INTRODUCTION
Apoptosis and necrosis are two morphologically and mechanistically distinct types of cell death
1. Apoptosis is executed by caspases without cytoplasmic membrane disruption, whereas necrosis is characterized by the early loss of cytoplasmic membrane integrity. Apoptosis, mediated by an evolutionarily conserved cell death pathway, is involved in regulating cell death during development, adult homeostasis and host defense responses. The activation of caspases, such as caspase-3 and caspase-8, as indicated by proteolytic cleavages in specific substrates after Asp residues, provides biomarkers for apoptosis. Apoptotic cells are removed by efferocytosis, which mediates rapid removal and degradation of apoptotic bodies
2,3. Necroptosis is a form of regulated necrotic cell death mediated by RIPK1, RIPK3 and MLKL
1. The activation of RIPK1, RIPK3 and MLKL, as marked by their specific phosphorylation events, provides biomarkers for necroptosis. In addition, the cleavage of gasdermins, such as GSDMD, mediates the execution of pyroptosis by disrupting the integrity of cytoplasmic membrane
4,5. The caspase-mediated cleavage of GSDMD provides the biomarker for pyroptosis
5,6. However, we still know very little about the terminal events during necroptosis or pyroptosis.
In addition to regulated cell death mechanisms mentioned above, cells may also die passively due to overwhelming stress, such as ferroptosis
7. Ferroptosis is induced by excessive lipid peroxidation, e.g., by depriving cells of the essential GSH precursor or by inhibiting GPX4, a key member of the glutathione peroxidase family which catalyzes the reduction of hydrogen peroxide and lipid hydroperoxides to protect cells against oxidative damage
8. However, while overwhelming lipid peroxidation can induce ferroptosis, lipid peroxidation can be a result of certain reversible metabolic processes and therefore, is not necessarily indicative of cell death
9. Accordingly, current approaches primarily report biochemical features associated with ferroptosis initiation rather than the terminal execution of cell death. The lack of specific markers for the terminal necrotic state shared by ferroptosis and other lytic cell death pathways makes it difficult to pinpoint when and where irreversible necrosis occurs in diseases.
Caspase-mediated cleavage of numerous protein substrates is the irreversible biochemical event responsible for apoptotic cell death process including phosphatidylserine exposure
10, apoptotic membrane blebbing and apoptotic DNA fragmentation
11,12. In contrast, we still know very little about the involvement of proteolytic cleavage during necrotic cell death. Here we investigated dynamic morphological changes and characteristic biochemical hallmarks of necrosis distinctive from those of apoptosis. We found that these necrotic hallmarks can be used to define necrosis, including necroptosis and ferroptosis, while pyroptosis is executed by either apoptosis or necrosis. Using mass spectrometry, we performed N-terminomic analyses to characterize proteolytic cleavage events in necrosis, both
in vitro and
in vivo. This approach identified distinctive necrotic cleavages occurring after arginine and lysine residues. We subsequently developed monoclonal antibodies against selective neo-N-termini to serve as biomarkers of necrosis. Furthermore, our study demonstrates the role of extracellular trypsin-like proteases in these cleavage events, which facilitate the phagocytosis of necrotic nuclei and the degradation of nuclear DNA to prevent autoimmunity. Taken together, our study provides insights into the terminal events in necrosis and biomarkers for characterizing necrosis
in vivo.
RESULTS
Necrotic cleavages of lamin-B1 and lamin-A/C in necrosis
We analyzed the morphological changes of apoptosis and necrosis using live cell panoramic super-resolution microscopy based on an improved label-free three-dimensional optical diffraction tomography (super resolution ODT) with high-speed tomographic scanning and enhanced spatial resolution
13. Super resolution ODT offers an unbiased and label-free view of the quantitative mass density distributions as refractive index in living cells and organisms. We quantified the dynamic changes of refractive index in nuclear envelope and nucleolus during cell death using the culture medium as a reference. Established protocols were used to induce necroptosis (with TNF/SM164/zVAD (TSZ)
14 and LPS/zVAD
15), apoptosis (with TNF/SM164/GSK872 (TSG)
16,17), RIPK1-dependent apoptosis (with LPS/5z7
18), RIPK1-independent apoptosis (with TNF/CHX (TC)
19), pyroptosis (with LPS/Nigericin
4), and ferroptosis (with RSL3
20,21). Necrosis is defined by early plasma membrane disruption
22. By super resolution ODT, we observed that the nuclear envelope and nucleolus of HT22 cells and bone marrow-derived macrophages (BMDMs) became diffusive (quantified by the reduction in their refractive index) in early-stage necrotic cell death after induction of necroptosis and ferroptosis, suggesting that necrosis can affect both nuclear envelope and nucleus (Fig. 1a, b). In contrast, apoptosis led to the formation of characteristic apoptotic bodies with the nuclear envelope staying relatively intact until disruption of cytoplasmic membrane integrity by blebbing (Fig. 1a, b). In BMDMs treated with LPS/Nigericin, a known paradigm that can induce pyroptosis
23, we observed individual dying cells with either apoptotic morphology characterized by apoptotic bodies, or necrotic morphology characterized by diffused nuclear envelope and nucleolus (Fig. 1b). Since LPS promotes the activation of caspase-1 and caspase-11 which can activate caspase-3 to promote apoptosis
24 and the cleavage of GSDMD to form pores on plasma membrane to promote necrosis
5, these observations suggest that treatment with LPS/Nigericin leads to a combination of apoptosis and necrosis, with individual cells may die by either apoptosis or necrosis.
We next characterized the necrotic nuclear morphology by immunofluorescence (IF) staining of lamin-B1 as a biomarker for nuclear envelope and Nup98 as a biomarker for nuclear pore complex. Interestingly, induction of necroptosis in HT22 cells, BMDMs and HT29 cells, as well as induction of ferroptosis in HT22 cells and HT1080 cells, all led to the loss of IF signal for lamin-B1, but not Nup98, suggesting that necrosis may lead to specific destruction of the nuclear envelope (Fig. 1c, d; Supplementary Fig. S1a, b). In contrast, apoptosis of HT22 cells, BMDMs, HT29 cells and HT1080 cells showed a deformed nuclear envelope marked by lamin-B1 and Nup98 (Fig. 1c, d; Supplementary Fig. S1a, b). With destruction of the nuclear envelope, we observed nuclear DNA loss in necroptosis of HT22 cells and ferroptosis of HT1080 and HT22 cells (Fig. 1c; Supplementary Fig. S1b, c). In BMDMs treated with LPS/Nigericin, we observed some dying cells with deformed nuclear envelope marked by lamin-B1 and Nup98 signal as that in apoptotic cells, as well as some dying cells without lamin-B1 IF as that in necroptotic cells, suggesting again that LPS/Nigericin can induce some cells to die by apoptosis and some by pyroptosis (Fig. 1d). The cleaved caspase-3 was found in BMDMs treated with TSG as well as LPS/Nigericin, supporting the involvement of apoptosis in both paradigms (Fig. 1e; Supplementary Fig. S1d). Thus, by using super resolution ODT and IF imaging, we discovered nuclear envelope destruction characterized by the loss of lamin-B1 but not Nup98 IF during necrotic cell death, including necroptosis and ferroptosis, which is distinguishable from nuclear envelope deformation marked by lamin-B1 and Nup98 IF during apoptosis. In addition, we discovered cells with either necrotic or apoptotic changes in nuclear envelope during pyroptosis, with necrotic dying cells losing lamin-B1 but not Nup98 IF signals, or apoptotic dying cells showing nuclear envelope deformation marked by lamin-B1 and Nup98.
