201 Most REG3βpos neutrophils belonged to the clusters of aged neutrophils, which was confirmed 202 by flow cytometric analysis, showing increased presence of neutrophil aging markers such as 203 CXR4, CD49D, CD24, and SiglecF in as R EG3βpos neutrophils and decreased presence of 204 CD62L, a marker for young neutrophils (Fig.3 o –t)19. A particular enrichment of REG3βpos 205 neutrophils was seen in aged cluster 5 (9% of REG3βpos vs. 2% of REG3βneg), which has the 206 highest scores for metallopeptidase activity, ROS production, and phagocytosis among all 207 neutrophil subsets (Fig. 3n, o; Suppl. Fig. 6i–p). To analyze whether activation and aging is a 208 prerequisite for binding and subsequent cytotoxicity of REG3β or a mere epiphenomenon, we 209 cultured isolated neutrophils from the bone marrow and exposed them to lipopolysaccharide 210 (LPS) to induce activation and aging (Suppl. Fig. 7 a–d). As expected, a ctivation of bone 211 marrow-derived neutrophils increased REG3β binding and REG3β-dependent cytotoxicity 212 (Suppl. Fig. 7e–i). Taken together, the data demonstrate that REG3β preferentially binds and 213 kills hyperactivated, aged neutrophils. 214 Translocation of granule-derived paucimannose-conjugated proteins to the cell 215 surface enables binding of REG3β 216 Since no conclusive evidence of a receptor for REG3β exists but specific interactions of REG 217 proteins with carbohydrate epitopes of peptidoglycans on the cellular surface of bacterial cells 218 have been reported15,23,24, we reasoned that R EG3β may exerts its action via binding to cell 219 surface N-glycans. Cell surface N-glycan analysis of sorted REG3βneg and REG3βpos peritoneal 220 neutrophils identified 28 N-glycans, which were grouped into 8 distinct glycan traits (Suppl. Fig. 221 8a–c; Suppl. Tab le 2). High mannose structures dominated the cell surface N -glycome of 222 REG3βneg and REG3βpos neutrophils, followed by presence of fucose, galactose, and complex 223 N-glycan traits on REG3βneg neutrophils (Suppl. Fig. 8c). In contrast, 18% of the cell surface 224 N-glycome of REG3βpos but only 2% of REG 3βneg neutrophils consisted of atypical small 225 paucimannose type N -glycans, which is an under-studied class of N -glycosylation in 226 mammalian cells (Fig. 4a; Suppl. Fig. 8 c)25,26,27. Immunofluorescence staining and flow 227 cytometric analysis using the paucimannose-reactive Mannitou antibody confirmed enrichment 228 for paucimannosylation of REG3βpos neutrophils28 (Fig. 4b, c). 229 Furthermore, activation of bone-marrow derived neutrophils with LPS elevated paucimannose 230 levels on the cell surface, suggesting that the ability of REG3β to bind to activated neutrophils 231 is mediated by paucimannosylation (Fig. 4d; Suppl. Fig. 7e). To test this hypothesis, we treated 232 activated neutrophils with the enzyme Peptide:N-glycosidase F (PNGase F), which cleaves all 233 N-linked glycans from glycoproteins . PNGase F treatment essentially abolished binding of 234 REG3β to activated neutrophils ( Fig. 4e) and also neutralized cytotoxic effects (Fig. 4e, f). 235 Importantly, treatment of activated neutrophils with endoglycosidase D (EndoD), specifically 236 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted February 20, 2026. ; https://doi.org/10.64898/2026.02.19.706868doi: bioRxiv preprint 9 cleaving Paucimannose type N-glycans, caused similar effects (Fig. 4e, f), whereas treatment 237 with Endo H or EndoF2, which cleave high mannose hybrid and complex saccharide s, 238 respectively, did neither affect REG3β binding nor cytotoxicity ( Fig. 4e, f). To further 239 corroborate these results, we performed competition assays with different monosaccharides 240 in millimolar concentrations. We found that mannose, a key component of paucimannosidic N-241 glycans, but not galactose inhibited binding of REG3β and prevented cell death (Fig. 4f, g). 242 To identify paucimannosylated glycoproteins interacting with REG 3β, we conducted a liquid 243 chromatography tandem mass s pectrometry (LC-MS/MS)-based glycoproteomic analysis of 244 peritoneal neutrophils. We found that the majority of glycoproteins were primarily decorated 245 with high mannose, followed by complex, paucimannose and hybrid glycans (Fig. 4h). In total, 246 21 paucimannosylated proteins were identified . Mannose2Fucose1N-Acetylglucosamine2 247 (M2F) was the most abundant paucimannose signature of REG 3βpos neutrophils, a 248 characteristic N-glycan of azurophile granule derived glycoproteins (Fig. 6a; Suppl. Fig. 9b)26. 249 The vast majority of paucimannosylated proteins (approx. 94% ) consisted of neutrophilic 250 granule protein (NGP), myeloperoxidase (MPO), neutrophil elastase (ELANE), integrin alpha 251 M (ITGAM) , and CD177 (CD177) ( Fig. 4 h; Suppl. Fig. 9 a, b ). Analysis of the putative 252 subcellular localization by neXtprot indicated that NGP, MPO, and ELANE reside in azurophilic 253 granules ( Fig. 4i; Suppl. Fig. 9c)25,27. Based on these results we reasoned that 254 paucimannosylated granular proteins of REG3βpos neutrophils are translocated to the cell 255 membrane following activation and subsequent degranulation of neutrophils. 256 Flow cytometric cell surface profiling for paucimannosylated proteins confirmed our 257 hypothesis, revealing elevated levels of NGP, MPO, and ELANE on the cell membrane of 258 cardiac tissue REG3βpos neutrophils (Suppl. Fig. 10 a–c), whereas ITGAM and CD177 were 259 either not altered or declined (Suppl. Fig. 10d, e). Furthermore, we detected a rapid increase 260 of paucimannose, NGP, MPO, and ELANE upon activation of neutrophils , which further 261 increased with aging and peaked in hyperactivated, aged neutrophils (Fig. 4k; Suppl. Fig. 10f–262 h). Concomitant increase of extracellular levels of the degranulation marker CD63 support the 263 idea that increased degranulation in activated and aging neutrophils maximizes translocation 264 of paucimannosylated proteins (Suppl. Fig. 10i ). Moreover, we found that inhibition of 265 neutrophil degranulation by Nexinhib20 diminished binding of REG3β to activated neutrophils 266 and abrogated cytotoxic effects ( Suppl. Fig. 10 j, k )29. To interrogate membrane-specific 267 interactions of REG3β with azurophile granule derived paucimannosylated proteins, we used 268 confocal microscopy, which revealed the presence of NGP, MPO, and ELANE on wheat germ 269 agglutinin (WGA) + plasma membranes of REG3βpos neutrophils, partially colocaliz ing with 270 REG3β (Fig. 4l). C ell surface c o-immunoprecipitation experiments using REG3β-treated 271 neutrophils uncovered interactions between REG 3β and NGP, MPO, and ELANE at the 272 plasma membrane (Fig. 4m; Suppl. Fig. 11 a–d, f–k). No interactions were detected between 273 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted February 20, 2026. ; https://doi.org/10.64898/2026.02.19.706868doi: bioRxiv preprint 10 REG3β and ITGAM and CD177, illustrating specificity of these