signaling in vivo.

44 Introduction 27 28 inflammatory diseases of the gut, skin and lung, as well as ischemic-reperfusion injuries of 29 the kidney, heart and brain. However, precise identification of the cells and tissues that 30

undergo necroptotic cell death in vivo has proven challenging in the absence of robust 31 reagents and protocols for immunohistochemical detection. Here, we provide automated 32 immunohistochemistry protocols to detect core necroptosis regulators – Caspase-8, This toolbox and methods will facilitate precise localisation and evaluation of necroptotic RIPK1, RIPK3 and MLKL – in formalin-fixed mouse and human tissues. We observed surprising heterogeneity in protein expression within tissues, whereby short-lived immune barrier cells were replete with necroptotic effectors, whereas long-lived cells exhibited 36 reduced RIPK3 or MLKL expression. Local changes in the expression of necroptotic effectors were observed in response to insults such as sterile inflammation, dysbiosis or 38 immune challenge, consistent with necroptosis being dysregulated in disease contexts.

The necroptotic cell death pathway leads to cell lysis and expulsion of cellular contents 46 into the extracellular milieu, which in turn provokes an innate immune response.

Necroptosis is considered to be an altruistic cell death pathway whose principal role is to protect the host from pathogens (1-8). Despite this, it is the aberrant functions of necroptosis associated with inflammatory diseases that have spurred interest in its underlying mechanisms and therapeutic prospects (9, 10). Studies of mice lacking the 51 terminal effectors of the pathway – RIPK3 (Receptor-interacting protein kinase-3) or MLKL 52 (Mixed-lineage kinase domain-like) – have led to the concept that excess necroptosis

The core signaling axis of the necroptotic pathway has been well-defined and can be activated in response to a variety of inflammatory cues including ligation of Death, Toll59 like or Pathogen-Pattern receptors (22-28). Caspase-8 is a critical negatively regulator of necroptotic signaling (28), whereby its deletion or loss-of-function promotes oligomerisation of RIPK1 (Receptor-interactor protein kinase-1), TRIF (TIR domaincontaining adapter molecule 1) and/or ZBP1 (Z-DNA-binding protein 1) (27). This oligomeric structure, otherwise known as the necrosome, promotes activation of the downstream effectors RIPK3 and MLKL (29). RIPK3 recruits the MLKL pseudokinase to 65 the necrosome, where it phosphorylates MLKL to provoke a conformational change, 66 release from the necrosome, oligomerization and trafficking to the plasma membrane (3035). At the plasma membrane, accumulation of activated MLKL to a critical threshold level 68 is required for membrane permeabilization via a poorly understood mechanism that brings 53

drives a range of inflammatory pathologies in organs including the skin, gut, brain, heart, 54 lung, kidney and testes (11-18). However, many of these attributions have been disputed 55 (19-21), likely reflecting an evolving understanding of the pathway and the limited availability of validated reagents to interrogate necroptosis in pathological specimens. about the cell’s demise (33, 36, 37).

As our understanding of the necroptosis pathway has grown, new tools and protocols have been developed to study necroptotic signaling in fixed cultured cells (35, 38-40). However, robust procedures for assessing the necroptotic pathway in tissues are still lacking, often leading to contradictory reports in the literature and misattributions of necroptotic pathologies. Here, we report automated immunostaining protocols for detecting Caspase-8, RIPK1, RIPK3 and MLKL, in mouse formalin-fixed paraffin-embedded tissues. These procedures have allowed assembly of an atlas of necroptotic pathway expression in mouse tissues under basal conditions and during innate immune challenge. While the necroptosis machinery is rarely expressed in cell types other than short-lived barrier cells 79 in mouse tissues, sterile inflammation increased RIPK3 expression in the gut and liver, broadly predisposing multiple cell types to necroptotic death. In contrast, elimination of the 81 intestinal microflora diminished RIPK3 expression and reduced necroptotic susceptibility.

RIPK3 is also uniquely upregulated in splenic germinal centres and may influence antigenspecific immune responses in a non-necroptotic manner. Furthermore, we present robust protocols for detecting human Caspase-8, RIPK1 and MLKL and illustrate their utility for detecting dysregulated necroptosis in biopsies from patients with inflammatory bowel disease (IBD). Collectively, these protocols will empower the definitive evaluation of where and when necroptosis occurs in vivo in health and disease.

Results 91 tissues 80 82 83 84 85 86 87 88 89 90 92 93 94 95 96 100 101 104 Standardised immunohistochemical detection of the necroptotic pathway in mouse We recently compiled a toolbox of immunofluorescence assays to detect necroptotic signaling in cells (38). This toolbox requires use of: 1) non-crosslinking fixatives and 2) gene knockouts to account for non-specific signals; requirements that often cannot be met when immunostaining tissues. Here we aimed to develop robust immunohistochemistry protocols to detect the necroptotic pathway in formalin-fixed paraffin-embedded mouse 97 tissues. Embedding and immunostaining was performed in an automated manner (see 98

Methods) to allow reliable and scalable detection of the necroptotic pathway, and to lessen 99 the future need to account for non-specific immunosignals using appropriate gene knockout controls. The specificity of thirteen monoclonal antibodies against Caspase-8, RIPK1, RIPK3 or MLKL was first tested by immunoblotting spleen homogenates from wild102 type versus knockout mice (Fig. S1). The intensity and specificity of antibodies for 103 immunohistochemistry was then iteratively optimised across 21 conditions (see Methods and Fig. S1). At each optimisation step, immunohistochemistry signals from the spleen of 105

wild-type versus knockout mice were quantified (Fig. 1Ai), ratioed (Fig. 1Aii) and 106 integrated to yield an index of performance (Fig. 1Aiii). For example, this pipeline 107 improved the detection of RIPK1 with the monoclonal antibody D94C12 by approximately 108 three orders of magnitude (Fig. S1). In total, seven automated immunohistochemistry protocols to detect mouse Caspase-8, RIPK1, RIPK3 or MLKL were developed (Fig. 1B). The detection of Caspase-8, RIPK1 and RIPK3 using these immunohistochemistry protocols (Fig. 1B) closely aligned with the abundance of these proteins across multiple 112 tissues as measured by high-resolution quantitative mass spectrometry (Fig. 1C), 113 indicating both specificity and sensitivity. Despite many rounds of optimisation with three specific anti-MLKL antibodies, mouse MLKL remained difficult to detect via 115 immunohistochemistry in all tissues except spleen (Fig. 1B,D and Fig. S1). 109 110 111 114 116 117 118 119 120 122 123 125 126 127 129 130

Close inspection of sites where the necroptotic pathway was constitutively expressed showed unexpected layers of spatial regulation (Fig. 2A, B). In the epithelial Basal expression of the necroptotic pathway is restricted to fast-cycling immune barriers.