Apoptosis is known to lead to the cleavage of lamin-B1 producing a product of ~40 kDa
25,26 and lamin-A/C producing a product of ~30 kDa
26,27, which was confirmed in apoptosis of HT22 cells and BMDMs induced by TSG, as well as apoptosis of BMDMs induced by LPS/5z7 (Fig. 1f, g). In contrast, induction of necroptosis in HT22 cells by TSZ, ferroptosis in HT22 cells by RSL3 and necroptosis in BMDMs by LPS/zVAD led to larger cleavage products of lamin-B1 at ~50 kDa and of lamin-A/C at ~55 kDa (Fig. 1f, g). Necrotic cleavages of lamin-B1 (50 kDa) and lamin-A/C (55 kDa) were blocked by treatment with RIPK1 inhibitor Nec-1s and RIPK3 inhibitor GSK872 in necroptosis of HT22 cells (induced by TSZ) and BMDMs (induced by LPS/zVAD), by
RIPK3 knockout in necroptosis of BMDMs (induced by LPS/zVAD), and by Liproxstatin-1 (Lip-1) in RSL3-induced ferroptosis of HT22 cells (Fig. 1f, g). The distinguishable apoptotic and necrotic cleavage products of lamin-B1 and lamin-A/C were also found in human HT29 cells and HT1080 cells undergoing apoptosis by treatment with TC, necroptosis by treatment with TSZ and ferroptosis by treatment with RSL3 (Supplementary Fig. S1e, f). Necrotic cleavages of lamin-B1 and lamin-A/C in necroptosis of HT29 cells (induced by TSZ) were inhibited by the treatment with Nec-1s, GSK872 and MLKL inhibitor NSA (Supplementary Fig. S1e). BMDMs treated with LPS/Nigericin led to the production of both apoptotic ~42 kDa lamin-B1 as well as necrotic 50 kDa lamin-B1, supporting the involvement of both apoptosis and necrosis (Fig. 1g, h). Consistent with the IF detection of cleaved caspase-3 during pyroptosis, western blotting analysis also confirmed the presence of cleaved caspase-3 in LPS/Nigericin-treated cells (Fig. 1h). Notably, caspase-3 activation was increased in correspondence with the cleaved caspase-1 in a Nigericin dose-dependent manner (Fig. 1h). These findings suggest that nuclear envelope destruction during necrotic cell death may involve the cleavage of nuclear envelope proteins, such as lamin family members including lamin-B1, by a mechanism distinct from that of apoptotic nuclear envelope destruction.
This 50 kDa necrotic cleavage product of lamin-B1 was also observed
in vivo in the mouse kidneys after ischemia-reperfusion injury (IRI)-induced acute kidney injury (AKI) known to involve necrosis
28 (Fig. 1i; Supplementary Fig. S1g). The
in vivo observation suggests that the enzymatic activity underlying necrotic lamin-B1 cleavage exists in intact tissues
in vivo and can be engaged during pathological injury, supporting the physiological relevance of this cleavage event beyond cultured cell systems. Taken together, the distinguishable cleavage products of lamin-B1 and lamin-A/C and distinctive changes in nuclear envelope morphology during apoptosis and necrosis suggest the involvement of different enzymatic mechanisms during the final stage of cell death execution in apoptosis and necrosis.
Inhibition of necrotic protein cleavage by leupeptin
We next characterized the enzymatic mechanism that mediates protein cleavages in necrosis using necrotic lamin-B1 cleavage as a biomarker. After screening with various protease inhibitors (Fig. 2a; Supplementary Fig. S2a), we found that treatment with a cocktail that contains 6 different protease inhibitors (AEBSF, Aprotinin, Bestatin, E64, leupeptin and Pepstatin A) could reduce the necrotic lamin-B1 cleavage product (50 kDa) in necroptosis of BMDMs without affecting the levels of phosphorylated MLKL (p-MLKL), and in pyroptosis without affecting the cleavage of GSDMD or the apoptotic lamin-B1 cleavage product (42 kDa) (Fig. 2a). We next examined the effects of these protease inhibitors individually on the necrotic cleavage of lamin-B1. Treatment with leupeptin alone was as effective as the triple combination of leupeptin, Aprotinin and AEBSF in inhibiting necrotic lamin-B1 cleavage in necroptosis of BMDMs induced by LPS/zVAD (Fig. 2b), necroptosis of HT22 cells induced by TSZ and ferroptosis of HT22 cells induced by RSL3 (Fig. 2c). Notably, leupeptin alone also showed comparable efficacy to the triple combination in suppressing necrotic lamin-B1 cleavage without affecting apoptotic cleavage of lamin-B1(42 kDa) in pyroptosis of BMDMs treated with LPS/Nigericin (Fig. 2b). Similarly, both leupeptin alone and the triple combination inhibited necrotic lamin-B1 cleavage in TSZ-induced necroptotic HeLa cells expressing RIPK3 (HeLa-RIPK3), as well as in digitonin-treated HeLa-RIPK3 cells (Fig. 2d). The presence of necrotic lamin-B1 cleavage after digitonin treatment — an acute, signaling-independent membrane rupture model — supports the conclusion that this proteolysis is a general consequence of plasma membrane rupture during necrotic cell death. In contrast, individual treatment with AEBSF, Aprotinin, E64, Bestatin or Pepstatin A had no effect on necrotic cleavage of lamin-B1 (Fig. 2b–d; Supplementary Fig. S2b, c).
We also performed time course analysis of necrotic cleavage of lamin-B1 in HT22 cells undergoing necroptosis and ferroptosis, as well as in BMDMs undergoing necroptosis and pyroptosis (Fig. 2e, f). A time-dependent increase in lamin-B1 necrotic cleavage was observed, which was effectively suppressed by treatment with leupeptin (Fig. 2e, f). Treatment with leupeptin alone or with the triple-combination of leupeptin, AEBSF, and Aprotinin had no effect on the progression of necroptosis, ferroptosis and pyroptosis (Supplementary Fig. S2d–f), suggesting that the proteases targeted by leupeptin in necrotic lamin-B1 cleavage do not mediate cell death.
Since leupeptin is known to inhibit a variety of proteases, including cathepsin B, calpain, trypsin, and plasmin
29, we next examined whether lysosomal cathepsins might be involved in mediating necrotic cleavage of lamin-B1. However, we found that treatment with cathepsin inhibitor Cathepsin inhibitor 1, Z-FY-CHO, LY2811376, lysosome inhibitor NH
4Cl, chloroquine, Bafilomycin A1, or lysosome protease inhibitor E64d could not inhibit necrotic cleavage of lamin-B1, suggesting that lysosomal protease is not involved in mediating necrotic lamin-B1 cleavage (Fig. 2a; Supplementary Fig. S2g).
Necrotic protein cleavage occurs after membrane rupture in necrosis
To define the timing of necrotic lamin-B1 cleavage, we next used SYTOX Green to mark and select cells that had already lost the integrity of cytoplasmic membrane by flow cytometry. In HT29 cells treated with TSZ to induce necroptosis, we detected necrotic lamin-B1 and lamin-A/C cleavage and p-MLKL in SYTOX Green positive (ST+) cells. In contrast, SYTOX Green negative (ST–) cells also had p-MLKL but not necrotic cleavage of lamin-B1 or lamin-A/C, suggesting that the cleavage of lamin-B1 and lamin-A/C occurs after the disruption of cytoplasmic membrane integrity (Fig. 2g; Supplementary Fig. S3a). Furthermore, p-MLKL alone may not be sufficient to disrupt the integrity of cytoplasmic membrane. Similarly, in cells treated with RSL3 to induce ferroptosis, necrotic lamin-B1 cleavage was only found in ST+ cells, but not in ST– cells (Fig. 2h; Supplementary Fig. S3b). In BMDMs treated with LPS/zVAD to induce necroptosis, necrotic lamin-B1 cleavage was only detected in ST+ cells, whereas p-MLKL was detected in both ST+ and ST– cells (Fig. 2i; Supplementary Fig. S3c). In cells treated with LPS/Nigericin, both apoptotic and necrotic cleavages of lamin-B1 were found in ST+ cells, while only apoptotic lamin-B1 cleavage was found in ST– cells (Fig. 2i; Supplementary Fig. S3d). Thus, apoptotic lamin-B1 cleavage can occur before cytoplasmic membrane disruption, while necrotic lamin-B1 cleavage occurs after the loss of cytoplasmic membrane integrity.
Necrotic cleavage by extracellular proteases
Since the necrotic lamin-B1 cleavage occurs after disruption of cytoplasmic membrane, we considered the possibility that the cleavage activity might come from extracellular sources. With cultured cells, we considered the contribution of proteases in serum or proteases released by cells to necrotic lamin-B1 and lamin-A/C cleavage. We found that the necrotic cleavage of lamin-B1 was reduced by exchanging the culture medium to Dulbecco’s modified Eagle medium (DMEM) without fetal bovine serum (FBS) before adding TSZ, which did not affect necroptosis as the levels of p-MLKL were not affected (Fig. 2j; Supplementary Fig. S3e). However, BMDMs treated with conditioned medium collected from HT29 cells or BMDMs after having been cultured in serum-free condition for 12 h also exhibited necrotic lamin-B1 cleavage upon TSZ treatment, which was inhibited by leupeptin (Fig. 2j). These results suggest a contribution of FBS-derived proteases to necrotic cleavage, while the remaining cleavage activity in FBS-free conditions implies that cells are capable of producing and releasing factors that mediate necrotic cleavage in a serum-independent manner.