interactions (Suppl. Fig. 11d, 274 e). Taken together, our results demonstrate that paucimannosylated proteins are redistributed 275 by degranulation to the cell surface of neutrophils in an activation- and age-dependent manner, 276 enabling binding of REG3β. 277 278 REG3β induces lysosome-mediated cell death of activated neutrophils 279 Induction of apoptosis by tumor necrosis factor alpha (TNFα) is an established mechanism to 280 accomplish programmed cell death of neutrophils , although data for cardiac neutrophils are 281 scarce30. To investigate whether R EG3β employs a similar machinery to kill neutrophils , we 282 directly compared effects of REG3β and TNFα on neutrophils. We observed similar kinetics of 283 LDH release kinetics and loss of Mitospy, a fluorescent reagent labelling mitochondria in viable 284 cells after treatment with either REG3β or TNFα (Fig. 5a, b). The ratio of 7AAD-/AnnV+ 285 neutrophils and formation of forward-side-scatter (FSC)low/AnnV+ small vesicles increased in 286 REG3β- or TNFα-treated neutrophil cultures, though the increase of 7AAD-/AnnV+ neutrophils 287 by R EG3β was less prominent and only became significant after 60 minutes ( Fig. 5c, d). 288 Importantly, we did not detect increased activation of the key executioner caspases 3 and 7 289 upon administration of REG3β to neutrophils at any time point, in stark contrast to the effects 290 of TNFα (Fig. 5e). We concluded that R EG3β does not induce classical apoptotic cell death 291 via activation of caspases 3 and 7. 292 To obtain further insights into the mechanism by which R EG3β eliminates neutrophils, we 293 analyzed REG3β-treated cultured peritoneal neutrophils by transmission (TEM) and scanning 294 electron microscopy (SEM) . We observed several morphological alterations including 295 cytoplasmatic vacuolization, loss of granular structures and membrane ruffling, which became 296 apparent already after 15 minutes and increased within 60 minutes after exposure to REG3β 297 (Fig. 5f). Characteristic features of caspases 3/7 -dependent apoptosis such as chromatin 298 compaction in crescent shaped masses at the nuclear periphery or membrane blabbing were 299 not detected. Due to the presence of REG3β in the cytoplasm of neutrophils, we speculated 300 that REG3β is taken up by activated neutrophils after binding to paucimannosylated membrane 301 proteins and induces cell death by an intracellular mechanism . Proteins are frequently taken 302 up via the endocytotic pathway, involving vesicle formation at the plasma membrane followed 303 by fusion with endosomes, which mature into lysosomes or fuse with preexisting 304 lysosomes31,32. Immunostaining combined with expansion microscopy detected REG3β in 305 early endosome antigen 1 (EEA1)+ endosomes within minutes upon administration as well as 306 in lysosomal-associated membrane protein 1 (LAMP1)+ lysosomes (Fig. 5g). A major endocytic 307 pathway in mammalian cells is c lathrin-mediated endocytosis (CME)33. To investigate its 308 involvement in the uptake of R EG3β, we treated neutrophils with chlorpromazine, a CME 309 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted February 20, 2026. ; https://doi.org/10.64898/2026.02.19.706868doi: bioRxiv preprint 11 blocker34. Treatment with chlorpromazine abrogated cytotoxic effects of REG3β on neutrophils 310 (Fig. 5h), indicating a critical role of CME for REG3β-dependent cell death. 311 To analyze putative direct effects of R EG3β on lysosomes, we stained REG3βneg and 312 REG3βpos cardiac neutrophils with Lysotracker, a cell -permeable fluorescent dye labelling 313 acidic lysosomes. Interestingly, R EG3β-binding neutrophils showed depletion of lysosomes. 314 Approximately 49% of REG3βpos neutrophils were negative for Lysotracker (Lysoneg), but only 315 21% of REG3βneg neutrophils, indicating depletion of lysosomes in REG3βpos neutrophils (Fig. 316 5i). Next, we treated Lysotracker-stained peritoneal neutrophils with REG3β, which resulted in 317 a rapid increase of Lysoneg neutrophils. The effects of REG3β were similar to L-leucyl-L-leucine 318 methyl ester (LLOme), a dipeptide, which polymerize s inside lysosomes and induce s 319 lysosomal membrane damage ( Fig. 5j)35. To corroborate REG3β-dependent damage of 320 lysosomal membranes, we employed the galectin puncta assay, which is based on 321 translocation of cytosolic galectins to the endolysosomal glycocalyx after lysosomal membrane 322 permeabilization36. We observed a punctuated localization of Galectin-1 (GAL-1), overlapping 323 with LAMP1+ lysosomes, whereas PBS -treated neutrophils showed diffuse cytosolic GAL -1 324 expression, characteristic for intact lysosomes (Fig. 5k). 325 Lysosomal membrane permeabilization leads to leakage of lysosomal hydrolases, especially 326 cathepsins, which induce cell death in various cell types 37,38. RNA-seq analysis identified 15 327 members of the large family of cathepsins in cardiac neutrophils after MI, with Cathepsin B 328 (CATB)- and Cathepsin D (CATD) showing the highest expression (Fig. 5l). Expansion 329 microscopy localized CATB and CAT D to lysosomal LAMP1 + compartments under basal 330 conditions (Fig. 5m, n). In contrast, administration of REG3β caused a much more diffuse 331 positioning of CATB and CAT D in neutrophils , indicating release from lysosomes into the 332 cytosol (Fig. 5m, n ). To confirm the role of cathepsin -release from lysosomes for REG3β-333 mediated cell death , we pretreated neutrophils with pan-cathepsin inhibitor E -64d, CAT B 334 inhibitor Pepstatin A , and CAT D inhibitor CA -074. All three inhibitors prevented R EG3β-335 induced cell death of activated neutrophils as indicated by reduced release of LDH (Fig. 5o). 336 Taken together our data indicate that REG3β induces death of activated neutrophils by 337 lysosomal membrane permeabilization and release of cathepsins after endocytotic uptake and 338 transport to lysosomes. 339 REG3β is required for clearance of neutrophils by efferocytosis 340 Removal of dying neutrophils from sites of injury is eventually achieved by efferocytosis, 341 employing professional phagocytes, including macrophages, and to a lesser extent by 342 monocytes, dendritic cells and neutrophils39,40. In addition to increased localization of AnnV on 343 the cell surface of R EG3βpos neutrophils, reflecting externalization of the ‘eat-me’ signal 344 phosphatidylserine, we detected increased expression of genes coding for efferocytotic 345 receptors and “bridging” ligands in REG3βpos neutrophils localized in the scRNA-seq cluster 5 346 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted February 20, 2026. ; https://doi.org/10.64898/2026.02.19.706868doi: bioRxiv preprint 12 (Fig. 2h; Fig. 6a, b). The cluster also contained macrophage-specific genes such as Ccr2, 347 Mrc1, Cd209a suggesting formation of macrophage-Reg3βpos neutrophil-hybrids. 