The immunohistochemical profile of Caspase-8, RIPK1, RIPK3 and MLKL across seven different organs suggested that expression of the necroptotic pathway is heavily restricted 121 in unchallenged mice (Fig. 1D). For example, RIPK3+ cells were scarce in the kidney and heart, and RIPK3 was undetectable in the brain (Fig. 1D). By comparison, co-expression of Caspase-8, RIPK1 and RIPK3 was evident in intestinal epithelial cells, some splenic 124 regions and Kupffer cells (Fig. 2A). The Tabula Muris single cell RNA sequencing dataset supports the conclusion that expression of the necroptotic pathway is highly restricted in mice (Fig. S2A). Expression of the necroptotic effectors, MLKL and RIPK3, was below detection limits in kidney epithelial, cardiac muscle and resident brain cells, but was 128 frequently detected in progenitor and immune barrier cell populations (Fig. S2A; (41)). barrier of the ileum, Caspase-8 expression was lower in crypts and higher at villi tips, whereas RIPK3 levels peaked in the transit amplifying region and decreased towards the villus tip (Fig. 2Ai). In the colonic epithelial barrier, both Caspase-8 and RIPK3 exhibited higher expression in the base of the crypt and decreased towards the tip of the crypt (Fig. 2Aii). It is noteworthy that expression patterns of Caspase-8 and RIPK3 differ between the small and large intestine because these organs exhibit distinct cell death responses to the same inflammatory stimuli (e.g. TNF) or genetic deficiency (e.g. deletion of Casp8 or Fadd) (42-45). In liver, Kupffer cells expressed Caspase-8, RIPK1 and RIPK3 (Fig. 2Aiii; arrowheads), whereas hepatocytes expressed Caspase-8 and RIPK1, but not RIPK3. 140 Interestingly, Caspase-8 levels were higher in pericentral hepatocytes than in periportal hepatocytes (Fig. 2Aiii; arrow). Zonation of the necroptotic pathway was also evident in 142 the spleen, with Caspase-8, RIPK1, RIPK3 and MLKL levels peaking in the marginal zone 131 132 133 134 135 136 137 138 139 141 143 144 145 146 147 148 149 151 153 154 where circulating antigens are trapped for immune presentation (Fig. 2A,B and Fig. S2B). Prior spatial transcriptomics data support the conclusion that necroptotic potential is zoned along the intestinal crypt-to-villus axis (Fig. S2C; (46)) and along the hepatic central-toportal axis (Fig. S2D; (47)). We also performed spatial transcriptomics on mouse spleen to confirm that necroptotic potential peaks in the marginal zone (Fig. S2E-G).

The expression of RIPK3 appears to be under particularly strict spatial control. For example, in the ileum, RIPK3 levels were high in fast-cycling epithelial progenitors, but low 155 in the Tabula Muris dataset, gene expression of cell cycle markers Top2a and Mki67 156 correlated with the expression of Ripk3, but not Ripk1 (Fig. 2C and Fig. S2A; (41)). Prior 150 in adjacent, terminally differentiated Paneth cells (Fig. 2Ai; open arrowhead). Published single cell transcriptomics data supports that Paneth cells express low levels of RIPK3 152 (48). As another example, RIPK3 levels were high in fast-cycling colonic epithelial cells, but undetectable in slow-cycling renal epithelial cells (Fig. 1D). These observations suggest that RIPK3 expression is linked to cell turnover. Indeed, across 103 cell ontologies cell cultures studies further suggest that the expression and function of RIPK3 fluctuates during the mitotic cell cycle (49, 50).

Altogether, by applying a set of optimised immunohistochemistry protocols to multiple organs, we have found that necroptotic pathway is preferentially expressed at 161 fast-cycling immune barriers under basal conditions. Such targeted expression is consistent with the evolutionary origin of necroptosis being an anti-pathogen defence measure (1, 2, 6). We further find that necroptotic potential is spatially graded along barriers such as the intestinal mucosa. These gradations in the availability of cell death mediators along barriers likely allow multiple cell death programs to be flexibly deployed against invading pathogens (51, 52). 168 Inflammation, dysbiosis or immune challenge trigger local changes in RIPK3 157 158 159 160 162 163 164 165 166 167 169 170 171 172 173 175 176 177 180 181 expression To demonstrate scalability, we used automated immunohistochemistry to characterise the expression of Caspase-8, RIPK1 and RIPK3 across six tissues during TNF-induced Systemic Inflammatory Response Syndrome (SIRS) – a widely-used model of RIPKdependent pathology (Fig. 3A; (19, 45, 53-55)). Littermate wild-type mice were 174 intravenously administered TNF, or vehicle, and tissues harvested 9 hours later when symptoms such as hypothermia were manifesting (Fig. 3B). No major changes to Caspase-8 or RIPK1 expression were observed after TNF administration, except for an unidentified population of RIPK1-expressing cells appearing at the onset of apoptosis in 178 lymphoid tissues (Fig3. C-D; arrowhead). By comparison, RIPK3 was upregulated in 179 intestinal epithelial cells (Fig. 3E-F), certain vascular beds (Fig. 3G-H) and in liver (Fig. 3IJ); the main sites where RIPK1- and RIPK3-signaling during SIRS has been implicated by knockin and knockout mouse studies (19, 45, 53). In contrast, RIPK3 levels were not 182 increased in resident cells of the kidney or heart in TNF-treated mice. Our data therefore suggest that targeted upregulation of RIPK3 in resident cells of the gut and liver underlies RIPK-mediated pathology in SIRS. TNF-treatment also changed the pattern of RIPK3 expression in the intestine, potentially skewing cell death responses in the inflamed gut (Fig. 3E-F). It was surprising that RIPK3 was detected in peri-portal hepatocytes after TNF administration, given that RIPK3 is epigenetically silenced in hepatocytes under basal conditions (56). Collectively, our immunohistochemical characterisation of the SIRS mouse model leads us to propose that the regulation of RIPK3 is akin to an acute phase reactant, with hepatic and intestinal expression that rapidly increases in response to inflammation.

Next, we addressed whether microbiota-depletion affects necroptotic pathway expression. This question was prompted by studies showing that antibiotics offer protection in various models of intestinal necroptosis (43, 57-59). As shown in Fig. 4A, a 183 184 185 186 187 188 189 190 191 192 193 195 196 197 198 199 200 204 205 206 207 208 epithelial Caspase-8 expression remaining constant, while RIPK1 levels were elevated and RIPK3 expression reduced in the crypt and transit amplifying regions of the ileum in antibiotic-treated mice (Fig. 4D). Unexpectedly, immunohistochemistry also showed that microbiota-depletion triggered cytoplasmic accumulations of RIPK1 and RIPK3 in enterocytes at villi tips (Fig. 4D; arrowheads). These clusters of RIPK1 and RIPK3 are 194 litter of wild-type mice was split into two cages and the water for one cage was supplemented with antibiotics for 6 days. As expected, the cecum of antibiotic-treated mice was enlarged and canonical anti-microbial factors such as lysozyme and REG-3β were reduced in the ileum, but not the spleen, of antibiotic-treated mice (Fig. 4B-C). These predictable changes to microbiota-depletion also coincided with a lowering of RIPK3 gene and protein expression in the ileum, but not the spleen (Fig. 4B-C). MLKL gene expression was also suppressed in the ileum, but not the spleen, of antibiotic-treated mice (Fig. 4B). 201 In contrast, Caspase-8 gene and protein expression in the ileum were unaffected by 202

microbiota-depletion, whereas ileal RIPK1 protein levels were increased by antibiotic 203 treatment (Fig. 4B-C). Similar trends were observed by immunohistochemistry, with 209 reminiscent of chylomicrons (60, 61), and were not observed in the colon of antibiotic210 treated mice. Overall, we find that expression of the necroptotic pathway locally responds 211 to changes in the microbiome. This response is spatially restricted to the ileum, zoned along the crypt-to-villus axis, and warrants further investigation given that dysbiosis often occurs in cell death-associated disorders such as Crohn’s disease (62).