To further characterize the effect of serum on necrotic cleavage, we established a delayed protocol with the time point of medium replacement from 1 h before TNF-α treatment to 1.5 h after TNF-α treatment (Supplementary Fig. S3f). With this delayed protocol, the necrotic cleavage of lamin-B1 and lamin-A/C was still inhibited in necroptotic HeLa-RIPK3 cells cultured without FBS and by the addition of heat-inactivated FBS; and furthermore, the addition of FBS led to concentration-dependent increases in the necrotic cleavage of lamin-B1 and lamin-A/C in necroptotic HeLa-RIPK3 cells (Supplementary Fig. S3g–i). The different inhibitory efficiencies of FBS-free conditions observed in Fig. 2j and Supplementary Fig. S3g, i may in part result from the different time available for cells to produce and release factors that mediate necrotic cleavage, which is shorter in the delayed medium replacement protocol. Constant presence of leupeptin was required to inhibit necrotic lamin cleavage as necrotic lamin-B1 cleavage still occurred when cells were cultured in 10% FBS after removal of leupeptin following 6 h of incubation (Supplementary Fig. S3j). However, heat-inactivation (at 60 °C or 80 °C) of FBS was sufficient to block necrotic cleavage, consistent with the role of serum proteases in mediating necrotic cleavage (Supplementary Fig. S3j). In addition, we found that TSZ-induced necroptosis of HeLa-RIPK3 cells cultured in the mouse serum freshly isolated from adult mice led to stronger necrotic cleavage of lamin-B1 and lamin-A/C than that of the cells cultured in commercial FBS, suggesting the contribution of proteases from freshly isolated mouse serum (Supplementary Fig. S3f, k). These alterations in serum addition mentioned above did not affect necroptosis progression other than the cleavage of lamin-B1 and lamin-A/C (Supplementary Fig. S3l).
In summary, extracellular proteases, including proteases in serum and proteases expressed and secreted by cultured cells, may enter cells upon plasma membrane rupture to mediate necrotic protein cleavage. The presence of such necrotic protein cleavage in a mouse model of AKI (Fig. 1i; Supplementary Fig. S1g) further indicates that such extracellular proteases, responsible for cleaving proteins in necrotic cells, are present within the tissue microenvironment.
Neo-N-terminomic analysis of necrotic protein cleavages
Since the above results suggest that necrotic protein cleavage occurs after the loss of cytoplasmic membrane integrity, we reasoned that lamin-B1 and lamin-A/C were unlikely to be the only proteins cleaved in necrosis. Thus, we next performed protein neo-N-terminal analysis by mass spectrometry to identify additional cleaved protein targets in necrosis. We adapted dimethyl protein N-terminal labeling method
30, which was previously used to label protein N-termini
in vitro, to unbiasedly identify neo-N-terminal peptides produced after induction of necrosis in cell models and kidney samples. Total proteins from control and necrotic cells were first treated with formaldehyde to conduct dimethyl labeling for existing protein N-termini as well as those newly exposed after necroptosis induction before trypsin digestion. The unlabeled N-termini generated by trypsin digestion were then labeled by HPG-ALD (~100 kDa) to increase the sizes of these peptides to over 100 kDa, allowing their retention by a 10 kDa ultrafiltration device and subsequent removal. The resulting pool of dimethyl-labeled N-terminal peptides enriched by ultrafiltration was analyzed by mass spectrometry (Supplementary Fig. S4a).
We determined and compared dimethyl-labeled N-terminal peptides from HT29 cells under control conditions, necroptosis induced by TSZ, inhibition of necroptosis by NSA and apoptosis induced by TC. By principal component analysis (PCA), the profiles of cleavage peptides identified by neo-N-terminomics are highly distinct between control and necroptosis, control and apoptosis, as well as necroptosis and apoptosis (Supplementary Fig. S4b). In apoptosis of HT29 cells induced by TC, we identified 1,308 cleavage peptides from 949 protein targets that were apoptosis-specific, with 94 cleavage peptides identical to those found in a previous analysis for apoptotic cleavage products
31 (Supplementary Fig. S4c, d). The predominant apoptotic cleavage sites were after Asp residue (Supplementary Fig. S4c), consistent with the cleavage specificity of caspases
32, thus validating the method for identifying neo-N-termini generated by distinct cellular mechanisms. We also performed neo-N-terminomic analysis in BMDMs treated with TSZ that induces necroptosis. PCA also revealed highly distinct neo-N-terminal profiles among control, necroptosis and necroptosis with leupeptin (Supplementary Fig. S4e). To characterize necrotic protein cleavages
in vivo, we next performed neo-N-terminomic analysis on kidney samples derived from a mouse AKI model
in vivo. By PCA, the peptide profiles identified by neo-N-terminomic analysis were highly distinct between control and AKI mouse kidneys (Supplementary Fig. S4f).
We next compared the neo-N-termini identified in control and necroptotic samples in mass spectrometry analysis and defined those neo-N-termini only present in necroptotic cells as specific protein cleavages in necroptosis. From necroptotic HT29 cells induced by TSZ, we identified 1,004 neo-N-terminal peptides from 691 proteins, which could not be identified in cells treated with MLKL inhibitor NSA (Fig. 3a). The protein cleavage sites identified in necroptosis were mainly after Arg/Lys residues (Fig. 3a). From necroptotic BMDMs induced by TSZ in mouse serum, we identified 6,931 neo-N-terminal peptides from 2,882 proteins in which 4,724 cleavage sites were after Arg/Lys residues (Fig. 3b). Among these 6,931 peptides, 3,054 peptides from 1,929 proteins could not be identified after leupeptin treatment (Supplementary Fig. S4g). In mouse AKI model induced by ischemia-reperfusion, we identified 8,354 neo-N-terminal peptides from 2,584 proteins that were predominantly cleaved at lysine and arginine (Fig. 3c). We compared the cleaved peptides identified during necroptosis in BMDMs with those detected in the AKI model and found a shared set of 1,253 peptides derived from 772 proteins which were identified under both conditions (Supplementary Fig. S4h). The cleavage sites of these peptides were also predominantly located after arginine and lysine residues (Fig. 3d). Since the Arg/Lys residues are the preferred cleavage sites for trypsin-like proteases which can be inhibited by leupeptin, these results suggest that trypsin-like serum proteases might play an important role in mediating necrotic protein cleavages, which happens after loss of cytoplasmic membrane integrity. Also, the shared protein cleavage targets and sites in necroptotic cells and AKI model support the involvement of necrotic proteolytic cleavages in vivo under pathological conditions.
Next, we selectively confirmed a subset of the necrotic protein cleavages identified in neo-N-terminomic analysis by western blotting. We were able to confirm the necrotic cleavage products of ABHD16A, LAP2β, PREB, ATP13A1, PELP1 and Syntaxin-18 identified by neo-N-terminomic analysis in necroptotic HT29 cells; the generation of these products was inhibited by NSA and leupeptin (Fig. 3e; Supplementary Fig. S4i). In the case of TC-induced apoptosis of HT29 cells, we identified apoptotic cleavage products of lamin-B1, lamin-B2, LAP2β and Syntaxin-18 that were distinct from those of necrotic cleavage products (Fig. 3e). We were also able to confirm by western blotting the necrotic cleavages of ABHD16A, LAP2β, PREB, ATP13A1, PELP1 and Syntaxin-18 in ferroptotic HT1080 cells induced by RSL3 which were inhibited by Lip-1 and leupeptin, while apoptosis of HT1080 cells led to distinctive apoptotic cleavage of lamin-B1, lamin-B2 and LAP2β (Fig. 3f). In addition, the necrotic cleavages of lamin-B1, ABHD16A, LAP2β and PREB and their inhibition by leupeptin were confirmed by western blotting in BMDMs induced by TSZ to undergo necroptosis or LPS/Nigericin to undergo pyroptosis, while apoptotic cleavages of lamin-B1 and LAP2β were also induced by LPS/Nigericin (Fig. 3g). We were also able to verify a subset of the
in vivo cleavage events identified in neo-N-terminomic analysis of AKI model by western blotting, including the cleavage of ABHD16A, GLUD1, TUBA4A, PCCB, CYP4B1, ACAT1, PON1, ST7, and ZNRF2 (Fig. 3h; Supplementary Fig. S4i–k). We next used a hepatic ischemia-reperfusion injury (HIRI) model
33 to further verify necrotic cleavage of proteins. The necrotic cleavages of lamin-B1, ABHD16A, GLUD1 and TUBA4A were also found in liver after ischemia-reperfusion injury (Fig. 3i).
Similar to apoptotic cleavage of protein targets in diverse subcellular compartments
34, necroptosis in HT29 cells, necroptosis in BMDMs and necrosis in AKI model
in vivo were also characterized by widespread proteolytic processing of proteins localized to diverse cellular compartments (Supplementary Fig. S5a–c), including proteins involved in diverse functions such as protein transport, mRNA processing, translation, proteolysis, and chromatin remodeling (Supplementary Fig. S5d–f). Thus, our neo-N-terminal analysis by mass spectrometry uncovered a systematic proteolytic cleavage event in necrosis.