348 Morphological analysis of s orted cardiac R EG3βneg versus REG3βpos neutrophils confirmed 349 this assumption (Fig. 6c, d). Flow cytometry analysis revealed that approx. 60% of REG3βpos 350 neutrophils formed hybrids with CD64hi/MERTKhi macrophages, whereas such hybrids were 351 essentially absent when analyzing REG3βneg neutrophils (Fig. 6e). 352 To functionally validate the role of REG3β for macrophage-mediated neutrophil clearance, we 353 employed Ly6G TdTomato mice, termed Catchup, which serve as reporter for neutrophils41. 354 Analysis of Ly6G TdTomato and Ly6G TdTomato//Reg3b-/-mice demonstrated a decreased ratio of 355 TdTomato+ (CD64hi/MERTKhi) macrophages in infarct regions of Ly6G TdTomato//Reg3b-/- 356 compared to Ly6GTdTomato mice (Fig. 6f). Accordingly, ratios of non-phagocytosed TdTomato+ 357 neutrophils were increased in Ly6GTdTomato//Reg3b-/-mice (Fig. 6g), indicat ing stimulation of 358 efferocytosis by R EG3β. Since efferocytosis promotes transition of macrophages to a more 359 reparative phenotype , we analyzed the phenotype of macrophages based on differential 360 surface expression of MHC -II and Ly6C in infarcted hearts of Catchup versus Catchup x 361 Reg3b-/- mice42,43,17. We observed a substantial increase of proinflammatory MHC-IIhi /LY6Clo 362 and a decline of reparative MHC-IIlo /LY6Clo macrophages in Ly6GTdTomato//Reg3b-/- compared 363 to Ly6GTdTomato mice. In contrast, the ratios of proangiogenic LY6C hi macrophages were not 364 affected ( Fig. 6h). In conclusion, o ur findings demonstrate a crucial role of REG3β in 365 orchestrating removal of neutrophil from infarcted hearts , thereby contributing to the 366 termination of inflammation after MI and enabling proper cardiac remodeling. 367 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted February 20, 2026. ; https://doi.org/10.64898/2026.02.19.706868doi: bioRxiv preprint 13

368 Containment of inflammation after MI is critical for myocardial healing and proper remodeling, 369 allowing damaged hearts to function despite the loss of contractile tissue. Neutrophils are 370 important drivers of inflammation, arriving first at the scene after tissue damage, making they 371 timely removal important to limit inflammation44. In addition, activated neutrophils exert direct 372 adverse effects on the myocardium, potentially causing electrical storms, further emphasizing 373 the need for their s wift clearance44,45. We discovered that REG3β specifically removes aged 374 and hyperactivated neutrophils, establishing a negative feedback loop within the heart to 375 restrain tissue damaging-activities of neutrophils and terminate the first phase of inflammation. 376 We postulate that this pathway establishes an active, tissue-instructed program of neutrophil 377 elimination. So far, t he function of cardiomyocytes in removal of neutrophils and limiting 378 inflammation has not been recognized before. Attention was mostly focused on cardiac stromal 379 cells and infiltrating macrophages, which also play important roles in restricting inflammation 380 via different mechanisms46,47. 381 REG3β enriches at the interface of infarcted and non -infarcted regions of ischemically 382 damaged hearts, where it presumably creates a barrier against further neutrophil expansion, 383 protecting the viable myocardium. The scenario in infarcted hearts resembles the situation in 384 the intestine, but for very different reasons. In the intestine, RegIIIγ acts as an antibacterial 385 lectin to establish a zone that separates microbiota from the epithelial surface48. In both cases, 386 REG proteins recognize the targets (hyperactivated neutrophils or microbiota) through the 387 presence of mannose-containing glycoconjugates, arguing for an evolutionary conserved type 388 of innate immune response48,24. Paucimannosylation has been broadly studied in lower 389 organisms such as insects and nematodes but more recently also detected in vertebrates as 390 an unconventional type of protein N -glycosylation49. Interestingly, human neutrophils contain 391 large amounts of paucimannosidic glycans that are enriched in azurophilic granules but the 392 role of paucimannosylation in disease processes has remained understudied50. 393 Specificity of REG3β for hyperactivated and aged neutrophils is achieved by enhanced 394 translocation of paucimannosylated proteins from granules to the surface. Progressive loss of 395 granule content is a hallmark of neutrophil activation and aging , although the functional 396 relevance of the translocation is not fully understood 26,51. We assume that the exposition of 397 azurophil granules-derived paucimannosylated proteases, including MPO, and ELANE, on the 398 surface of aged neutrophils serves a dual purpose: (i) facilitate migration of neutrophils through 399 inflamed cardiac tissue and digestion of disposable extracellular material 52; and (ii) label 400 hyperactive neutrophils for removal by REG3β. After binding to p aucimannosylated proteins 401 REG3β enters neutrophils by clathrin-mediated endocytosis before accumulation within 402 lysosomes and induction of lysosomal membrane permeabilization. We currently do not 403 understand the exact molecular mechanism by which REG3β permeabilizes lysosomal 404 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted February 20, 2026. ; https://doi.org/10.64898/2026.02.19.706868doi: bioRxiv preprint 14 membranes but it is tempting to speculate that a similar mode of action is employed as in 405 bacteria. In bacteria, RegIIIα binds peptidoglycan carbohydrate s on membranes before 406 forming a hexameric membrane -permeabilizing oligomeric pore that kills the targe t53. Why 407 REG3β does not form a pore within the cell membrane of neutrophils but presumably in the 408 lysosomal membrane is unclear and requires further investigations. 409 REG3β initiates lysosome-mediated cell death for removal of activated and aged neutrophils, 410 which differs from other ligand-dependent modes of killing, including pyroptosis, necroptosis, 411 ferroptosis, NETOsis, and apoptosis54. We found that permeabiliz ation of lysosomal 412 membranes by REG3β releases cathepsins. Lysosomal membrane permeabilization and 413 discharge of cathepsin are found in most types of cell death , including apoptosis, often 414 amplifying ongoing cell death routines55. However, release of cathepsins does not necessarily 415 lead to apoptosis but depends on the cellular context. W e did not detect cleavage of the 416 effector caspases-3/7, making it unlikely that activation of caspases is the principal executioner 417 of cell death in REG3β-treated neutrophils. On the other hand, we observed higher ratios of 418 the apoptotic marker AnnexinV on REG3βpos neutrophils and enhanced display of 419 phosphatidylserine, allowing efficient efferocytosis. The condition is reminiscent of a process 420 coined lysoptosis, which occurs in the absence of caspase activation but requires cytoplasmic 421 release of lysosomal cysteine peptidases 55. Further studies are required to unravel the 422 interactions between lysosomal cell death and apoptosis, although a mechanistic dissection is 423 difficult given the rapid dynamics, complexity and crosstalk between individual cell death 424 pathways56. 425 The discovery of REG3β-induced cell death of hyperactivated neutrophils paves the way for 426 potential therapeutic interventions, allow ing manipulation of neutrophil-driven inflammation, 427 e.g. preventing neutrophil -dependent damage of cardiomyocyte membranes that cause 428 arrhythmias and cell death45. Timing will be critical for such manipulations since hyperactivated 429 neutrophils are not necessarily detrimental, acting in a highly stage -dependent manner. 430 Exploitation of changes in glycoconjugates such as p aucimannosylation may provide further 431 tools for manipulation of specific subsets of neutrophils. 432 433 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted February 20, 2026. ; https://doi.org/10.64898/2026.02.19.706868doi: bioRxiv preprint 15