Lastly, we used automated immunohistochemistry to uncover a potential nonnecroptotic role for RIPK3 in adaptive immunity. We immunised wild-type mice with the model ligand, NP-KLH (4-hydroxy-3-nitrophenylacetyl hapten conjugated to keyhole limpet hemocyanin), and harvested tissues 14 days later when antigen-specific antibody 212 213 214 215 216 217 221 222 223 224 225 226 228 229 230 231 232 233

RIPK1 or RIPK3 were noted in the ileum of immunised mice and no marked differences in 220 the expression of Caspase-8, RIPK1 or MLKL were observed in the spleen of immunised mice. However, RIPK3 levels were markedly elevated in Ki67+ germinal centres (Fig. S3B; arrowheads). This finding suggests that RIPK3 may have a non-necroptotic role in antibody production. To investigate this possibility, Ripk3+/+, Ripk3+/- and Ripk3-/- littermate mice were immunised with NP-KLH and humoral immune responses measured in blood and spleen 14 days later (Fig. S3C-H). Ripk3-deficiency did not alter circulating white blood cell counts (Fig. S3C), or the total number of class-switched B cells in the spleen 227 (Fig. S3D,E), or the amount of circulating NP-specific antibodies after immunisation (Fig.

S3G,H). Thus, consistent with prior studies (63), RIPK3 does not overtly affect early antigen-specific antibody responses. However, the number of CD138+ NP-specific plasma cells was reduced in the spleen of immunised Ripk3-/- mice (Fig. S3F), suggesting that RIPK3 may have an ancillary role during antigen-driven immunity. Thus, future studies should explore a role of RIPK3 in splenic germinal centres, especially given that RIPK3 has an undefined non-necroptotic role during lymphoproliferative disease (64). 234

In summary, by employing a toolbox of automated immunohistochemical stains, we 235 find that expression of the necroptotic pathway, in particular RIPK3, is responsive to 236 inflammation, dysbiosis or immunisation. These context-specific changes are tightly regulated across space and time, underscoring the need for robust, scalable, in situ assays to pinpoint necroptotic pathway expression and activation.

Automated immunohistochemical detection of the human necroptotic pathway 241 Important differences exist between the human and mouse necroptotic pathways (29, 30, 65-69). For instance, the primary sequence of RIPK3 and MLKL are poorly conserved between species (65), and humans uniquely express Caspase-10 which likely negates necroptotic signaling (70, 71). Thus, in parallel to developing assays for the murine necroptotic pathway, sixteen antibodies against Caspase-8, Caspase-10, RIPK1, RIPK3 or MLKL were tested on wild-type versus knockout formalin-fixed paraffin-embedded HT29 human cells via immunoblot, and then iteratively optimised for immunohistochemistry (see Methods and Fig. S4). While RIPK3 and Caspase-10 remained refractory to 237 238 239 240 242 243 244 245 246 247 248 253 254 255 256 257 258 249 immunohistochemical detection (Fig. S4), five automated immunohistochemistry protocols 250

were developed for human Caspase-8, RIPK1, or MLKL (Fig. S4). These automated 251 immunohistochemistry protocols can detect the low levels of Caspase-8, RIPK1 and MLKL 252 found in human cells during TNF-induced necroptosis, while simultaneously detecting the pool of Caspase-8, RIPK1 and MLKL held at necrosomes (Fig. 5; arrowhead). Indicative of specificity, Caspase-8+ RIPK1+ MLKL+ necrosomes were observed during TNF-induced necroptosis, but not during treatment with TNF alone or during TNF-induced apoptosis (Fig. 5). Notably, the translocation of Caspase-8 and RIPK1, but not MLKL, to necrosomes could also be detected in mouse dermal fibroblasts undergoing TNF-induced necroptosis (Fig. S5). This species-dependent difference is likely due to dissimilarities in the 259 interaction between RIPK3 and MLKL, which is thought to be more transient in mouse 260 than human cells (68). Next, by combining automated immunohistochemistry with high261 resolution digital slide scanning (~250nm resolution) and customised image segmentation, we show that necrosomes can be detected and quantified across a large population of cells in an unbiased manner (Fig. 5B-C). We observed that the accuracy of segmenting human Caspase-8 or MLKL at necrosomes is higher than that of RIPK1, because the small puncta formed by necrosomal RIPK1 are near the resolution limit of existing brightfield slide scanners (Fig. 5C). Nonetheless, as necrosomes are a pathognomonic 267 feature of necroptotic signaling, we propose that machine-based detection of necrosomes 262 263 264 265 266 268 269 270 271 272 273 274 275 276 277 278 284 285 could be developed into a diagnostic assay for pinpointing necroptosis in formalin-fixed human patient biopsies.

Necrosome immunodetection in patients with IBD Ulcerative colitis (UC) and Crohn’s disease (CD) are the main types of inflammatory bowel disease (IBD) (72, 73). The causes of adult onset IBD are multifactorial (74, 75). While many studies show that excess necroptosis promotes IBD-like pathology in mice (18, 42, 58, 76-79), few studies have examined the prevalence of necroptosis in IBD patients (8082). One completed Phase II trial of a RIPK1 inhibitor in UC failed to demonstrate clinical efficacy (83). Another 7 clinical and preclinical trials of RIPK1 inhibitors in IBD are underway. Thus, the role of necroptosis in IBD requires further investigation. We collected 279 intestinal biopsies from adults with UC, CD and non-IBD patients (Fig. 6A). To capture the 280

chronology of disease, biopsies were collected from endoscopically ‘non-inflamed’, 281 ‘marginally inflamed’ and ‘inflamed’ intestinal tissue from each IBD patient. The grading of 282 inflammation was verified by blinded histopathology scores (Fig. 6B). Cell death signaling 283 in biopsies from each location and endoscopic grade was assessed by immunoblot (Fig. 6C and Table S1). To assist interpretation, patient samples (Fig. 6C; blue annotations) were immunoblotted alongside lysates from HT29 cells undergoing apoptosis or arrowheads). Necroptotic signaling was inferred from increases in the abundance of phosphorylated RIPK3 and MLKL, relative to their non-phosphorylated forms (Fig. 6C; asterisks). This approach showed that cell death signaling is elevated in intestinal tissue 291 from IBD patients relative to non-IBD patients, especially in inflamed intestinal biopsies 292 from IBD patients (Fig. 6C). However, marked heterogeneity in the prevailing form of cell death was apparent in both UC and CD patients; with apoptosis dominant in some IBD cases (Fig. 6C; patients B and F), and necroptosis dominant in others (Fig. 6C; patients D and H). Given the ongoing development of RIPK1 inhibitors, it is noteworthy that phosphorylated MLKL coincided with phosphorylated RIPK1 in some, but all not IBD patients (Fig. 6C; patients D and H). The reason why cell death mechanisms vary between patients is currently unknown. Collectively, we find that cell death signaling increases in 299 the inflamed gut, supporting the idea that cell death inhibitors are a potential treatment 286

necroptosis (Fig. 6C; red annotations). Apoptotic signaling was inferred from increases in 287 the conversion of pro-Caspase-3, -8 and -10 into their active cleaved forms (Fig. 6C; open 288 289 290 293 294 295 296 297 298 300 301 302 303 304 305 307 308 310 311 option for IBD. Whether apoptosis or necroptosis manifests in an individual IBD patient appears to be highly variable, highlighting the need for diagnostic approaches, such as automated immunohistochemistry, to identify patients that may benefit from antinecroptotic therapy.