Necrotic cleavages mediated by Trypsin-like proteases
The neo-N-terminomic analysis described above suggests the involvement of proteases with preference for cleaving after Arg/Lys residues in necrosis. Thus, we next characterized proteases present in the culture medium of BMDMs, HEK293T cells and HT29 cells, FBS and mouse serum by mass spectrometry and identified 10 candidate proteases with Arg/Lys cleavage specificity. We established a complementation system to identify proteases that can complement the necrotic cleavage activity for necroptotic BMDMs (induced by TSZ) cultured in FBS-free condition which does not affect the progression of necroptosis (Supplementary Fig. S6a). Using necrotic cleavages of lamin-B1 and ABHD16A as biomarkers, among 10 candidate proteases tested, PRSS2 and plasminogen (PLG), both encoding zymogens of trypsin-like proteases, can promote necrotic protein cleavage of lamin-B1 (Supplementary Fig. S6b, c). PRSS1, PRSS2, PRSS3, TRY4 and TRY5 share about 90% sequence similarity, and the peptidase S1 domain of PLG shares about 54% sequence similarity with trypsin family (Supplementary Fig. S6d). We found that PRSS1 has the highest activity promoting necrotic protein cleavage in necroptosis (Supplementary Fig. S6e, f), while PRSS2, TRY4 and TRY5 have similar activity and PRSS3 has the lowest activity in ATP13A1 cleavage (Supplementary Fig. S6e, f). We were able to reproduce necrotic protein cleavage of GLUD1 observed in an
in vivo AKI model by applying conditioned medium from PRSS1-expressing HEK293T cells to BMDMs (Supplementary Fig. S6g). Consistent with the activation of trypsin-like proteolytic enzymes
in vivo, we detected the cleavage of not only substrates, such as ABHD16A and GLUD1, but also trypsin itself in mouse kidney ischemia-reperfusion samples by western blotting (Supplementary Fig. S6h). Furthermore, we identified a site known to be a marker of trypsin activation conserved among different PRSS family members
35 in the tissue lysates of AKI model by mass spectrometry (Supplementary Fig. S6i). Taken together, these data demonstrate the activation of extracellular trypsin-like proteases after tissue damage
in vivo and during necrosis of cultured cells, and that these proteases can mediate necrotic protein cleavages.
To further define the tissue distribution of these trypsin-like proteases, we examined the expression of PRSS1, PRSS2, PRSS3, TRY4, TRY5, and PLG across a panel of mouse tissues, using HEK293T cells as a reference (Supplementary Fig. S7a, b). PRSS1, PRSS2, PRSS3, TRY4, and TRY5 were highly expressed in the stomach, pancreas, and intestine; detectable in the liver, spleen, and lung; low in the kidney, brain, skeletal muscle, and testis; and undetectable in the heart (Supplementary Fig. S7a). In contrast, PLG was expressed in the liver, spleen, kidney, stomach, intestine, skeletal muscle, and testis, but was not detected in the heart, lung, or brain (Supplementary Fig. S7b). This tissue expression pattern is consistent with our observation of necrotic protein cleavage in injured kidney and liver.
To assess whether this phenomenon can be extended to human tissue samples, we first confirmed the presence of close human homologs of PRSS1, PRSS2, PRSS3, and PLG (Supplementary Fig. S7c). We then retrieved RNA expression profiles of
PRSS1,
PRSS2,
PRSS3, and
PLG across human tissues from the Human Protein Atlas
36,37 (Supplementary Fig. S7d–g).
PRSS1,
PRSS2, and
PRSS3 showed extremely high transcript levels in the pancreas, and were also expressed at relatively high levels — albeit to varying extents — across digestive tissues, including the colon, duodenum, rectum, salivary gland, small intestine, and stomach (Supplementary Fig. S7d–f). In addition,
PRSS1 and
PRSS2 exhibited moderate expression in adipose tissue, blood vessel, heart, muscle, and liver, whereas
PRSS3 showed higher expression in the amygdala, cerebellum, cerebral cortex, and skin (Supplementary Fig. S7d–f). Overall, beyond the pancreas and digestive organs,
PRSS1,
PRSS2 and
PRSS3 transcripts were detectable in multiple non-digestive tissues where trypsin-family proteases have not typically been considered functional (Supplementary Fig. S7d–f). In contrast, PLG displayed a more restricted pattern, with high transcript levels in the kidney and liver and little to no detectable expression in most other tissues (Supplementary Fig. S7g). Collectively, these analyses establish that human tissues express close homologs of the trypsin-like proteases implicated in our mouse and cell-based studies, supporting the feasibility of extending this phenomenon to human tissue samples. They also reveal substantial inter-tissue differences in the protease expression and composition, suggesting that the extent and specificity of necrotic protein cleavage — and potentially the associated autoimmune risk — may vary across tissues, an important area for future investigation.
Necrotic protein cleavage promotes nuclear damage in necrosis
We next characterized the contribution of necrotic protein cleavages to morphological and biochemical changes in necrosis. We found that in contrast to that of apoptotic BMDMs induced by TSG, which leads to classical DNA laddering
38, necroptosis of BMDMs and HT22 cells induced by TSZ and ferroptosis of HT22 cells induced by RSL3 led to the degradation of genomic DNA which was reduced by treatment with leupeptin (Fig. 4a, b). In contrast, leupeptin had no effect on apoptotic-like DNA fragmentation in BMDMs treated with LPS/Nigericin or apoptotic DNA fragmentation in BMDMs and HT22 cells treated with TSG (Fig. 4a, b). The ability of leupeptin and heat-inactivated FBS to reduce the necrotic loss of genomic DNA was also demonstrated by total genomic DNA quantification (Supplementary Fig. S8a).
Consistent with the protection against the loss of genomic DNA in necrosis, treatment with leupeptin can rescue the morphological changes of nuclear envelope and nucleolus in necroptosis of HT22 cells and BMDMs induced by TSZ and ferroptosis of HT22 cells induced by RSL3 (Fig. 4c, d). Fresh mouse serum exhibited stronger necrotic protein-cleavage activity than that of commercial FBS (Supplementary Figs. S3k, S8b). Given our previous finding that these cleavage events are mediated by trypsin-like proteases, we further compared leupeptin with several other trypsin-like protease inhibitors, including TLCK, soybean trypsin inhibitor, ulinastatin, benzamidine, 4-aminobenzamidine, patamostat, and camostat mesylate; however, none of these inhibitors performed better than leupeptin (Supplementary Fig. S8b). Consistently, compared with FBS, the addition of mouse serum led to more extensive disruption of nuclear envelope and nucleolar structures in necroptotic BMDMs, and this damage was prevented by leupeptin (Fig. 4d, e). These results further support that the observed nuclear envelope and nucleolar morphological changes are driven by protease-mediated cleavage. Treatment with leupeptin also led to inhibition of the lamin-B1 loss in necroptosis, ferroptosis and pyroptosis (Fig. 4f, g; Supplementary Fig. S8c, d). To examine the functional consequence of necrotic cleavage on the nuclear envelope, we characterized changes of fluorescent protein-tagged nuclear envelope proteins including lamin-B1, lamin-B receptor (LBR) which is an integral protein in the inner nuclear envelope, as well as Emerin, and nucleoporin Nup50
39,40, which were all identified as necrotic cleavage substrates in neo-N-terminomic analysis (Supplementary Fig. S8e). In BMDMs expressing mGFP-lamin-B1, mGFP-LBR, mGFP-Emerin or mGFP-Nup50 and treated with LPS/zVAD for necroptosis induction, we observed a loss of mGFP signals on nuclear envelope which was inhibited by the addition of leupeptin (Fig. 4h; Supplementary Fig. S8f–h). These results suggest that protease-mediated necrotic cleavage of nuclear envelope proteins contributes to nuclear envelope disruption during necrosis, and inhibition of these proteases preserves nuclear integrity.
Necrotic protein cleavage promotes phagocytosis of necrotic cells
Phagocytosis is important for the clearance of dead cells
41. To investigate whether necrotic protein cleavage might serve to facilitate phagocytosis of necrotic cells, we assessed whether leupeptin inhibits phagocytosis of necrotic cells by healthy BMDMs (Supplementary Fig. S9a). Necroptotic cells marked by nuclear Histone4-RFP and Hoechst were incubated with healthy BMDMs expressing GFP. We found that the numbers of Histone4-RFP
+ or the Hoechst-labeled necrotic nuclei in GFP
+ healthy BMDMs were reduced by leupeptin (Fig. 5a; Supplementary Fig. S9b). In contrast, phagocytosis of necroptotic cells labeled with ER-localized Sec61β-RFP, which marks the uptake of cytoplasm, was unaffected by leupeptin (Supplementary Fig. S9b). Thus, necrotic protein cleavages might be particularly important for the efficient engulfment of nuclear components during clearance of necroptotic cells.