434 We thank Dr. Ul rich Gärtner and Anika Seipp f rom Justus Liebig University Giessen for 435 transmission and scanning electron microscopy. We thank Kikhi Khrievono for help with flow 436 cytometry. The help of Kenny Mattonet in imaging experiments is greatly acknowledged. 437 438 SOURCES OF FUNDING 439 This work was supported by the collaborative research center SFB 1531 (TP B08) and SFB 440 1213 (TP A02 and B02), the Transregional Collaborative Research Centre 267 (TP A05), and 441 332 (TP C06), the Excellence Cluster Cardiopulmonary Institute (CPI) and the German Center 442 for Cardiovascular Research DZHK StartUP grant 81X3200301. 443 444 DISCLOSURES 445 None. 446 447 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted February 20, 2026. ; https://doi.org/10.64898/2026.02.19.706868doi: bioRxiv preprint 16

448 1 Frangogiannis, N. G. The inflammatory response in myocardial injury, repair, and 449 remodelling. Nat Rev Cardiol 11, 255-265 (2014). 450 https://doi.org:10.1038/nrcardio.2014.28 451 2 Sreejit, G. et al. Neutrophil-Derived S100A8/A9 Amplify Granulopoiesis After Myocardial 452 Infarction. Circulation 141, 1080-1094 (2020). 453 https://doi.org:10.1161/CIRCULATIONAHA.119.043833 454 3 Bouvain, P. et al. Non-invasive mapping of systemic neutrophil dynamics upon 455 cardiovascular injury. Nat Cardiovasc Res 2, 126-143 (2023). 456 https://doi.org:10.1038/s44161-022-00210-w 457 4 Yang, L. et al. Neutrophils in Tissue Injury and Repair: Molecular Mechanisms and 458 Therapeutic Targets. MedComm (2020) 6, e70184 (2025). 459 https://doi.org:10.1002/mco2.70184 460 5 Bratton, D. L. & Henson, P. M. Neutrophil clearance: when the party is over, clean-up 461 begins. Trends Immunol 32, 350-357 (2011). https://doi.org:10.1016/j.it.2011.04.009 462 6 Frodermann, V. & Nahrendorf, M. Neutrophil-macrophage cross-talk in acute myocardial 463 infarction. Eur Heart J 38, 198-200 (2017). https://doi.org:10.1093/eurheartj/ehw085 464 7 Hsu, A. Y. et al. Neutrophil-derived vesicles control complement activation to facilitate 465 inflammation resolution. Cell 188, 1623-1641 e1626 (2025). 466 https://doi.org:10.1016/j.cell.2025.01.021 467 8 Lorchner, H. et al. Myocardial healing requires Reg3beta-dependent accumulation of 468 macrophages in the ischemic heart. Nat Med 21, 353-362 (2015). 469 https://doi.org:10.1038/nm.3816 470 9 Arruda-Olson, A. M., Reeder, G. S., Bell, M. R., Weston, S. A. & Roger, V. L. Neutrophilia 471 predicts death and heart failure after myocardial infarction: a community-based study. 472 Circ Cardiovasc Qual Outcomes 2, 656-662 (2009). 473 https://doi.org:10.1161/CIRCOUTCOMES.108.831024 474 10 Morrissey, S. M. et al. Multifaceted roles of neutrophils in cardiac disease. J Leukoc Biol 475 117 (2025). https://doi.org:10.1093/jleuko/qiaf017 476 11 Margraf, A., Lowell, C. A. & Zarbock, A. Neutrophils in acute inflammation: current 477 concepts and translational implications. Blood 139, 2130-2144 (2022). 478 https://doi.org:10.1182/blood.2021012295 479 12 Silvestre-Roig, C., Braster, Q., Ortega-Gomez, A. & Soehnlein, O. Neutrophils as 480 regulators of cardiovascular inflammation. Nat Rev Cardiol 17, 327-340 (2020). 481 https://doi.org:10.1038/s41569-019-0326-7 482 13 Soehnlein, O., Steffens, S., Hidalgo, A. & Weber, C. Neutrophils as protagonists and 483 targets in chronic inflammation. Nat Rev Immunol 17, 248-261 (2017). 484 https://doi.org:10.1038/nri.2017.10 485 14 Perez-Figueroa, E., Alvarez-Carrasco, P., Ortega, E. & Maldonado-Bernal, C. 486 Neutrophils: Many Ways to Die. Front Immunol 12, 631821 (2021). 487 https://doi.org:10.3389/fimmu.2021.631821 488 15 Chen, Z., Downing, S. & Tzanakakis, E. S. Four Decades After the Discovery of 489 Regenerating Islet-Derived (Reg) Proteins: Current Understanding and Challenges. Front 490 Cell Dev Biol 7, 235 (2019). https://doi.org:10.3389/fcell.2019.00235 491 16 Lavine, K. The heart is put at risk of rupture by cells bordering damaged tissue. Nature 492 633, 45-46 (2024). https://doi.org:10.1038/d41586-024-02436-z 493 17 Lorchner, H. et al. Reg proteins direct accumulation of functionally distinct macrophage 494 subsets after myocardial infarction. Cardiovasc Res 114, 1667-1679 (2018). 495 https://doi.org:10.1093/cvr/cvy126 496 18 Mayadas, T. N., Cullere, X. & Lowell, C. A. The multifaceted functions of neutrophils. 497 Annu Rev Pathol 9, 181-218 (2014). https://doi.org:10.1146/annurev-pathol-020712-498 164023 499 19 Vafadarnejad, E. et al. Dynamics of Cardiac Neutrophil Diversity in Murine Myocardial 500 Infarction. Circ Res 127, e232-e249 (2020). 501 https://doi.org:10.1161/CIRCRESAHA.120.317200 502 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted February 20, 2026. ; https://doi.org/10.64898/2026.02.19.706868doi: bioRxiv preprint 17 20 Calcagno, D. M. et al. SiglecF(HI) Marks Late-Stage Neutrophils of the Infarcted Heart: A 503 Single-Cell Transcriptomic Analysis of Neutrophil Diversification. J Am Heart Assoc 10, 504 e019019 (2021). https://doi.org:10.1161/JAHA.120.019019 505 21 Evrard, M. et al. Developmental Analysis of Bone Marrow Neutrophils Reveals 506 Populations Specialized in Expansion, Trafficking, and Effector Functions. Immunity 48, 507 364-379 e368 (2018). https://doi.org:10.1016/j.immuni.2018.02.002 508 22 Calcagno, D. M. et al. The myeloid type I interferon response to myocardial infarction 509 begins in bone marrow and is regulated by Nrf2-activated macrophages. Sci Immunol 5 510 (2020). https://doi.org:10.1126/sciimmunol.aaz1974 511 23 Brown, G. D., Willment, J. A. & Whitehead, L. C-type lectins in immunity and 512 homeostasis. Nat Rev Immunol 18, 374-389 (2018). https://doi.org:10.1038/s41577-018-513 0004-8 514 24 Lehotzky, R. E. et al. Molecular basis for peptidoglycan recognition by a bactericidal 515 lectin. Proc Natl Acad Sci U S A 107, 7722-7727 (2010). 