Having identified necroptotic signaling in IBD ‘patient D’ relative to non-IBD ‘patient C’, we applied our automated immunohistochemistry panel to biopsies collected from 306 these same patients, taken at the same time and from the same intestinal locations (Table S1). No obvious changes to epithelial RIPK1 and MLKL were evident between patients C and D. By comparison, cytoplasmic clusters of Caspase-8 were evident in the epithelial 309 layer of the inflamed biopsy from patient D (Fig. 6D; arrowheads). As these cytoplasmic clusters of Caspase-8 were reminiscent of the Caspase-8+ necrosomes in necroptotic HT29 cells (Fig. 5), we used the same high-resolution digital slide scanning and unbiased 312 image segmentation approach as before. This quantitation showed that the number of 313 intraepithelial Caspase-8+ clusters was low in non-IBD patient C and increased with 314 inflammation in IBD patient D (Fig. 6E). This trend mirrors the levels of necroptotic signaling in IBD patient D (Fig. 6B), suggesting that cytoplasmic clusters of Caspase-8+ may represent bona fide necrosomes. This case study provides proof-of-principle that automated immunohistochemical detection of necrosomes is feasible and could be developed into a diagnostic assay for pinpointing necroptosis in clinical practice. 315 316 317 318 319 320 321 323 324 325 327 328 329 330 331 332 333 334 335 336

Discussion

The difficulty of reliably detecting necroptotic signaling in fixed tissues has been a 322 longstanding issue, generating confusion and conflicting results in the literature. To address this problem, we optimised the immunohistochemical detection of Caspase-8, RIPK1, RIPK3 and MLKL in formalin-fixed paraffin-embedded samples. While our prior studies immunostained non-crosslinking fixed monolayers (33, 38), here we used formalin326 fixed paraffin-embedded specimens consistent with standard practice in hospital pathology and research departments around the world. In total, over 280 different immunostaining conditions were tested, yielding 12 automated immunohistochemistry protocols that we anticipate will be of broad utility to the cell death community and drive new insight into the causes, circumstances, and consequences of necroptosis. To assess the reliability of our automated protocols, we benchmarked our immunostaining results against data obtained using other methodologies. For instance, our automated immunohistochemistry protocols produced results that were comparable to public resources generated using proteomic, single cell transcriptomic, and spatial transcriptomic approaches (41, 46, 47, 84). Confidence was also taken from the similar staining patterns produced between three Caspase-8 antibodies (clones D53G2, 3B10 and 1G12), between two RIPK3 antibodies 337 (clones 8G7 and 1H12), and between mouse and human intestinal tissue using the same

RIPK1 antibody (clone D94C12). Thus, multiple lines of evidence suggest that the automated protocols described herein are specific and sensitive.

Our approach for optimising and interpreting immunohistochemical signals relied upon quantitative analyses that carries important technical considerations. First and 342 foremost, the signals produced by immunohistochemistry are non-linear (85). Moreover, before quantitation, we digitally unmixed immunosignals from the haematoxylin counterstain, which is another non-linear transformation of signal intensity (86). Thus, only 345 relative changes in expression levels were inferred from changes in immunohistochemical signal intensity. Because of this caveat, we only compared and quantified immunosignals between closely matched specimens, such as corresponding wild-type and knockout samples where both samples were sectioned at the same time, mounted on the same slide, and stained and imaged contemporaneously. To aid quantitation, all our 350 immunohistochemistry protocols were developed to produce unsaturated signals. One final salient point is that this study used automated embedding and immunostaining procedures, with all quantification using macros that analyse a high number (typically 353 thousands) of cells per sample. Thus, the automated immunohistochemistry protocols 338 339 340 341 343 344 346 347 348 349 351 352 354 355 356 358 359 360 361 362 363 357 to use knockout samples as a control for specificity.

described herein can be used for quantitative purposes, but only when comparing closely matched specimens, and ideally with supporting data from alternative methodologies such as spatial transcriptomics. These recommendations reduce, but do not eliminate, the need

The phosphorylated forms of RIPK1, RIPK3 and MLKL are the most widely used markers of necroptotic signaling (65). Accordingly, prior attempts to detect necroptotic signaling in fixed specimens have focussed on phospho-RIPK1, -RIPK3 and -MLKL (7, 16, 21, 33, 35, 38-40, 87-89). However, there are drawbacks with this approach: 1) antibodies against phosphorylated epitopes in RIPK1/3 and MLKL exhibit much poorer signal-to-noise properties than do antibodies against unphosphorylated epitopes in RIPK1/3 and MLKL (38); 2) while knockout samples are sufficient for verifying non-phospho-signals, multiple controls are needed to authenticate phospho-signals (e.g. resting, knockout and phosphatase pre-treated samples; and 3) because only a small fraction of RIPK1/3 and MLKL is phosphorylated during necroptosis these phospho-species are inherently more difficult to detect than their unphosphorylated counterparts. Indeed, as is standard practice when detecting necroptosis via immunoblot, the immunohistochemical detection of phospho-RIPK1, -RIPK3 and -MLKL can only be interpretated when their unphosphorylated forms are also detected. For these reasons, we focused on the 372 immunohistochemical detection of unphosphorylated RIPK1, RIPK3 and MLKL, and used 373 the immunodetection of necrosomes as a marker of necroptotic signaling.

It was surprising that expression of the necroptotic pathway was heavily restricted under steady-state conditions in mice. That fast-cycling progenitors and immune barrier cells are the dominant expressors of RIPK3 and/or MLKL suggests that the existential role of necroptosis is to protect the host from invading pathogens. The phenotypes and cell 378 types affected in mice carrying activation-prone polymorphisms in Mlkl also supports the 364 365 366 367 368 369 370 371 374 375 376 377 379 380 381 384 385 387 388 389 rapidly derepressed in hepatocytes during inflammation supports the notion that long-lived cells actively suppress necroptotic signaling, but only in the absence of challenge. The derepression of RIPK3 may reconcile the contribution of necroptosis in inflammatory view that the ancestral role of necroptosis lies in innate immunity (90-92). The absence of necroptotic pathway expression in slow-cycling cell populations was equally striking, with RIPK3 and/or MLKL undetectable in resident cells of the heart (except for certain 382 fibroblasts), the brain (except for leptomeningeal vessels), the kidney and the liver (except 383 for Kupffer cells) in mice under basal conditions. These observations challenge a vast body of literature, which we believe highlights the need for robust well-controlled methodologies. These results lead us to propose that slow-cycling cell populations ensure 386 longevity by avoiding inadvertent necroptosis. The observation that RIPK3 expression is 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 disorders such as hepatocellular carcinoma (93), acute myocardial infarction (13), acute ischemic stroke (94) and kidney ischemia-reperfusion injury (15). Notably, the Human Protein Atlas lacks immunohistochemical data for RIPK3 and MLKL, and therefore it remains unknown whether the necroptotic pathway is similarly restricted in humans (95).

To exemplify the utility of our approach, we applied our full immunohistochemistry panel to determine whether inflammation, dysbiosis or immunisation alter necroptotic pathway expression. Each of these challenges altered necroptotic pathway expression in a manner that chiefly involved local shifts in RIPK3 expression. This profiling of the necroptotic pathway yielded

many unexpected observations, including: 1) RIPK3 expression is disinhibited in hepatocytes after TNF administration; 2) RIPK3 expression is suppressed in the gut during antibiotic administration with RIPK1/3 coalescing into unidentified cytoplasmic clusters in epithelial cells at the villus tip; and 3) RIPK3 expression is uniquely upregulated in splenic germinal centres after immunisation. While the mechanistic basis for these findings warrants future attention, their initial description here illustrates the benefits of studying the necroptotic pathway using automated immunohistochemistry.

We provide proof-of-principle that the immunodetection of necrosomes can be used as an in situ marker of necroptotic signaling in IBD patients. To reach this conclusion, we developed a suite of automated immunohistochemical protocols to detect the relocation of necroptotic effectors into necrosomes and validated the presence of necroptotic signaling in closely matched biopsies from IBD patients using immunoblot. Collectively, these experiments show that Caspase-8+ RIPK1+ MLKL+ necrosomes can be readily detectable under idealised cell culture conditions, that elevated necroptotic signaling occurs in a subset of IBD patients, and that cytoplasmic clustering of Caspase-8 correlated with necroptotic signaling across a set of biopsies. Why clusters of Caspase-8, but not clusters of RIPK1 or MLKL, were detectable in biopsies with active necroptotic 416

signaling is unknown, but may be due to technical limitations (e.g. resolution limit) or gaps 417 in our understanding of how necroptosis manifests in vivo. Another important issue is whether the immunohistochemical detection of Caspase-8+ clusters can be used to quantify necroptotic signaling in larger cohorts of IBD patients and in patients with other clinical indications. Notwithstanding these issues, the detection of in situ changes in necroptotic pathway expression in a scalable, quantitative, and automated manner 422 represents a major leap forward in the capacity to pinpoint when and where necroptosis arises in health and disease.