To investigate the functional significance of necrotic protein cleavages
in vivo, we performed intraperitoneal injection of control PBS (i.p. PBS), necrotic BMDMs (i.p. BMDM/TSZ), necrotic BMDMs co-injected with leupeptin (i.p. BMDM/TSZ+200 mg/kg leupeptin), leupeptin-pretreated necrotic BMDMs co-injected with leupeptin (i.p. BMDM/(TSZ+leupeptin)+200 mg/kg leupeptin; Supplementary Fig. S9c). We found that intraperitoneal injection of necrotic BMDMs in which necrotic protein cleavage had been inhibited by leupeptin led to splenomegaly (Fig. 5b). Using an established autoimmunity assay
42, we found that the serum from mice injected with leupeptin-inhibited necrotic BMDMs combined with leupeptin exhibited the strongest autoimmunity signals compared to other groups (Fig. 5c; Supplementary Fig. S9d).
Using an established protocol to visualize phagocytosis of necrotic cells
in vivo43,44, we intraperitoneally injected mice with MEFs expressing Histone4-RFP and Sec61β-mBaojin
45 (MHS) that were treated with either TSZ, or TSZ/Leupeptin, with or without leupeptin co-administration, and collected peritoneal CD11b
+ myeloid cells at 4 h and 17 h post-injection (Supplementary Fig. S9e, f). Both the signals of Sec61β-mBaojin and Histone4-RFP in CD11b
+ myeloid cells were increased 4–7 h after injection of TSZ-induced necroptotic MEFs, indicating the efficient phagocytosis of necroptotic MEFs by myeloid cells
in vivo (Fig. 5d). Blocking necrotic cleavage
in vitro is important for reducing phagocytosis as treatment of necroptotic MHS cells with leupeptin
in vitro led to reduced phagocytosis, regardless of leupeptin injection
in vivo, as indicated by both Sec61β-mBaojin and Histone4-RFP levels in CD11b
+ myeloid cells 4 h post injection (purple vs green/red quantification dots in Fig. 5d); while systemic treatment of leupeptin alone with necroptotic MHS cells after in vitro necrotic cleavage had no effect (green vs red quantification dots in Fig. 5d). After a longer time point of 17 h, the systemic leupeptin injection led to increases in the signals of necroptotic Sec61β-mBaojin and Histone4-RFP-MEFs in CD11b
+ myeloid cells compared to that without leupeptin, suggesting that systemic leupeptin injection was able to reduce the degradation of necroptotic cells (red vs green quantification dots in Fig. 5e). Similar to results observed 4 h post injection, after 17 h, systemic leupeptin injection of leupeptin-treated necroptotic MHS cells led to reduced phagocytic uptake compared to systemic leupeptin injection of necroptotic MHS cells without
in vitro leupeptin treatment (purple vs green quantification dots in Fig. 5e). Collectively, these results suggest that necrotic protein cleavage
in vivo may facilitate phagocytic clearance of necrotic cells, reducing the exposure of intracellular epitopes that otherwise can trigger autoimmune response.
Development of monoclonal antibodies specifically recognizing necrotic protein cleavages
Our results described above suggest that in the context of necroptosis and pyroptosis, the detection of p-MLKL or cleaved GSDMD, respectively, does not necessarily indicate that disruption of cytoplasmic membrane, the final execution event of cell death, has occurred. Thus, we still lack reliable methods to specifically mark cells that have executed necrosis, especially in a tissue context in vivo. Since the necrotic protein cleavages identified in our study represent irreversible modifications that occur specifically after plasma membrane rupture, these cleavage events can serve as markers for identifying necrotic cells. We selected ABHD16A (cleaved after R178) and GLUD1 (cleaved after K397) for developing necrotic cleavage-specific monoclonal antibodies as biomarkers of necrosis (Supplementary Fig. S10a, b). Lamin-B1 was excluded because our neo-N-terminomic analysis did not detect a cleavage site consistent with the size of the necrotic lamin-B1 fragment observed by western blotting (Supplementary Fig. S4i). Both ABHD16A and GLUD1 are broadly expressed across multiple tissues (Supplementary Fig. S10c).
To generate cleavage-specific antibodies, we immunized rabbits with four peptides corresponding to the N- and C-terminal neoepitopes of ABHD16A cleaved after residue R178 and GLUD1 cleaved after residue K397, respectively. Monoclonal antibodies were first screened by western blotting for their ability to detect necrotic cleavage products of ABHD16A or GLUD1. In murine HT22 cells and BMDMs as well as in human HT1080 cells and HT29 cells with induction of necroptosis, ferroptosis, or pyroptosis, specific antibodies were able to recognize the N- and C-termini of cleaved ABHD16A, ABHD16A-N (20 kDa) and ABHD16A-C (42 kDa), which were blocked by the treatment with leupeptin (Fig. 6a–c). The cleavage of ABHD16A was consistently detected in samples undergoing ferroptosis induced by multiple ferroptosis inducers (ML162, ML210, FINO2, FIN56, and Erastin) and was inhibited by treatment with Liproxstatin-1 and leupeptin (Supplementary Fig. S10d, e). Moreover, ABHD16A-N and ABHD16A-C antibodies detected the corresponding cleavage products in vivo, including in kidneys from mice subjected to AKI and in livers from mice subjected to HIRI model (Fig. 6a–d).
The ABHD16A-N and ABHD16A-C products, whose generations were inhibited by leupeptin, were also detected in TSG-induced apoptosis of HT22 cells and BMDMs, likely reflecting secondary necrotic protein cleavage following the loss of cytoplasmic membrane integrity at late apoptotic stages (Fig. 6a, b). Consistently, time-course analysis revealed that the generation of these ABHD16A cleavage products became detectable only at late apoptotic stages induced by TSG, appearing 6–8 h after apoptosis induction (Supplementary Fig. S10f).
The antibody functionality in recognizing necrotic cleavages was further assessed by IF and immunohistochemistry (IHC) staining. By IF, the ABHD16A-C antibody produced strong signals in necroptotic BMDMs labeled with propidium iodide (PI), a robust marker of cell death (Supplementary Fig. S11a–d). Treatment with leupeptin or knockdown of ABHD16A markedly reduced the IF signals of cleaved ABHD16A in PI+ cells during necroptosis (Supplementary Fig. S11a–c), confirming the specificity of the antibody. Compared with PI, the cleaved ABHD16A antibody labeled approximately 70% (147/209) of PI+ cells during necroptosis, which was effectively reduced by treatment with leupeptin (Supplementary Fig. S11b). Similarly, the ABHD16A-C antibody labeled approximately 72% (54/74) PI+ cells during pyroptosis (Supplementary Fig. S11d). However, we also detected a subset of cleaved caspase-3-positive cells that showed little, if any, immunostaining with ABHD16A-C antibody (Supplementary Fig. S11d), suggesting that these LPS/Nigericin treated cells are undergoing apoptosis and have not yet lost the integrity of their cytoplasmic membrane. These results indicate that the ABHD16A-C antibody by IF can effectively identify the dying cells that have lost plasma membrane integrity.
For GLUD1, a C-terminal GLUD1 antibody detected the cleaved GLUD1 product (20 kDa) in necroptotic BMDMs cultured with mouse serum, as well as in kidneys and livers from mice subjected to
in vivo models AKI and HIRI, respectively (Fig. 6c, d; Supplementary Fig. S10g). Using IF, the GLUD1-C antibody detected dying necroptotic BMDMs cultured in the presence of mouse serum (Supplementary Fig. S11e). Moreover, both ABHD16A-C antibody and GLUD1-C antibody detected dying cells by IF staining in the mouse kidneys following AKI and in the mouse intestine with TAK1-ablation-induced ileitis
46 (Fig. 6e; Supplementary Fig. S11f). Both antibodies were also able to effectively mark necrotic cells in AKI kidney and HIRI liver by IHC (Fig. 6f, g). TAK1-ablation-induced ileitis could be rescued by RIPK1-D138N kinase-dead knockin mutation in combination with TRADD-IEC knockout, and correspondingly, the abundance of necrotic cells detected by IHC using ABHD16A-C and GLUD1-C antibodies was also reduced following combined RIPK1-D138N knockin and TRADD-IEC knockout (Fig. 6h). Finally, application of these antibodies to human diabetic nephropathy (DN) samples stained necrotic renal tubular epithelial cells but no signal could be detected when these antibodies were applied to control samples, indicating that DN progression is associated with increased necrosis (Fig. 6i). Collectively, these data demonstrate that we have successfully developed necrosis cleavage-specific antibodies that reliably detect necrotic cells across multiple experimental systems and species, providing valuable tools for assessing the terminal events in necrosis and for investigating the contribution of necrosis to disease pathogenesis and physiological processes.