516 https://doi.org:10.1073/pnas.0909449107 517 25 Thaysen-Andersen, M. et al. Human neutrophils secrete bioactive paucimannosidic 518 proteins from azurophilic granules into pathogen-infected sputum. J Biol Chem 290, 8789-519 8802 (2015). https://doi.org:10.1074/jbc.M114.631622 520 26 Loke, I., Ostergaard, O., Heegaard, N. H. H., Packer, N. H. & Thaysen-Andersen, M. 521 Paucimannose-Rich N-glycosylation of Spatiotemporally Regulated Human Neutrophil 522 Elastase Modulates Its Immune Functions. Mol Cell Proteomics 16, 1507-1527 (2017). 523 https://doi.org:10.1074/mcp.M116.066746 524 27 Reiding, K. R., Lin, Y. H., van Alphen, F. P. J., Meijer, A. B. & Heck, A. J. R. Neutrophil 525 azurophilic granule glycoproteins are distinctively decorated by atypical pauci- and 526 phosphomannose glycans. Commun Biol 4, 1012 (2021). https://doi.org:10.1038/s42003-527 021-02555-7 528 28 Zipser, B., Diestel, S., Bello-DeOcampo, D., Tai, M. H. & Schmitz, B. Mannitou 529 monoclonal antibody uniquely recognizes paucimannose, a marker for human cancer, 530 stemness and inflammation. J Biotechnol 161, 5-5 (2012). 531 https://doi.org:10.1016/j.jbiotec.2012.07.160 532 29 Johnson, J. L. et al. Identification of Neutrophil Exocytosis Inhibitors (Nexinhibs), Small 533 Molecule Inhibitors of Neutrophil Exocytosis and Inflammation: DRUGGABILITY OF THE 534 SMALL GTPase Rab27a. J Biol Chem 291, 25965-25982 (2016). 535 https://doi.org:10.1074/jbc.M116.741884 536 30 Geering, B. & Simon, H. U. Peculiarities of cell death mechanisms in neutrophils. Cell 537 Death Differ 18, 1457-1469 (2011). https://doi.org:10.1038/cdd.2011.75 538 31 Jeger, J. L. Endosomes, lysosomes, and the role of endosomal and lysosomal biogenesis 539 in cancer development. Mol Biol Rep 47, 9801-9810 (2020). 540 https://doi.org:10.1007/s11033-020-05993-4 541 32 Birgisdottir, A. B. & Johansen, T. Autophagy and endocytosis - interconnections and 542 interdependencies. J Cell Sci 133 (2020). https://doi.org:10.1242/jcs.228114 543 33 Mettlen, M., Chen, P. H., Srinivasan, S., Danuser, G. & Schmid, S. L. Regulation of 544 Clathrin-Mediated Endocytosis. Annu Rev Biochem 87, 871-896 (2018). 545 https://doi.org:10.1146/annurev-biochem-062917-012644 546 34 Bitsikas, V., Correa, I. R., Jr. & Nichols, B. J. Clathrin-independent pathways do not 547 contribute significantly to endocytic flux. Elife 3, e03970 (2014). 548 https://doi.org:10.7554/eLife.03970 549 35 Tan, J. X. & Finkel, T. A phosphoinositide signalling pathway mediates rapid lysosomal 550 repair. Nature 609, 815-821 (2022). https://doi.org:10.1038/s41586-022-05164-4 551 36 Aits, S. et al. Sensitive detection of lysosomal membrane permeabilization by lysosomal 552 galectin puncta assay. Autophagy 11, 1408-1424 (2015). 553 https://doi.org:10.1080/15548627.2015.1063871 554 37 Yuan, J. & Ofengeim, D. A guide to cell death pathways. Nat Rev Mol Cell Biol 25, 379-555 395 (2024). https://doi.org:10.1038/s41580-023-00689-6 556 38 Xie, Z. et al. Cathepsin B in programmed cell death machinery: mechanisms of execution 557 and regulatory pathways. Cell Death Dis 14, 255 (2023). https://doi.org:10.1038/s41419-558 023-05786-0 559 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted February 20, 2026. ; https://doi.org/10.64898/2026.02.19.706868doi: bioRxiv preprint 18 39 Boada-Romero, E., Martinez, J., Heckmann, B. L. & Green, D. R. The clearance of dead 560 cells by efferocytosis. Nat Rev Mol Cell Biol 21, 398-414 (2020). 561 https://doi.org:10.1038/s41580-020-0232-1 562 40 Ramos, C. & Oehler, R. Clearance of apoptotic cells by neutrophils in inflammation and 563 cancer. Cell Death Discov 10, 26 (2024). https://doi.org:10.1038/s41420-024-01809-7 564 41 Hasenberg, A. et al. Catchup: a mouse model for imaging-based tracking and modulation 565 of neutrophil granulocytes. Nat Methods 12, 445-452 (2015). 566 https://doi.org:10.1038/nmeth.3322 567 42 Watanabe, S., Alexander, M., Misharin, A. V. & Budinger, G. R. S. The role of 568 macrophages in the resolution of inflammation. J Clin Invest 129, 2619-2628 (2019). 569 https://doi.org:10.1172/JCI124615 570 43 Chen, R. et al. Macrophages in cardiovascular diseases: molecular mechanisms and 571 therapeutic targets. Signal Transduct Target Ther 9, 130 (2024). 572 https://doi.org:10.1038/s41392-024-01840-1 573 44 Filep, J. G. Targeting Neutrophils for Promoting the Resolution of Inflammation. Front 574 Immunol 13, 866747 (2022). https://doi.org:10.3389/fimmu.2022.866747 575 45 Kumowski, N. et al. Resistin-like molecule gamma attacks cardiomyocyte membranes 576 and promotes ventricular tachycardia. Science 389, 1043-1048 (2025). 577 https://doi.org:10.1126/science.adp7361 578 46 Cadosch, N., Gil-Cruz, C., Perez-Shibayama, C. & Ludewig, B. Cardiac Fibroblastic 579 Niches in Homeostasis and Inflammation. Circ Res 134, 1703-1717 (2024). 580 https://doi.org:10.1161/CIRCRESAHA.124.323892 581 47 Dick, S. A. et al. Self-renewing resident cardiac macrophages limit adverse remodeling 582 following myocardial infarction. Nat Immunol 20, 29-39 (2019). 583 https://doi.org:10.1038/s41590-018-0272-2 584 48 Vaishnava, S. et al. The antibacterial lectin RegIIIgamma promotes the spatial 585 segregation of microbiota and host in the intestine. Science 334, 255-258 (2011). 586 https://doi.org:10.1126/science.1209791 587 49 Loke, I., Kolarich, D., Packer, N. H. & Thaysen-Andersen, M. Emerging roles of protein 588 mannosylation in inflammation and infection. Mol Aspects Med 51, 31-55 (2016). 589 https://doi.org:10.1016/j.mam.2016.04.004 590 50 Venkatakrishnan, V. et al. Glycan analysis of human neutrophil granules implicates a 591 maturation-dependent glycosylation machinery. J Biol Chem 295, 12648-12660 (2020). 592 https://doi.org:10.1074/jbc.RA120.014011 593 51 Hua, Y. et al. Proteomic Portrait of Degranulation Program in Human Circulating 594 Neutrophils Upon Multi-Inflammatory and Infectious Activation. Mol Cell Proteomics 25, 595 101078 (2026). https://doi.org:10.1016/j.mcpro.2025.101078 596 52 Lerchenberger, M. et al. Matrix metalloproteinases modulate ameboid-like migration of 597 neutrophils through inflamed interstitial tissue. Blood 122, 770-780 (2013). 598 https://doi.org:10.1182/blood-2012-12-472944 599 53 Mukherjee, S. et al. Antibacterial membrane attack by a pore-forming intestinal C-type 600 lectin. Nature 505, 103-107 (2014). https://doi.org:10.1038/nature12729 601 54 Chen, T., Ren, Q. & Ma, F. New insights into constitutive neutrophil death. Cell Death 602 Discov 11, 6 (2025). https://doi.org:10.1038/s41420-025-02287-1 603 55 Luke, C. J. et al. Lysoptosis is an evolutionarily conserved cell death pathway moderated 604 by intracellular serpins. Commun Biol 5, 47 (2022). https://doi.org:10.1038/s42003-021-605 02953-x 606 56 Rosazza, T., Warner, J. & Sollberger, G. NET formation - mechanisms and how they 607 relate to other cell death pathways. FEBS J 288, 3334-3350 (2021). 608 https://doi.org:10.1111/febs.15589 609 57 Pullen, A. B., Kain, V., Serhan, C. N. & Halade, G. V. Molecular and Cellular Differences 610 in Cardiac Repair of Male and Female Mice. J Am Heart Assoc 9, e015672 (2020). 