418 419 420 421 423 424 425 426 427 428 429 431 433 434 436 437 438 439 440 441

Cat#3493), rabbit anti-RIPK3 (clone 18H1L23; RRID: AB_2866471; 0.5g/L; Thermo Fisher Scientific Cat#703750), rabbit anti-phospho-RIPK3 (clone D6W2T; RRID:AB_2800206; Cell Signaling Technology Cat#93654), rat anti-RIPK3 (clone 1H12; 2g/L produced inhouse (38)), rat anti-RIPK3 (clone 8G7; RRID: RRID:AB_2940810; 2g/L produced inhouse (6) and available from Millipore Cat#MABC1595), rabbit anti-phospho-MLKL (clone EPR9514; RRID:AB_2619685; Abcam Cat#ab187091; (35)), mouse anti-MLKL (clone Primary antibodies were rat anti-mouse Caspase-8 (clone 3B10; RRID:AB_2490519; 1g/L produced in-house (96) and available from AdipoGen Cat#AG-20T-0138), rat anti-mouse Caspase-8 (clone 1G12; RRID:AB_2490518; 1g/L produced in-house (96) and available

RRID:AB_10545768; Cell Signaling Technology Cat#4790), rabbit anti-phospho-RIPK1 432 (clone D813A; RRID:AB_2799268; Cell Signaling Technology Cat#44590S), mouse antiRIPK1 (clone 38/RIP; RRID:AB_397831; 0.25g/L BD Biosciences Cat#610459), mouse anti-RIPK1 (clone 334640; RRID:AB_2253447; 0.5g/L; R&D Systems Cat#MAB3585), 435 rabbit anti-RIPK1 (clone D94C12; RRID:AB_2305314; Cell Signaling Technology 3D4C6; RRID:AB_2882029; 1.957g/L; Proteintech Cat#66675-1-IG), rabbit anti-mouse MLKL (clone D6W1K; RRID:AB_2799118; Cell Signaling Technology Cat#37705), rat antimouse MLKL (clone 5A6; RRID:AB_2940800; 50g/L produced in-house (38) and available produced in-house (32) and available from Millipore Cat# MABC604), mouse antiCaspase-10 (clone 4C1; RRID:AB_590721; 1g/L; MBL International Cat# M059-3), mouse anti-Caspase-8 (clone B.925.8; RRID:AB_10978471; 0.619g/L Thermo Fisher Scientific Cat# MA5-15226), mouse anti-Caspase-8 (clone 5D3; RRID:AB_590761; 1g/L; MBL 450 International Cat#M058-3), rat anti-human RIPK3 (clone 1H2; RRID:AB_2940816; 2g/L produced in-house (6) and available from Millipore Cat# MABC1640), rabbit anti-human RIPK3 (clone E1Z1D; RRID:AB_2687467; Cell Signaling Technology Cat# 13526), rat antihuman MLKL (clone 7G2; RRID:AB_2940818; 2g/L Millipore Cat# MABC1636), rat antihuman MLKL (clone 10C2; RRID:AB_2940821; 2g/L Millipore Cat# MABC1635), rabbit anti-MLKL (clone 2B9; RRID:AB_2717284; 1g/L; Thermo Fisher Scientific Cat#MA5457 ab184718), mouse anti-GAPDH (clone 6C5; RRID:AB_2107445; 1g/L; Millipore Cat# 442 443 444 446 447 448 449 451 452 453 454 455 458 459 460 461 462 463 464 465 466

MAB374), rabbit anti-smooth muscle actin (clone D4K9N; RRID:AB_2734735; Cell Signaling

Technology

Cat#9245S), rabbit anti-lysozyme (clone EPR2994(2); RRID:AB_10861277; Abcam Cat#108508), cleaved-caspase 3 (Cell Signaling Technology, #9661), Ki67 (Cell Signaling Technology, #12202). The concentration of antibodies from Cell Signaling Technology is often not provided and thus was not listed here.

Secondary antibodies for immunoblotting were horseradish peroxidase (HRP)-conjugated goat anti-rat immunoglobulin (Ig) (Southern BioTech Cat#3010-05), HRP-conjugated goat anti-rabbit Ig (Southern BioTech Cat#4010-05) and HRP-conjugated goat anti-mouse Ig 467 (Southern BioTech Cat#1010-05). Reagents for immunohistochemistry were HRP-conjugated anti-rabbit Ig (Agilent Cat# K400311-2), HRP-conjugated anti-mouse Ig (Agilent Cat# K400111-2), MACH4 universal HRP-Polymer (Biocare Medical Cat#M4U534L), ImmPRESS HRP-conjugated anti-rat IgG 472 for human samples (Vector Laboratories CatcVEMP740450), HRP-conjugated anti rat IgG 473 for mouse samples (R&D Systems Cat#VC005-125), Rabbit Linker (Agilent Cat#GV809112). Epitope Retrieval Solution 1 (Leica Cat#AR9961), Epitope Retrieval Solution 2 (Leica Cat#AR9640), Retrieval Solution Low pH (Agilent Cat#GV80511-2), Retrieval Solution High pH (Agilent Cat#GV80411-2), 3,3'-diaminobenzidine (DAB) substrate (Agilent Cat#GV82511-2 or GV92511-2), Dako REAL Peroxidase-blocking reagent (Agilent S202386-2), bluing reagent (Leica, 3802915), Background Sniper (Biocare Medical Cat#BS966L), Dako Protein Block (Agilent Cat#X0909), ‘Normal’ block (Agilent Cat#S202386-2), EnVision FLEX TRS High pH (Agilent Cat# GV80411-2), EnVision FLEX TRS Low pH (Agilent Cat# GV80511-2), MACH4 universal HRP polymer (Biocare Medical Cat#M4U534L), and DPX (Trajan Cat#EUKITT).

All experiments were approved by The Walter and Eliza Hall Institute (WEHI) Animal Ethics Committee, Australia, in accordance with the Prevention of Cruelty to Animals Act Practice for the Care and Use of Animals for Scientific Purposes (1997), and with the ARRIVE guidelines (97). Mice were housed at the WEHI animal facility under specific pathogen-free, temperature- and humidity-controlled conditions and subjected to a 12 h 491 light/dark cycle with ad libitum feeding. Mice without functional MLKL alleles (Mlkl-/-) have been described previously (32). Mice without functional RIPK3 alleles (Ripk3-/-) have been described previously (98). Mice without functional alleles of RIPK1, RIPK3 and Caspase-8 494 (Ripk1-/- Ripk3−/−Casp8−/− triple knockout mice) and Ripk3-/-Casp8-/- double knockout mice 495 496 507 Immunoblot 508

For Fig. S1 and Fig. S4, RIPA lysates were boiled for 10min in Laemmli sample buffer 509 (126 mM Tris-HCl, pH 8, 20% v/v glycerol, 4% w/v sodium dodecyl sulfate, 0.02% w/v 510

bromophenol blue, 5% v/v 2-mercaptoethanol) and fractionated by 4-15% Tris-Glycine gel 511 (Bio-Rad Cat#5678084) using Tris-Glycine running buffer (0.2M Tris-HCl, 8% w/v SDS, 0.15M glycine). After transfer onto nitrocellulose, membranes were blocked in 1% w/v bovine serum albumin (BSA) or 5% w/v skim cow’s milk in TBS+T (50mM Tris-HCl pH7.4, primary antibodies or 1:1000 for other primary antibodies; supplemented with 0.01% w/v sodium azide; see Materials above) overnight at 4oC, washed twice in TBS+T, probed with an appropriate HRP-conjugated secondary antibody (see Materials above), washed four 518 times in TBS+T and signals revealed by enhanced chemiluminescence (Merck Cat#WBLUF0100) on a ChemiDoc Touch Imaging System (Bio-Rad). Between probing have been described previously (99) and were derived from reported mouse strains (63, 100, 101).