DISCUSSION
Our study described above demonstrated widespread proteolytic cleavage of intracellular proteins mediated by leupeptin-inhibitable extracellular proteases after the loss of cytoplasmic membrane integrity in diverse models of necrosis, including necroptosis, ferroptosis, a subset of pyroptosis and secondary necrosis after apoptosis (Fig. 7). Thus, regardless of how membrane rupture occurs during primary necrosis or during late apoptotic progression to secondary necrosis, breach of the plasma membrane barrier is sufficient to allow extracellular proteases, particularly trypsin-like proteases, to enter cell corpses and drive intracellular protein cleavage. Such trypsin-like proteolytic activity drives characteristic morphological alterations in nuclear envelopes and nucleoli during necrotic cell death modalities such as necroptosis, ferroptosis, and pyroptosis. Our neo-N-terminal analysis for necrosis identified a substantial list of necrotic cleavage substrates in diverse cellular compartments, including cytoplasm, membrane, Golgi, ER and particularly nucleus. Thus, extracellular proteases such as trypsin-like proteases perform functions analogous to caspase-mediated protein cleavage during apoptosis for the degradation of necrotic cells.
The clearance of apoptotic cells, known as efferocytosis, is mediated by the caspase-mediated cleavage of phosphatidylserine (PtdSer) flippases, leading to the exposure of PtdSer on cell surface to act as an 'eat me' signal to attract macrophages
47. Defects in efferocytosis lead to autoimmunity
2. Similarly, our findings indicate that proteolytic processing of necrotic cells also contributes to their efficient clearance and may play an important role in maintaining tissue homeostasis by limiting autoimmune responses. Our data suggest that phagocytosis of necrotic cells may proceed in at least two separable phases. Engulfment of cytoplasmic material is leupeptin-insensitive (Supplementary Fig. S9b), whereas engulfment of nuclear components is inhibited by this protease inhibitor (Fig. 5a). Because necrotic cells lose plasma membrane integrity and therefore, may not be cleared as intact 'whole-cell' corpses as those in apoptosis; instead, stepwise removal of cellular contents may be required for the clearance of necrotic cell corpses. We therefore speculate that necrosis-associated proteolysis may generate nuclear neo-epitopes that function as 'eat-me' signals to be recognized by phagocytic receptors — a possibility that warrants future investigation.
Notably, direct interrogation of necrotic proteolysis in vivo is constrained by both pharmacological and biological limitations. Systemic inhibition using the broad-spectrum protease inhibitor leupeptin is restricted by a dose dependent toxicity, as concentrations sufficient to suppress tissue protease activity are lethal, whereas lower doses are ineffective. In addition, we identify multiple proteases capable of generating necrosis-associated cleavage (including trypsin-like proteases and PLG), yet ~20–35% of N-terminomic cleavage sites are non-K/R, implying additional protease specificities operating in parallel (Fig. 3a–c). Thus, necrotic proteolysis appears to be driven by a broad, partially redundant extracellular protease milieu rather than a single dominant enzyme, which limits the interpretability of single-gene loss-of-function approaches in vivo. Accordingly, genetic dissection would likely require combinatorial perturbations across many candidate extracellular proteases, which is currently impractical. Consequently, we employed a chemical approach in which leupeptin-treated necrotic cells were injected into mice to probe the functional consequences of impaired necrotic cleavage. While this strategy reveals a potential role for necrotic proteolysis in promoting necrotic cell clearance and preventing autoimmunity, further studies will be required to define the full spectrum of functions and the physiological relevance of necrotic proteolysis in vivo.
We also demonstrated necrotic cleavages of ABHD16A and GLUD1 in cell culture system, kidney samples after AKI, hepatic samples after ischemia liver damage, rodent models of intestinal inflammation as well as human diabetic nephropathy samples. These observations suggest that a subset of necrosis-associated proteolytic events is conserved across diverse cellular and tissue contexts despite distinct microenvironments. Although the precise cleavage site in lamin-B1 was not identified, its consistent processing further supports its inclusion among conserved necrotic substrates. Notably, the human samples analyzed in this study are limited in size due to restricted availability of well-annotated biopsy specimens that meet the requirements for diagnosis, appropriate controls and standardized handling. We therefore regard the human diabetic nephropathy data as an initial proof-of-concept demonstrating that necrosis-associated cleavage signatures identified in mouse and cell models can also be detected in human disease tissue. Larger, independent cohorts and broader sampling across necrosis-associated diseases will be necessary to further establish the generalizability and translational relevance of these biomarkers. Because our mechanistic analyses largely rely on in vitro death models, the kinetics and substrate selectivity of membrane rupture-permissive extracellular proteolysis may differ in multifactorial, ‘multi-hit’ disease contexts, as suggested by differences between our cultured-cell and mouse kidney N-terminomic profiles. Consistent with this notion, necrotic protein cleavages may exhibit substantial inter-tissue heterogeneity due to differential protein expression profiles, distinct tissue-resident protease repertoires and different local inhibitory environments. Such tissue-specific protease landscapes could, in principle, influence disease susceptibility by modulating the extent and specificity of necrotic proteolysis.
We show that the activation of pyroptosis may lead to cell death by either apoptosis or necrosis. The rapid execution of apoptotic cell death is mediated by the sequential activation of caspases in a cascade manner
48; while pyroptosis is activated after the caspase-mediated cleavage of gasdermin D which releases its N-terminal pore-forming domain from the C-terminal repressor domain to allow the formation of pores on cytoplasmic membrane
49. In addition, membrane repair mechanisms can rescue cells from necrotic demise following limited membrane damage
50,51. Thus, whether an individual cell dies by apoptosis or necrosis may depend on the expression levels of caspases and gasdermins and membrane repair mechanism, i.e., whether the activation of caspase cascade reaches the end point for apoptotic demise before sufficient numbers of necrotic pores can be formed to promote necrosis. Our findings also suggest that the execution modes of cell death
in vivo must be interrogated on single cell basis in a complex tissue environment.
Our findings provide a conceptual framework for understanding necrotic proteolysis as a membrane rupture-enabled process with broad relevance across pathological settings. The interaction of dying cells with their tissue surrounding milieu might be important not only for how they die, but also for how they are cleaned up. Regardless of microenvironmental heterogeneity or the complexity of upstream death triggers, membrane rupture is expected to enable this class of proteolytic events, although both activity and substrate selectivity may vary across tissues and disease conditions. Importantly, the necrotic cleavage-specific antibodies against ABHD16A and GLUD1 generated by this study will provide tools to further explore the involvement of necrotic cleavages in human diseases.
MATERIALS AND METHODS
Mice
WT C57BL/6 mice aged 8 weeks, 12 months and 20 months were purchased from Ling Chang Biology (Shanghai, China). All mice were maintained in a specific pathogen-free environment and housed with no more than five animals per cage under controlled light (12 h light and 12 h dark cycle), temperature (25 ± 2 °C) and humidity (50 ± 10%) conditions and provided ad libitum access to food and water throughout all experiments.
Cell lines and cell cultures
HT-29 cells were cultured in McCoy’s 5A medium (Gibco, Grand Island, NY, USA), supplemented with 10% (vol/vol) FBS (Gibco) and 100 units/mL penicillin/streptomycin. HT1080 cells, HT22 cells, BMDMs, HeLa cells, HEK293T cells (ATCC) and MEFs were cultured in DMEM (Gibco) with 10% (vol/vol) FBS (Gibco) and 100 units/mL penicillin/streptomycin. For experiments under different serum conditions, cells were switched to medium containing the specified concentration and type of serum (e.g., mouse serum or serum-free) prior to the experiment. Serum heat-inactivation at 53 °C, 60 °C, 80 °C, and 100 °C was performed by incubating the serum in a metal bath for 20 min (100 °C) or 2 h (53 °C, 60 °C, 80 °C). Cells were cultured at 37 °C in a humidified atmosphere with 5% CO2.