611 https://doi.org:10.1161/JAHA.119.015672 612 58 Lorchner, H. et al. Concomitant Activation of OSM and LIF Receptor by a Dual-Specific 613 hlOSM Variant Confers Cardioprotection after Myocardial Infarction in Mice. Int J Mol Sci 614 23 (2021). https://doi.org:10.3390/ijms23010353 615 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted February 20, 2026. ; https://doi.org/10.64898/2026.02.19.706868doi: bioRxiv preprint 19 59 Swamydas, M., Luo, Y., Dorf, M. E. & Lionakis, M. S. Isolation of Mouse Neutrophils. Curr 616 Protoc Immunol 110, 3 20 21-23 20 15 (2015). 617 https://doi.org:10.1002/0471142735.im0320s110 618 60 Bolger, A. M., Lohse, M. & Usadel, B. Trimmomatic: a flexible trimmer for Illumina 619 sequence data. Bioinformatics 30, 2114-2120 (2014). 620 https://doi.org:10.1093/bioinformatics/btu170 621 61 Dobin, A. et al. STAR: ultrafast universal RNA-seq aligner. Bioinformatics 29, 15-21 622 (2013). https://doi.org:10.1093/bioinformatics/bts635 623 62 Liao, Y., Smyth, G. K. & Shi, W. featureCounts: an efficient general purpose program for 624 assigning sequence reads to genomic features. Bioinformatics 30, 923-930 (2014). 625 https://doi.org:10.1093/bioinformatics/btt656 626 63 Love, M. I., Huber, W. & Anders, S. Moderated estimation of fold change and dispersion 627 for RNA-seq data with DESeq2. Genome Biol 15, 550 (2014). 628 https://doi.org:10.1186/s13059-014-0550-8 629 64 Le, S., Josse, J. & Husson, F. FactoMineR: An R Package for Multivariate Analysis. 630 Journal of Statistical Software 25, 1-18 (2008). https://doi.org:10.18637/jss.v025.i01 631 65 Xie, C. et al. KOBAS 2.0: a web server for annotation and identification of enriched 632 pathways and diseases. Nucleic Acids Res 39, W316-322 (2011). 633 https://doi.org:10.1093/nar/gkr483 634 66 Wolf, F. A., Angerer, P. & Theis, F. J. SCANPY: large-scale single-cell gene expression 635 data analysis. Genome Biol 19, 15 (2018). https://doi.org:10.1186/s13059-017-1382-0 636 67 McInnes, L., Healy, J. & Melville, J. UMAP: Uniform Manifold Approximation and 637 Projection for Dimension Reduction. arXiv:1802.03426 (2028). 638 https://doi.org:10.48550/arXiv.1802.03426 639 68 Traag, V. A., Waltman, L. & van Eck, N. J. From Louvain to Leiden: guaranteeing well-640 connected communities. Sci Rep 9, 5233 (2019). https://doi.org:10.1038/s41598-019-641 41695-z 642 69 Schultheis, H. et al. SC-Framework: A robust and FAIR semi-interactive environment for 643 single-cell resolution datasets. bioRxiv https://doi.org/10.1101/2025.11.11.687874 (2025). 644 https://doi.org:10.1101/2025.11.11.687874 645 70 Setty, M. et al. Characterization of cell fate probabilities in single-cell data with Palantir. 646 Nat Biotechnol 37, 451-460 (2019). https://doi.org:10.1038/s41587-019-0068-4 647 71 Aiyetan, P., Zhang, B., Zhang, Z. & Zhang, H. XGlycScan: An Open-source Software For 648 N-linked Glycosite Assignment, Quantification and Quality Assessment of Data from Mass 649 Spectrometry-based Glycoproteomic Analysis. MOJ Proteom Bioinform 1 (2014). 650 https://doi.org:10.15406/mojpb.2014.01.00004 651 72 Merz, S. F. et al. Contemporaneous 3D characterization of acute and chronic myocardial 652 I/R injury and response. Nat Commun 10, 2312 (2019). https://doi.org:10.1038/s41467-653 019-10338-2 654 73 Cardiff, R. D., Miller, C. H. & Munn, R. J. Manual hematoxylin and eosin staining of 655 mouse tissue sections. Cold Spring Harb Protoc 2014, 655-658 (2014). 656 https://doi.org:10.1101/pdb.prot073411 657 74 Bissen, D., Kracht, M. K., Foss, F. & Acker-Palmer, A. Expansion microscopy of mouse 658 brain organotypic slice cultures to study protein distribution. STAR Protoc 3, 101507 659 (2022). https://doi.org:10.1016/j.xpro.2022.101507 660 75 Zhou, E. et al. Simultaneous and Positively Correlated NET Formation and Autophagy in 661 Besnoitia besnoiti Tachyzoite-Exposed Bovine Polymorphonuclear Neutrophils. Front 662 Immunol 10, 1131 (2019). https://doi.org:10.3389/fimmu.2019.01131 663 664 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted February 20, 2026. ; https://doi.org/10.64898/2026.02.19.706868doi: bioRxiv preprint 20 FIGURES 665 666 Figure 1: Compromised removal of neutrophils in Reg3b-deficient mice after MI 667 increases mortality and impairs heart function . a, Flow cytometry quantification of 668 neutrophils in wild-type (WT) and Reg3b deficient (Reg3b-/-) hearts 1, 2, 3, and 4 days after 669 MI. Each data point represents an individual mouse. b, Flow cytometry quantification of blood 670 neutrophils in WT and Reg3b-/- mice 4 days after MI. n = 6 for both groups c, Representative 671 flow cytometry dot plots and quantification of viable and dead neutrophils in WT and Reg3b-/- 672 hearts 4 days after MI, identified by 7-Aminoactinomycin D (7AAD) and Annexin V (AnnV) 673 staining. n = 7 for both groups d, e, f, Immunofluorescent images of LY6G+ neutrophils 674 (magenta) in longitudinal sections and zoom-in view of infarcted zone (IZ) and non -infarcted 675 remote zone (RZ) of WT and Reg3b-/- hearts 4 days after MI. e, f, Immunofluorescent 676 quantification of Ly6G+ neutrophils in serial longitudinal sections of WT and Reg3b-/- hearts 1, 677 2, 3, and 4 days after MI. n = 7 mice per group and timepoint. Phalloidin staining was used to 678 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted February 20, 2026. ; https://doi.org/10.64898/2026.02.19.706868doi: bioRxiv preprint 21 distinguish infarcted from non-infarcted tissue. Scale bars, 1000μm and 100μm in magnified 679 sections. g, Schematic outline of treatment with control (IgG) or anti-Ly6G (αLy6G) antibodies 680 after MI. h, Kaplan-Meier survival curves of WT (n = 26), Reg3b-/- +IgG (n = 25), and Reg3b-/- 681 +αLy6G (n = 31) after MI. i, j, MRI-based analysis of end-systolic volume, and ejection fraction 682 of WT (n = 6), Reg3b-/- +IgG (n = 6), and Reg3b-/- +αLy6G (n = 9) mice 14 days after MI. Data 683 are mean ± s.e.m. Two-way ANOVA followed by Sidak's multiple comparison test (a), one-way 684 ANOVA followed by Sidak's multiple comparison test (c, i, j), two-way ANOVA followed by 685 Tukey's multiple comparison test (e, f), Kaplan-Meier survival analysis (h). All experiments 686 were conducted with male mice. 687 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted February 20, 2026. ; https://doi.org/10.64898/2026.02.19.706868doi: bioRxiv preprint 22 688 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted February 20, 2026. ; https://doi.org/10.64898/2026.02.19.706868doi: bioRxiv preprint 23 Figure 2 : REG3β directly binds neutrophils and initiates rapid cell death . a, 689 Representative 3D visualization of LY6G+ neutrophils (yellow) and REG3β protein localization 690 (magenta) by light sheet fluorescence microscopy within whole hearts and zoom-in view of WT 691 mice at day 2 after MI. Muscle autofluorescence is white. Left atrium (LA), right atrium (RA), 692 left ventricle (LV), and right