Mouse tissues were homogenised with a stainless steel ball bearing in a Qiagen TissueLyzer II (30 Hz, 1min) in ice-cold RIPA buffer (10mM Tris-HCl pH 8.0, 1mM EGTA, 2mM MgCl2, 0.5% v/v Triton X100, 0.1% w/v sodium deoxycholate, 0.5% w/v sodium dodecyl sulfate (SDS) and 90mM NaCl) supplemented with 1x Protease & Phosphatase 504 Inhibitor Cocktail (Cell Signaling Technology Cat#5872) and 100 U/mL Benzonase (SigmaAldrich Cat#E1014). 1mL of RIPA buffer per 25mg of tissue was used for homogenisation. 1224 1225 1226 1227 1228 1229 1230 1231 1232 1233 1234 1235 1236 1237 1238 1239 1240 1241 1242 1243 124 4 1245 1246 1247 1248

A. G. Bassuk, B. W. Darbro, M. A. Fiatarone Singh, N. Vlahovich, D. Hughes, M. Kozlovskaia, D. B. Ascher, K. Warnatz, N. Venhoff, J. Thiel, C. Biben, S. Blum, J. Reveille, M. S. Hildebrand, C. G. Vinuesa, P. McCombe, M. A. Brown, B. T. Kile, C. McLean, M. Bahlo, S. L. Masters, H. Nakano, P. J. Ferguson, J. M. Murphy, W. S. Alexander, J. Silke, A missense mutation in the MLKL brace region promotes lethal neonatal inflammation and hematopoietic dysfunction.

Nat Commun 11, 3150 (2020). 91.

S. E. Garnish, K. R. Martin, M. Kauppi, V. E. Jackson, R. Ambrose, V. V. Eng, S. Chiou, Y. Meng, D. Frank, E. C. Tovey Crutchfield, K. M. Patel, A.V. Jacobsen, G. K. Atkin-Smith, L. Di Rago, M. Doerflinger, C. R. Horne, C. Hall, S. N. Young, M. Cook, V. Athanasopoulos, C. G. Vinuesa, K. E. Lawlor, I. P. Wicks, G. Ebert, A. P. Ng, C. A. Slade, J. S. Pearson, A. L. Samson, J. Silke, J. M. Murphy, J. M. Hildebrand, A common human MLKL polymorphism confers resistance to negative regulation by phosphorylation. Nat Commun 14, 6046 (2023). 92.

X. Zhu, N. Yang, Y. Yang, F. Yuan, D. Yu, Y. Zhang, Z. Shu, N. Nan, H. Hu, X. Liu, S. Chen, L. Sun, H. Wang, Spontaneous necroptosis and autoinflammation are blocked by an inhibitory phosphorylation on MLKL during neonatal development. Cell Res 32, 407-410 (2022). 93.

M. Vucur, A. Ghallab, A. T. Schneider, A. Adili, M. Cheng, M. Castoldi, M. T. Singer, V. Buttner, L. S. Keysberg, L. Kusgens, M. Kohlhepp, B. Gorg, S. Gallage, J. E. Barragan Avila, K. Unger, C. Kordes, A. L. Leblond, W. Albrecht, S. H. Loosen, C. Lohr, M. S. Jordens, A. Babler, S. Hayat, D. Schumacher, M. T. Koenen, O. Govaere, M. V. Boekschoten, S. Jors, C. VillacortaMartin, V. Mazzaferro, J. M. Llovet, R. Weiskirchen , J. N. Kather, P. Starlinger, M. Trauner, M. Luedde, L. R. Heij, U. P. Neumann, V. Keitel, J. G. Bode, R. K. Schneider, F. Tacke, B. Levkau, T. Lammers, G. Fluegen, T. Alexandrov, A. L. Collins, G. Nelson, F. Oakley, D. A. Mann, C. Roderburg, T. Longerich, A. Weber, A. Villanueva, A. L. Samson, J. M. Murphy, R. Kramann, F. Geisler, I. G. Costa, J. G. Hengstler, M. Heikenwalder, T. Luedde, Sublethal necroptosis signaling promotes inflammation and liver cancer. Immunity 56, 1578-1595 e1578 (2023). 94.

A. Degterev, Z. Huang, M. Boyce, Y. Li, P. Jagtap, N. Mizushima, G. D. Cuny, T. J. Mitchison, M. A. Moskowitz, J. Yuan, Chemical inhibitor of nonapoptotic cell death with therapeutic potential for ischemic brain injury. Nat Chem Biol 1, 112-119 (2005).

M. Uhlen, L. Fagerberg, B. M. Hallstrom, C. Lin dskog, P. Oksvold, A. Mardinoglu, A. Sivertsson, C. Kampf, E. Sjostedt, A. Asplund, I. Olsson, K. Edlund, E. Lundberg, S. Navani, C. A. Szigyarto, J. Odeberg, D. Djureinovic, J. O. Takanen, S. Hober, T. Alm, P. H. Edqvist, H. Berling, H. Tegel, J. Mulder, J. Rockberg, P. Nilss on, J. M. Schwenk, M. Hamsten, K. von Feilitzen, M. Forsberg, L. Persson, F. Johansson, M. Zwahlen, G. von Heijne, J. Nielsen, F.

Ponten, Proteomics. Tissue-based map of the human proteome. Science 347, 1260419 1249 1250 1251 1252 1253 1254 1255 1256 1257 1258 1259 1260 1261 1262 1263 1264 1265 1266 1267 1268 1269 1270 1271 1272

L. A. O'Reilly, U. Divisekera, K. Newton, K. Scalzo, T. Kataoka, H. Puthalakath, M. Ito, D. C. Huang, A. Strasser, Modifications and intracellular trafficking of FADD/MORT1 and caspase8 after stimulation of T lymphocytes. Cell Death Differ 11, 724-736 (2004). 97.

N. Percie du Sert, V. Hurst, A. Ahluwalia, S. Alam, M. T. Avey, M. Baker, W. J. Browne, A. Clark, I. C. Cuthill, U. Dirnagl, M. Emerson, P. Garner, S. T. Holgate, D. W. Howells, N. A. Karp, S. E. Lazic, K. Lidster, C. J. MacCallum, M. Macleo d, E. J. Pearl, O. H. Petersen, F. Rawle, P. Reynolds, K. Rooney, E. S. Sena, S. D. Silberberg, T. Steckler, H. Wurbel, The ARRIVE guidelines 2.0: Updated guidelines for reporting animal research. PLoS Biol 18, e3000410 98.

E. C. Tovey Crutchfield, S. E. Garnish, J. Day, H. Anderton, S. Chiou, A. Hempel, C. Hall, K. M. Patel, P. Gangatirkar, K. R. Martin, C. S. N. Li Wai Suen, A. L. Garnham, A. J. Kueh, I. P. Wicks, J. Silke, U. Nachbur, A. L. Samson, J. M. Murphy, J . M. Hildebrand, MLKL deficiency protects against low-grade, sterile inflammation in aged mice. Cell Death Differ 30, 1059-1071 99. RIPK1 regulates RIPK3-MLKL-driven systemic inflammation and emergency hematopoiesis.