Induction of cell death
Unless otherwise indicated, the concentrations of compounds used to induce various forms of cell death across different cell types are as specified below. BMDMs: 100 ng/mL LPS (MedChemExpress (MCE), HY-D1056), 15 μM Nigericin (MCE, HY-100381), 100 ng/mL human-TNFα (Novoprotein, C008), 200 nM SM164 (custom synthesized), 100 μM zVAD (MCE, HY-16658), 20 μM GSK872 (MCE, HY-101872), 100 nM 5z7 (Sigma, O9890), 10 μM Nec-1s (custom synthesized), 100 μM leupeptin (MCE, HY-18234A). HT22 cells: 100 ng/mL human-TNFα, 100 nM SM164, 60 μM zVAD, 10 μM GSK872, 1 μM RSL3 (MCE, HY-100218A), 200 nM Liproxstatin-1 (MCE, HY-12726), 10 μM Nec-1s, 100 μM leupeptin. HeLa cells expressing HA-RIPK3: 50 ng/mL human-TNFα, 100 nM SM164, 80 μM zVAD, 100 μM leupeptin. HT29 cells: 50 ng/mL human-TNFα, 100 nM SM164, 80 μM zVAD, 10 μg/mL CHX (Sigma, C7698), 100 μM leupeptin. HT1080 cells: 50 ng/mL human-TNFα, 100 nM SM164, 80 μM zVAD, 10 μg/mL CHX, 1 μM RSL3, 200 nM Liproxstatin-1, 100 μM leupeptin. MEFs expressing Histone4-RFP and Sec61β-mBaojin: 100 ng/mL human-TNFα, 100 nM SM164, 60 μM zVAD, 100 μM leupeptin.
For treatments requiring replacement of serum conditions, unless otherwise specified, the culture medium was replaced with the medium containing specified serum concentration concurrently with the application of pretreatment compounds. For TNFα/SM164/zVAD (TSZ) treatment: SM164 and zVAD were added 1 h prior. For TNFα/SM164/GSK872 (TSG) treatment: SM164 and GSK872 were added 1 h prior. For LPS/Nigericin treatment: LPS was added 4 h before Nigericin. For TNFα/CHX (TC) treatment: CHX was added 1 h prior. For LPS/5z7 or LPS/zVAD treatment: 5z7 or zVAD was added 1 h prior. Nec-1s, GSK872, and Liproxstatin-1 were added 30 min before induction of cell death. Leupeptin was added 30 min after the initiation of cell death with TNFα, RSL3, or Nigericin.
Construction of plasmids and stable cell lines
pMSCV-blasticidin (BSD), pMSCV-puromycin (PURO) or pLenti-blasticidin were used as the plasmid backbone in our study. Indicated mouse genes were amplified using Q5® High-Fidelity DNA Polymerase (NEB, M0491L) from a mouse tissue cDNA library and cloned into the indicated vector using ClonExpress® II One Step Cloning Kit (Vazyme, C112). All plasmids were verified by DNA sequencing.
The following plasmids were made and used in this study: pMSCV-PURO-mouse Histone4-linker-RFP, pMSCV-PURO-mouse Sec61β-linker-RFP, pMSCV-PURO-HA-human Ripk3, pMSCV-PURO-mGFP, pMSCV-BSD-mGFP-linker-mouse Lamin-B1, pMSCV-BSD-mGFP-linker-mouse Nup50, pMSCV-BSD-mGFP-linker-mouse Emerin, pMSCV-BSD-mGFP-linker-mouse LAP2β, pLenti-BSD-mouse PRSS1-3×Flag, pLenti-BSD-mouse PRSS2-3×Flag, pLenti-BSD-mouse PRSS3-3×Flag, pLenti-BSD-mouse TRY4-3×Flag, pLenti-BSD-mouse TRY5-3×Flag, pLenti-BSD-mouse PRSS23-3×Flag, pLenti-BSD-mouse PRSS29-3×Flag, pLenti-BSD-mouse PLG-3×Flag, pLenti-BSD-mouse Adam10-3×Flag, pLenti-BSD-mouse C1rb-3×Flag, pLenti-BSD-mouse HTRA3-3×Flag, pLenti-BSD-mouse HTRA1-3×Flag, pLenti-BSD-mouse F12-3×Flag, pLenti-BSD-mouse F11-3×Flag, pLenti-BSD-mouse F10-3×Flag, pLenti-BSD-mouse F9-3×Flag, pLenti-BSD-mouse F7-3×Flag, pLenti-BSD-mouse F2-3×Flag, pLenti-BSD-mouse CTSD-3×Flag, pLenti-BSD-mouse CTSB-3×Flag, pLenti-BSD-mouse CTSE-3×Flag, pLenti-BSD-mouse CTSS-3×Flag, pLenti-BSD-mouse Masp1-3×Flag, pLenti-BSD-mouse Masp2-3×Flag.
The pMSCV plasmid was co-transfected with the two packaging plasmids, VSVG and GAG, into HEK293T cells. Similarly, the pLenti plasmid was co-transfected with the two packaging plasmids, psPAX2 and pMD2G, into HEK293T cells. The supernatant was collected 48 h post-transfection, filtered through a 0.45 μm membrane, and used for infection. Following viral infection, stable cell lines were selected using blasticidin (Invitrogen, R21001) or puromycin (Sigma, S7417) antibiotics according to the resistance conferred by the respective plasmids.
Sample collection and western blotting analysis
Cell or tissue sample collection
For collecting viable cell samples, the supernatant of culture medium was aspirated and discarded. For non-viable cell samples, the supernatant was collected into a 1.5 mL EP tube or 15 mL centrifuge tube on ice, followed by centrifugation at 1,000× g for 2–3 min to pellet cell debris. After discarding the supernatant, the cell pellet was lysed with lysis buffer. The remaining adherent cells in the culture dish were lysed, and the lysates were combined into a 1.5 mL EP tube for further processing. All cell samples were lysed with 4% SDS cell lysis buffer (100 mM Tris-HCl, pH 8.6, 4% (w/v) SDS, 20% (v/v) glycerol). All tissue samples were minced with fine scissors on ice and were lysed at 4 °C in RIPA buffer (10 mM Tris pH 8.0, 140 mM NaCl, 1% TritonX-100, 1 mM EDTA, 0.5 mM EGTA, 0.1% sodium deoxycholate, 0.1% SDS) with phosphatase inhibitors (2 mM Na3VO4, 2 mM NaF) and Cocktail protease inhibitors (1:100, Bimake, B14001) with 3–5 grinding beads at 4 °C by tissue homogenizer first, followed by the addition of 10% SDS lysis buffer (250 mM Tris-HCl, pH 8.6, 10% (w/v) SDS, 50% (v/v) glycerol) to a final concentration of 4% SDS. DNA was fragmented either by sonication at 4 °C or boiling for 10 min, depending on whether membrane proteins were being analyzed. Protein concentration was determined using a BCA assay (Thermo Fisher Scientific, 23225).
Western blotting analysis
Primary antibodies against the following proteins were used in western blotting: Lamin-B1 (66095-1-Ig), Lamin-B2 (10895-1-AP), Lamin-A/C (10298-1-AP), PREB (10146-2-AP), Tmpo (67157-1-Ig), ACIN1 (23937-1-AP), ATP13A (16244-1-AP), Syntaxin-18 (16013-1-AP), PELP1 (30135-1-AP), ZNRF2 (20200-1-AP), PON1 (18155-1-AP), GLUD1 (14299-1-AP), CYP4B1 (11771-1-AP), SRXN1 (14273-1-AP), PCCB (11139-1-AP), and PLG (17462-1-AP) were from Proteintech. ACAT1 (44276), Cleaved Caspase-3 (Asp175) (9661S), Phospho-MLKL (Ser345) (37333S), Phospho-MLKL (Ser358) (91689S), Caspase-11 (14340), and GSDMD (39754) were from Cell Signaling Technology. Flag (F7425) was from Sigma. ST7 (PA5-95405), TUBA4A (MA5-32738), ATP11A (PA5-20995), Trypsin Pan (PA5-46939) and Trypsin (PA5-106876) were from Invitrogen. ABHD16A/BAT5 (ab185549), Trypsin (ab200997), Trypsin (ab211491) were from Abcam. Caspase-1 (sc-56036) was from Santa Cruz Biotechnology. β-Actin (I10813) was from TransGen. The anti-MLKL antibody was homemade. Antibodies for cleaved ABHD16A-N, cleaved ABHD16A-C and cleaved Glud1-C were customly-made by Biolynx (HangZhou, China), the peptide sequences used to generate cleaved ABHD16A-C and cleaved GLUD1-C were GVALLRPEPLHRGC and AKIIAEGANGPTTPEADKc, respectively. Secondary antibodies, including Goat anti-Rabbit IgG (H+L) Cross-Adsorbed Secondary Antibody, HRP (Invitrogen, G-21234), Goat anti-Mouse IgG, IgM (H+L) Secondary Antibody, HRP (Invitrogen, 31444), Goat anti-Rat IgG (H+L) Secondary Antibody, HRP (Invitrogen, A18865), Donkey anti-Sheep IgG (H+L) Secondary Antibody, HRP (Invitrogen, A16041), were used. The signals were detected by Immobilon ECL Ultra Western HRP Substrate (Millipore, WBKLS0500). If needed, the membranes were re-probed after being quenched with buffer containing 0.02–0.05% NaN3 to inactivate previous secondary antibodies. All non-commercial antibodies utilized in this paper are available from the lead contact upon request with proper MTA agreements.