ventricle (RV). Scale bars, 10 00μm and 5 00μm in magnified 693 sections. b, Co-localization analysis of LY6G and REG3β signal intensities and calculation of 694 correlation coefficient by linear regression . n = 3 . c, Immunofluorescent images of LY 6G+ 695 neutrophils (yellow) and REG3β (magenta) protein localization in hearts of WT mice 2 days 696 after MI. Alpha-actinin-2 (ACTN2): grey and 4′,6-Diamidin-2-phenylindol (DAPI): blue. Scale 697 bars, 10μm. d, Immunohistochemical 3,3'-Diaminobenzidine (DAB) visualization of CD66b+ 698 neutrophils and REG3A in myocardial biopsies of humans with MI. Arrows indicate REG3A + 699 neutrophils. Counter staining with hematoxylin. Scale bars, 100μm. e, Representative flow 700 cytometric dot plots of REG3β-negative ( REG3βneg, blue) and REG3β-positive ( REG3βpos, 701 green) neutrophil subsets from bone marrow, blood and heart of WT mice 2 days after MI. 702 Mean ± sem of REG3βpos neutrophils in % of neutrophils are shown in each dot plot. n = 10. f, 703 Flow cytometric quantification of REG3βpos neutrophils in % of myeloid leukocytes in bone 704 marrow, blood and heart of WT mice before (n = 7) and 2 days after MI (n = 9). g, Heatmap of 705 Reg3b gene expression in REG3βneg and REG3βpos neutrophils from cardiac tissue of WT mice 706 2 days after infarct. Gapdh was used as reference. n = 4. h, Flow cytometric quantification of 707 viable and dead REG3βneg and REG3βpos neutrophils by Annexin V and 7AAD staining in WT 708 mice 2 days after MI. n = 10. i, Pearson correlation analysis between REG3βpos and cell death. 709 n = 9. j, Cell death kinetics of peritoneal neutrophils from WT mice , treated with REG3β at 710 indicated concentrations, determined by incorporation of Cytotox. Treatment with PBS served 711 as control. n = 7 for all groups. k, Lactate dehydrogenase ( LDH) release of isolated bone 712 marrow, blood and peritoneal neutrophils from WT mice after treatment with REG3β (100ng 713 ul-1). Administration of PBS and interferon gamma (IFNγ, 100ng ml-1) was used as controls. n 714 = 8 for all groups. Data are mean ± s.e.m. Pearson correlation ( b, i), nonparametric 715 Kolmogorov-Smirnov test (d), one-way ANOVA followed by Sidak's multiple comparison test 716 (f, h), two-way ANOVA mixed-effects analysis followed by Sidak's multiple comparison test (j), 717 two-way ANOVA followed by Tukey's multiple comparison test (k). All experiments were 718 conducted with male mice. 719 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted February 20, 2026. ; https://doi.org/10.64898/2026.02.19.706868doi: bioRxiv preprint 24 720 Figure 3: REG3β binds to hyperactive, aged neutrophils in the infarcted heart. a, Principal 721 component analysis of REG3β-negative ( REG3βneg, blue) and REG3β-positive ( REG3βpos, 722 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted February 20, 2026. ; https://doi.org/10.64898/2026.02.19.706868doi: bioRxiv preprint 25 green) neutrophils obtained from WT hearts 2 days after MI. (n = 4 for both groups). b, Volcano 723 plot depi cting fold changes (FC) and False Discovery Rate (FDR) -adjusted p -values of 724 differentially expressed genes (DEG) in REG3βneg and REG3βpos cardiac tissue neutrophils. c, 725 KOBAS (KEGG Orthology-Based Annotation System) pat hway enrichment bubble plot of 726 DEGs of REG3βneg compared with REG3βpos heart neutrophils. Selected significantly enriched 727 pathways are shown (P<0.05). The dashed line marks the P value of 0.05. Outer gray circles 728 represent the total number of genes in each pathway. Centered colored circles represent the 729 number of DEG in each pathway. Diameter and color of circles indicate the number of genes 730 and significance of enrichment, respectively. d–h, Geometric mean fluorescence intensity 731 (gMFI) of CD54 (d, n = 11), TLR2 (e, n = 10), CD63 (f, n = 9), TREM2 (g, n = 9), and NOX2 732 (h, n = 11) on REG3βneg, Pan and REG3βpos neutrophils from WT hearts 2 days after MI. i–m, 733 gMFI of MMP-9 (i, n = 8), and ADAM9 (j, n = 10), DHR123 (k, n = 11), pHrodo (l, n = 10), and 734 cardiac troponin (m, n = 10) in REG 3βneg, pan, and REG 3βpos neutrophils from WT hearts 2 735 days after MI. n, Uniform manifold approximation and projection (UMAP) of neutrophil gene 736 expression obtained from WT hearts at day 2 after MI. o, Density plots of REG3βneg, pan and 737 REG3βpos neutrophils from WT hearts 2 days after MI. Cluster ratios for REG 3βneg, pan and 738 REG3βpos heart neutrophils are also shown. p–t, gMFI of CXCR4 (p, n = 10), CD62L (q, n = 739 10), CD49D (r, n = 10), CD24 (s, n = 9), and SiglecF (t, n = 10) on REG3βneg, pan and REG3βpos 740 neutrophils from WT hearts 2 days after MI. Data are mean ± s.e.m. One-way ANOVA followed 741 by Sidak's multiple comparison test (d-m, p-r, t), Kruskal-Wallis 1-way ANOVA followed by 742 Dunn's multiple comparison test (s). All experiments were conducted with male mice. 743 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted February 20, 2026. ; https://doi.org/10.64898/2026.02.19.706868doi: bioRxiv preprint 26 744 Figure 4: Binding of REG3β to neutrophil subsets is enabled by translocation of granule-745 derived paucimannose-conjugated proteins. a, Cell surface N -glycan analysis of sorted 746 REG3β-negative (REG3βneg) and REG3β-positive (REG3βpos) peritoneal neutrophils from WT 747 mice. Zoomed-in view chromatograms of Paucimannose-type N-glycans from REG3βneg and 748 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted February 20, 2026. ; https://doi.org/10.64898/2026.02.19.706868doi: bioRxiv preprint 27 REG3βpos neutrophils are shown. mAU, milliabsorbance units. b, Immunofluorescent images 749 of paucimannose (yellow) in REG3βneg and REG3βpos peritoneal neutrophils from WT mice . 750 Wheat germ agglutinin (WGA ): grey . Scale bars, 5 μm. c, Geometric mean fluorescence 751 intensity (gMFI) of paucimannose on REG3βneg and REG3βpos cardiac tissue neutrophils from 752 WT mice 2 days after MI. n = 10. d, gMFI of Paucimannose on basal versus activated bone 753 marrow neutrophils. n = 7. e, gMFI of REG3β on REG3βpos bone marrow derived neutrophils 754 pretreated with peptide:N-glycosidase F (PNGase F), endoglycosidases H (EndoH), F2 755 (EndoF2), and D (EndoD) (all used at 1U ml -1) for 30 minutes after stimulation with REG 3β 756 (100 ng ml-1) for 15 minutes. PBS served as control. n = 6. f, Lactate dehydrogenase (LDH) 757 release of activated bone marrow derived -neutrophils pretreated with PNGase F, EndoH, 758 EndoF2, and EndoD (all used at 1U ml-1) for 30 minutes following stimulation with REG3β (100 759 ng ml-1) for 30 minutes. Cleavage specificity of each glycosidase is indicated. PBS served as 760 control. n = 6. g, h, gMFI of REG3β on REG3βpos bone marrow-derived neutrophils (g, n = 6) 761 and LDH release of activated bone marrow derived -neutrophils ( h, n = 6) treated with 762 Mannose, Galactose (both 100 mM) and REG3β (100 ng ml-1) for 30 minutes. PBS served as 763 control. i, Relative