Cell 157, 1175-1188 (2014).

M. A. Kelliher, S. Grimm, Y. Ishida, F. Kuo, B. Z. Stanger, P. Leder, The death domain kinase RIP mediates the TNF-induced NF-kappaB signal. Immunity 8, 297-303 (1998). 101.

L. Salmena, B. Lemmers, A. Hakem, E. Matysiak-Zablocki, K. Murakami, P. Y. Au, D. M. Berry, L. Tamblyn, A. Shehabeldin, E. Migon, A. Wakeham, D. Bouchard, W. C. Yeh, J. C. McGlade, P. S. Ohashi, R. Hakem, Essential role for caspase 8 in T-cell homeostasis and T-cellmediated immunity. Genes Dev 17, 883-895 (2003). 102. Tomancak, A. Cardona, Fiji: an open-source platform for biological-image analysis. Nat 103.

Y. Hao, S. Hao, E. Andersen-Nissen, W. M. Mauck, 3rd, S. Zheng, A. Butler, M. J. Lee, A. J. Srivastava, T. Stuart, L. M. Fleming, B. Yeung, A. J. Rogers, J. M. McElrath, C. A. Blish, R. Gottardo, P. Smibert, R. Satija, Integrated analysis of multimodal single-cell data. Cell 184, 3573-3587 e3529 (2021). 1888-1902 e1821 (2019). 104.

T. Stuart, A. Butler, P. Hoffman, C. Hafemeister, E. Papalexi, W. M. Mauck, 3rd, Y. Hao, M.

Stoeckius, P. Smibert, R. Satija, Comprehensive Integration of Single-Cell Data. Cell 177, 1297 1298 1299 1300 1301 1302 1303 1304 1305 1306 1307 1308 1309 1310 1311 1312 1313 1314 1315 1316 1317 1318 1319 1320 1321 data across different conditions, technologies, and species. Nat Biotechnol 36, 411-420 (2018). Transcription Factor T-bet in B Cells Modulates Germinal Center Polarization and Antibody Affinity Maturation in Response to Malaria. Cell Rep 29, 2257-2269 e2256 (2019). 108. A. Chen, S. Liao, M. Cheng, K. Ma, L. Wu, Y. Lai, X. Qiu, J. Yang, J. Xu, S. Hao, X. Wang, H. Lu, X. Chen, X. Liu, X. Huang, Z. Li, Y. Hong, Y. Jiang, J. Peng, S. Liu, M. Shen, C. Liu, Q. Li, Y. Yuan, X. Wei, H. Zheng, W. Feng, Z. Wang, Y. Liu, Z. Wang, Y. Yang, H. Xiang, L. Han, B. Qin, P. Guo, G. Lai, P. Munoz-Canoves, P. H. Maxwell, J. P. Thiery, Q. F. Wu, F. Zhao, B. Chen, M. Li, X. Dai, S. Wang, H. Kuang, J. Hui, L. Wang, J. F. Fei, O. Wang, X. Wei, H. Lu, B. Wang, S. Liu, Y. Gu, M. Ni, W. Zhang, F. Mu, Y. Yin, H. Yang, M. Lisby, R. J. Cornall, J. Mulder, M. Uhlen, M. A. Esteban, Y. Li, L. Liu, X. Xu, J. Wang, Spatiotemporal transcriptomic atlas of mouse organogenesis using DNA nanoball-patterned arrays. Cell 185, 1777-1792 e1721 (2022). 109. M. Martin, Cutadapt removes adapter sequences from high-throughput sequencing reads.

2011 17, 3 (2011). genotyping with HISAT2 and HISAT-genotype. Nat Biotechnol 37, 907-915 (2019). assigning sequence reads to genomic features. Bioinformatics 30, 923-930 (2014). alignment and quantification of RNA sequencing reads. Nucleic Acids Res 47, e47 (2019). 113.

M. D. Robinson, A. Oshlack, A scaling normalization method for differential expression analysis of RNA-seq data. Genome Biol 11, R25 (2010).

analysis tools for RNA-seq read counts. Genome Biol 15, R29 (2014). Approach to Multiple Testing. Journal of the Royal Statistical Society: Series B (Methodological) 57, 289-300 (1995).

Simpson, X. Dong, L. Whitehead, J. C. Tempany, A. J. Farchione, A. A. Sheikh, J. R. Groom, K. L. Rogers, M. J. Herold, V. L. Bryant, M. E. Ritchie, S. N. Willis, R. W. Johnstone, P. D. Hodgkin, S. L. Nutt, S. J. Vervoort, E. D. Hawkins, Epigenetic modulators of B cell fate identified through coupled phenotype-transcriptome analysis. Cell Death Differ 29, 2519117. K. G. Smith, A. Light, L. A. O'Reilly, S. M. Ang, A. Strasser, D. Tarlinton, bcl-2 transgene expression inhibits apoptosis in the germinal center and reveals differences in the selection of memory B cells and bone marrow antibody-forming cells. J Exp Med 191, 475-484 2530 (2022).

(2000). 118. C. W. Walker, S. Nelson, M. K. Vandervoort, V. Frisbie, M. A. Samaan, V. Jairath, D. K. Driman, K. Geboes, M. A. Valasek, R. K. Pai, G. Y. Lauwers, R. Riddell, L. W. Stitt, B. G.

Levesque, Development and validation of a histological index for UC. Gut 66, 50-58 (2017). 119. A. V. Jacobsen, C. L. Pierotti, K. N. Lowes, A. E. Au, Y. Zhang, N. Etemadi, C. Fitzgibbon, W. J.

A. Kersten, A. L. Samson, M. F. van Delft, D. C. S. Huang, H. J. Sabroux, G. Lessene, J. Silke, J. M. Murphy, The Lck inhibitor, AMG-47a, blocks necroptosis and implicates RIPK1 in signalling downstream of MLKL. Cell Death Dis 13, 291 (2022). 1347 1348 1349 1350 1351 1352 1353 1354 1355 120. C. Bossen, K. Ingold, A. Tardivel, J. L. Bodmer, O. Gaide, S. Hertig, C. Ambrose, J. Tschopp, P.

Schneider, Interactions of tumor necrosis factor (TNF) and TNF receptor family members in the mouse and human. J Biol Chem 281, 13964-13971 (2006).

Chunduru, S. M. Condon, M. McKinlay, R. Brink, M. Leverkus, V. Tergaonkar, P. Schneider, B. A. Callus, F. Koentgen, D. L. Vaux, J. Silke, IAP antagonists target cIAP1 to induce TNFalphadependent apoptosis. Cell 131, 682-693 (2007). cellular segmentation. Nat Methods 18, 100-106 (2021). 123. R. Haase, L. A. Royer, P. Steinbach, D. Schmidt, A. Dibrov, U. Schmidt, M. Weigert, N.

Maghelli, P. Tomancak, F. Jug, E. W. Myers, CLIJ: GPU-accelerated image processing for everyone. Nat Methods 17, 5-6 (2020). 124. A. Pedersen, A. Valla, M. Bofin, J. P. De Frutos, I. Reinersten, E. Smistad, FastPathology: An Open-Source Platform for Deep Learning-Based Research and Decision Support in Digital Pathology. IEEE Access 9, 58216-58229 (2021). 125. H. S. Pettersen, I. Belevich, E. S. Royset, E. Smistad, M. R. Simpson, E. Jokitalo, I. Reinertsen, I. Bakke, A. Pedersen, Code-Free Development and Deployment of Deep Segmentation Models for Digital Pathology. Front Med (Lausanne) 8, 816281 (2021).

We thanks James Vince, Jiyi Pang, David Tarlington, Angus Stock, Najoua Lalaoui, Asha Jois, Marcel Doerflinger, Quentin Gouil and Bruce Rosengarten for constructive feedback and reagents during the preparation of this manuscript. We thank WEHI Monoclonal antibody facility for producing several antibodies used in this study. We thank the WEHI Histology team for high level support with immunohistochemistry. We thank the WEHI Bioservices team for high level support for animal experimentation, welfare, and ethics.

This work was supported by National Health and Medical Research Council of Australia 1383 (grants 1172929 to JMM, 2008652 to EDH, 2002965 to ALS, and the Independent Research Institutes Infrastructure Support Scheme 9000719), by the Kenneth Rainin Foundation (award to JMM, BC, AHA, ALS), and by the Victorian State Government Operational Infrastructure Support scheme. SC is supported by the WEHI Handman PhD scholarship, and AHA by the Avant Foundation and by Crohn’s and Colitis Australia. JMM and JS received research funding from Anaxis Pharma Pty Ltd, from which the salaries of KMP, AH, AL and PG were paid.

KMP, SNY, AH, AL, PG, CRH, JD, IPW, JMH, JS, JMM and ALS have contributed to the development of necroptosis pathway inhibitors in collaboration with Anaxis Pharma Pty Ltd. All other authors have no additional financial interests.

Author contributions 1397 1398 1399 1400 1401 1402

Data collection and experimental assistance: SC, AHA, YP, KMP, AL, SNY, MB, LZ, AH, AL, CH, PG, WC, JMH, ET, ALS. Data analysis and interpretation: SC, AHA, YP, LWW, IYK, AL, CS, RKHY, PR, JAR, JMH, EDH, BC, JMM, ALS. Conceptualization: JMM, ALS. Essential reagents: CH, AVJ, CRH, DSH, LJI, JD, IPW, LAO, JS, EDH, JMM. Manuscript writing: SC, AHA, JMM, ALS. Reagents, supervision, and infrastructure: DH, IPW, CL, MER, RB, LAO, JS, LG-R, ET, KLR, EDH, BC, JMM, ALS.

C8 (Caspase-8), C10 (Caspase-10), cl. C3 (cleaved Caspase-3), TNF (Tumour Necrosis Factor), RIPK1 (Receptor-interacting serine/threonine-protein kinase-1), pRIPK1 (phosphorylated RIPK1), RIPK3 (Receptor-interacting serine/threonine-protein kinase-3), pRIPK3 (phosphorylated RIPK3), MLKL (Mixed lineage kinase domain-like protein), pMLKL (phosphorylated MLKL), R1 (Receptor-interacting serine/threonine-protein kinase1), R3 (Receptor-interacting serine/threonine-protein kinase-1), ML (Mixed lineage kinase domain-like protein), WEHI (Walter and Eliza Hall Institute of Medical Research), WT (wildtype), KO (knockout), IBD (inflammatory bowel disease), UC (ulcerative colitis), CD (Crohn’s disease), SMA (smooth muscle actin), Kpf (Kupffer cell), endo (endothelial cell), p.o. (oral administration), i.p. (intraperitoneal administration), CV (central vein), PV (portal vein), BD (bile duct), CA (splenic central artery), WP (splenic white pulp), MZ (splenic marginal zone), RP (splenic red pulp), CRP (C-reactive protein), TS (Tumour Necrosis Factor and SMAC mimetic), TSI (Tumour Necrosis Factor and SMAC mimetic and IDN6556), Ab (antibody), Uniform Manifold Approximation and Projection (UMAP), NP-KLH (4hydroxy-3-nitrophenylacetyl hapten conjugated to keyhole limpet hemocyanin).

Figure legends:

Fig. 1. Automated immunohistochemistry shows constitutive necroptotic pathway expression is restricted. A To gauge immunohistochemistry performance, 1424 immunosignals from wild-type (WT) versus knockout (KO) tissue were deconvoluted, (i) 1425

pixel intensities plotted, (ii) ratioed to yield a signal-to-noise (S/N) histogram, and then (iv) 1426 integrated. B Heatmap shows relative integrated S/N values from 7 automated 1427 immunohistochemistry protocols across 7 tissues. Column headers indicate the antibody 1428 target clone name. Data are representative of n≥3 for each target and tissue. C Heatmap depicts relative protein abundance values as measured by (84). D Immunosignals of Caspase-8, RIPK1, RIPK3 and MLKL in wild-type versus the appropriate knockout (KO) 1431 tissue from Mlkl-/- or Casp8-/-Ripk3-/- or Casp8-/-Ripk1-/-Ripk3-/-. Data are representative of 1422 1423 1429 1430 n≥3 for each target and tissue. Scale bars are 500µm. Related to Fig. S1.

Fig. 2. Necroptotic potential is spatially graded across tissue zones. A 1435 Immunosignals of Caspase-8, RIPK1 and RIPK3 from wild-type mouse ileum (i), colon (ii), 1436 liver (iii) and spleen (iv). The crypt base (crypt), villi/crypt tip (tip), central vein (CV), portal vein (PV), bile duct (BD), central artery (CA), white pulp (WP), marginal zone (MZ) and red pulp (RP) are annotated. Inset of immunostaining in the ileum shows lower RIPK3 expression in Paneth cells (open arrowhead) relative to neighbouring cells. Arrow shows peri-central hepatocytes that express higher levels of Caspase-8. Closed arrowheads show Caspase-8+ RIPK1+ RIPK3+ Kupffer cells. Scale bars are 50µm, except for 10µm scale bar in inset. Data are representative of n≥3 for each target and tissue. B Relative expression levels of Caspase-8, RIPK1, RIPK3 (and splenic MLKL; Fig. S2B) along the 1444 indicated tissue axes. Red datapoints indicate immunosignal intensities, and the overlaid dark blue line indicating the LOWESS best-fit along N=20 axes per tissue. Best-fit curves are superimposed in the left-most column. Dashed line indicates the boundary between splenic white pulp and marginal zone. Data are representative of n>3 mice per target per 1448 tissue. C Scatterplots where each dot represents a different cell ontology from the Tabula Muris dataset (41). The percent of cells within each ontology that expressed Mki67, Ripk1 or Ripk3 was plotted against that of Top2a. Pearson correlation coefficient values are shown. Related to Fig. S2.

Fig. 3. RIPK3 expression is rapidly altered during systemic inflammation. A Experimental design. B Core temperatures of vehicle- and TNF-injected mice (n=3 mice per group). C Immunosignals for cleaved Caspase-3 (cl. C3), Caspase-8, RIPK1 or RIPK3 1456 from the spleen of vehicle- or TNF-injected mice. Insets show unidentified RIPK1high cells 1457 that associate with apoptotic bodies in splenic white pulp. D Graph of white pulp area occupied by cleaved Caspase-3+ material in vehicle- and TNF-treated mice. Each red datapoint represents one white pulp lobule (N=20 lobules/mouse). Blue datapoints indicate 1460 the median value per mouse (n=3 mice/treatment). Black bars represent the mean value per group. *p<0.05 by unpaired 2-tailed t-test. E RIPK3 immunosignals in colon of vehicleor TNF-treated mice. F Best-fit curves of RIPK3 immunosignals along the crypt-to-tip axis 1463 from N=10 axes per mouse (n=3 mice/group). ***p<0.001 by multiple unpaired 2-tailed t1464 test. G RIPK3 (pink) and smooth muscle actin (brown) immunosignals in intestinal submucosa of vehicle- or TNF-treated mice. Insets show vessel cross-sections.