abundance of major N -glycan groups in peritoneal neutrophils. Relative 764 abundance of the five most abundant Paucimannose -conjugated proteins including 765 neutrophilic granule protein (NGP), myeloperoxidase (MPO), neutrophil elastase (ELANE), 766 integrin alpha M (ITGAM) and CD177 antigen (CD177). j, Subcellular localization of 767 paucimannose-conjugated proteins in peritoneal neutrophils. k, R atio of gMFI of 768 paucimannose, NGP, MPO, and ELANE on basal, activated, aged, and activated/aged bone 769 marrow derived-neutrophils in relation to basal -state neutrophils. Ratios are shown as fold 770 change, whereas levels in basal cells is set to 1 and illustrated by a red line. n = 6. l, 771 Immunofluorescent images of NGP, MPO, and ELANE (all in yellow) in non-permeabilized 772 REG3βpos peritoneal neutrophils from WT mice. REG3β: magenta and WGA: grey. Scale bars, 773 5μm and 1 μm in magnified sections. m, Immunoblot analysis of input and cell surface co -774 immunoprecipitated samples from peritoneal neutrophils using antibodies against REG 3β, 775 NGP, MPO and ELANE. Isotype controls (IgG) were used as controls. Arrows indicate bands 776 co-precipitated with the primary antigen. Data are mean ± s.e.m. T wo-sided unpaired t tests 777 (c, d), one-way ANOVA followed by Sidak's multiple comparison test (d–g). All experiments 778 were conducted with male mice. 779 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted February 20, 2026. ; https://doi.org/10.64898/2026.02.19.706868doi: bioRxiv preprint 28 780 Figure 5 : REG3β-induced release of lysosomal cathepsins causes cell death. a –e, 781 Lactate dehydrogenase (LDH) release (a, n = 10 for all groups), flow cytometric quantification 782 of Mitospylow (b, n = 8 for all groups), quantification of 7AAD-/AnnV+ (c, n = 6 for all groups ), 783 quantification of FSC low/AnnV+ small vesicles ( d, n = 6 for all groups), and quantification of 784 cleaved caspase 3/7 activity (e, n = 7 for all groups) of REG3β-treated (100ng ul-1) peritoneal 785 neutrophils from WT mice at indicated time points. PBS and tumor necrosis factor alpha (TNFα, 786 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted February 20, 2026. ; https://doi.org/10.64898/2026.02.19.706868doi: bioRxiv preprint 29 100ng ml-1) were used as controls. f, Transmission (TEM, left panel) and scanning electron 787 microscopy (SEM, right panel) images of peritoneal neutrophils from WT mice treated with 788 REG3β (100ng ul-1) at indicated time points. PBS served as control. Scale bars, 2μm for TEM 789 and 2.5 μm for SEM. g, Immunofluorescent images of early endosome a ntigen 1 (EEA1) 790 combined with REG3β (magenta) and lysosomal-associated membrane protein 1 (LAMP1, 791 yellow) combined with REG3β in REG3β-treated (100ng ul-1) peritoneal neutrophils from WT 792 mice at indicated time points. Pearson coefficient of colocalization of EEA1 and LAMP1 with 793 REG3β. n = 18. Scale bars, 5 μm. h, LDH release of peritoneal neutrophils from WT mice 794 pretreated with chlorpromazine (1μM) for 30 minutes, following stimulation with REG 3β (100 795 ng ml -1) for 15 minutes . DMSO served as control. n = 7. i, F low cytometric dot plots and 796 quantification of REG3βneg (blue) and REG3βpos (green) neutrophils from WT mice 2 days after 797 MI, separated into Lyso pos and Lyso neg subsets. n = 7 . j, Flow cytometric quantification of 798 Lysoneg cells of peritoneal neutrophils from WT mice after treatment with Reg3β (100ng ul-1) for 799 15 minutes. Administration of PBS and recombinant L-leucyl-L-leucine methyl ester (LLOme, 800 100ng ml-1) were used as controls. n = 8 for all groups. k, Immunofluorescent images of LAMP1 801 (yellow) and Galectin-1 (GAL-1, magenta) localization in peritoneal neutrophils from WT mice 802 after treatment with Reg3β (100ng ul-1) at indicated time points. PBS served as control. Scale 803 bars, 10μm. l, Heatmap of mean cathepsin family gene expression counts in neutrophils from 804 WT hearts 2 days after MI. n = 4. m, Immunofluorescent images of LAMP1 (yellow) and 805 Cathepsin B (CATB, magenta), and n, LAMP1 (yellow) and Cathepsin D (CATD, magenta) 806 localization in peritoneal neutrophils from WT mice upon treatment with recombinant Reg3β 807 (100ng ul-1) for 15 minutes. Administration of PBS served as negative control. Scale bars, 5μm. 808 o, LDH release of peritoneal neutrophils from WT mice pretreated with Pan cathepsin inhibitor 809 E-64d, CATD inhibitor Pepstatin A, and CATB inhibitor CA -074Me (all used at 1μM) for 30 810 minutes following stimulation with recombinant REG 3β (100 ng ml -1) for 15 minutes . DMSO 811 served as control. n = 10. Data are mean ± s.e.m. Two-way ANOVA followed by Tukey's 812 multiple comparison test (b–f,), one-way ANOVA followed by Sidak's multiple comparison test 813 (h–j, n). All experiments were conducted with male mice. 814 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted February 20, 2026. ; https://doi.org/10.64898/2026.02.19.706868doi: bioRxiv preprint 30 815 Figure 6 : REG3β promotes neutrophil clearance via macrophage -mediated 816 efferocytosis. a, b, Bubble plot showing gene expression of efferocytotic receptors and 817 efferocytotic ligands across neutrophil clusters (see Fig.3). Bubble sizes represent percentage 818 of cells expressing corresponding genes and bubble colors represents expression levels. c, 819 Feature plot showing expression of Ccr2, Mrc1, and Cd209a across neutrophil clusters. d, 820 Cytospin morphology of sorted REG3βneg and REG3βpos cardiac tissue neutrophils from WT 821 mice 2 days after MI . Arrows indicate macrophages. Scale bars, 1 0μm e, Flow cytometry-822 based identification and quantification of macrophage -neutrophil-hybrid doublets obtained 823 from WT hearts 2 days after MI. CD64 hi/MERTKhi macrophages (grey), REG3βneg (blue) and 824 REG3βpos (green) neutrophils are shown. n = 5. f, Flow cytometric density plots and 825 quantification of TdTomato+ CD64hi/MERTKhi macrophages of Catchup (n = 8) and Catchup x 826 Reg3b-/- (n = 7) hearts 2 days after MI. g, Flow cytometric density plots and quantification of 827 TdTomato+ neutrophils of Ly6GTdTomato (Catchup, n = 8) and Catchup x Reg3b-/- (n = 7) hearts 828 4 days after MI. h, Representative flow cytometric density plots and quantification of MHC -829 IIhi/LY6Clo, MHC -IIlo/LY6Clo, and LY6C lo macrophage subset ratios in Catchup (n = 8) and 830 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted February 20, 2026. ; https://doi.org/10.64898/2026.02.19.706868doi: bioRxiv preprint 31 Catchup x Reg3b-/- (n = 7) hearts 4 days after MI. Data are mean ± s.e.m.: T wo-sided Mann 831 Whitney test (e), two-sided unpaired t tests (f, g), one-way ANOVA followed by Sidak's multiple 832 comparison test (h). All experiments were conducted with male mice. 833 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted February 20, 2026. ; https://doi.org/10.64898/2026.02.19.706868doi: bioRxiv preprint