Live-cell imaging with super resolution ODT
HT22 cells or BMDMs were seeded into 35 mm dishes with glass bottoms (Cellvis, D35-20-1.5H). Prior to seeding, the glass-bottom dishes were coated with Celltak (Corning, 354240) and incubated at room temperature for 30 min. Cells were seeded only in the central glass-bottom region. During seeding, a small bubble was gently introduced to create a clear area for acquiring background reference images during ODT imaging. The culture medium was replaced with phenol red-free DMEM (Gibco, 21063029) with the indicated compound treatments, and the volume was adjusted to 3 mL. Immediately before ODT imaging, the final treatment compound (e.g., TNFα, RSL3, or Nigericin) was added, and the total volume was brought to 5 mL. After compound treatment, the dish was securely fixed on the microscope stage. The upper objective (immersed in water; care was taken to avoid bubbles) and the lower objective were adjusted to locate the cell plane along the z-axis. The laser was focused at the center of the field, and the system was checked for bubbles. A reference image was acquired either by locating a cell-free region or by raising the z-axis by approximately 40 μm. Five regions of interest were selected and their positions recorded. Time-lapse imaging was recorded at 5 min intervals for 12 h using live cell panoramic super-resolution microscopy (Pellucid Optics Technology (Nantong) Co., LTD), operated with MH-PanoView software. ODT image analyses were performed using Intelligent Data Segmentation and Tracking Software (IntellySeg) and Image J version 1.53u. The refractive index (RI) of the nucleolus and nuclear envelope was quantified from ODT images acquired before and after the induction of cell death pathways, including necroptosis, ferroptosis, and pyroptosis. For each condition, multiple cells within the same field of view were analyzed. The nucleolus and nuclear envelope regions were manually delineated on the reconstructed tomographic images to ensure accurate targeting of these subcellular structures. The average refractive index values of these manually segmented regions were computed across the cell population. The sensitivity and resolution of ODT allowed reliable detection of these subtle but significant RI differences corresponding to structural damage in the nuclear envelope and nucleolus during early-stage necrosis. The percentages of cell death were verified by standard cell death quantification protocol using SYTOX Green.
Cell IF staining
Cells were seeded onto coverslips placed in 24-well plates. After specified compound treatments, the culture medium was aspirated and discarded. For samples containing dead cells, 24-well plates were centrifuged at 800× g for 3 min before culture medium was aspirated and discarded. Cells were fixed with 4% paraformaldehyde (PFA) at room temperature or pre-cooled methanol at 4 °C for 15 min, followed by three PBS washes. Specifically, pre-cooled methanol was applied for fixation when immunostaining for cleaved ABHD16A-C and cleaved GLUD1-C. Cells were permeabilized with 0.2% PBST if fixed by PFA (PBS containing 0.2% Triton X-100) for 15–30 min. For nuclear protein staining or high-density cell cultures, the permeabilization time was extended. Blocking was performed using PBS containing 5% heat-inactivated (60 °C) goat serum and 2% BSA for 1 h. Primary antibody (diluted in PBS containing 5% goat serum and 2% BSA) was applied at 20 μL per coverslip. Coverslips were inverted onto the lid of a 24-well plate and incubated in a humidified chamber at 4 °C overnight. After three PBS washes, coverslips were incubated with secondary antibody (diluted 1:1,000 in PBS containing 2% goat serum and 1% BSA) at room temperature for 1–1.5 h in the dark. Coverslips were washed three times with 0.05% PBST (PBS containing 0.05% Tween-20) for 20 min each. Nuclei were stained with 1 μg/mL Hoechst (Invitrogen, H21492) solution for 10 min at room temperature, followed by three PBS washes (5 min each). Coverslips were mounted using ProLong antifade mountant (Invitrogen, P36984). Images were acquired using a Leica SP8 confocal microscope, LightSheet system or OLYMPUS FV4000 with a 100× oil immersion objective. Primary antibodies, including Lamin-B1 (Proteintech, 66095-1-Ig), Nup98 (CST, 2598T), Cleaved Caspase-3 (Asp175) (CST, 9661S), cleaved ABHD16A-C and cleaved GLUD1-C (Biolynx), were used. Secondary antibodies, including Goat anti-Rabbit IgG (H+L) Cross-Adsorbed Secondary Antibody Alexa Fluor® 568 (Invitrogen, A11011), Goat anti-Mouse IgG (H+L) Secondary Antibody Alexa Fluor® 488 (Invitrogen, A11001), were used.
Tissue IF and IHC staining
IF staining
Following cardiac perfusion with cold PBS, mouse tissues were collected and fixed with either 4% PFA or pre-chilled methanol, then embedded using Neg-50 frozen section medium. Intestinal tissue sections were cut at a thickness of 15 μm, and kidney sections at 10 μm. Sections stored at −80 °C were placed at room temperature for at least 1 h. These sections were then washed three times with 0.025% PBST (containing Tween-20), followed by permeabilization with 0.2% PBST (containing Triton X-100) for 15 min. After blocking with 5% heat-inactivated (60 °C) goat serum and 2% BSA for 1 h, the sections were incubated with primary antibodies overnight at 4 °C. Following three washes with 0.025% PBST (Tween-20), the sections were incubated with secondary antibodies at room temperature for 80 min. After another three washes with 0.025% PBST (Tween-20), the nuclei were stained with 1 μg/mL Hoechst for 15 min at room temperature. The sections were then washed three times with PBS and mounted with ProLong® antifade reagent (Invitrogen, P36984). Secondary antibodies, included Goat anti-Rabbit IgG (H+L) Cross-Adsorbed Secondary Antibody Alexa Fluor® 568 (Invitrogen, A11011), Goat anti-Mouse IgG (H+L) Secondary Antibody Alexa Fluor® 488 (Invitrogen, A11001), were used.
IHC staining
Following cardiac perfusion with cold PBS, mouse tissues were isolated and prepared for IHC by fixing with 4% PFA at 4 °C for 24 h, followed by dehydration and paraffin embedding. Kidney tissue sections were cut at a thickness of 5 μm. Sections were baked at 60 °C for 1–2 h and left at room temperature overnight, then deparaffinized and rehydrated. Antigen retrieval was performed by boiling in 10 mM Tris, 1 mM EDTA, 0.02% Tween-20 (pH 9.0) buffer. Endogenous peroxidase activity was quenched with 3% hydrogen peroxide. After blocking with 5% heat-inactivated (60 °C) goat serum for 1.5 h at room temperature, the sections were incubated with primary antibodies overnight at 4 °C. Following three washes with 0.02% PBST (Tween-20), the sections were incubated with anti-rabbit secondary antibody (Vector Laboratories, BA-1000, 1:500) for 60 min at room temperature. Sections were subsequently developed with DAB substrate (Vector Laboratories, SK-4105). The development time was optimized using positive control samples and applied uniformly to all remaining sections. Sections were then counterstained with hematoxylin, dehydrated, and mounted with neutral balsam (Solarbio, G8590).
Primary antibodies that against lamin-B1 (Proteintech, 66095-1-Ig), cleaved ABHD16A-C and cleaved GLUD1-C (Biolynx), were used.
Screening for protease inhibitors that block necrotic protein cleavage
The following inhibitors were used: Leupeptin hemisulfate (HY-18234A), Pepstatin (HY-P0018), Bestatin (HY-B0134), Aprotinin (HY-P0017), AEBSF hydrochloride (HY-12821), Chloroquine (HY-17589A), Aloxistatin (E64d) (HY-100229), FMK 9a (HY-100522), Ilomastat (HY-15768), Bardoxolone methyl (HY-13324), E-64 (HY-15282), Cathepsin inhibitor 1 (HY-100231), LY2811376 (HY-10472), Z-FY-CHO (HY-128140), Calpeptin (HY-100223), and Bafilomycin A1 (HY-100558), TLCK (HY-112716), Trypsin inhibitor soybean (HY-126388), Ulinastatin (HY-134616), Benzamidine hydrochloride hydrate (HY-W087937), 4-Aminobenzamidine dihydrochloride (HY-W018723), Patamostat (HY-114080), Camostat mesylate (HY-13512) were from MCE. NEM (E3876) was from Sigma. MG132 (S2619) was from Selleckchem. Suc-AAPF-pNA (NJP33415) was from NJPeptide (NanJing, China). Cocktail (B14001) was from Bitool. The working concentrations of these compounds are shown in Supplementary Figs. S2a, S8b.
Cell death detection using SYTOX Green
Cells were seeded in black, light-impermeable 384-well plates or in black, light-impermeable 24-well plates with glass bottoms (Cellvis, P24-1.5H-N). SYTOX Green nucleic acid stain (Invitrogen, S7020) was added simultaneously with the final treatment at a final concentration of 1 μM. For 24-well plate experiments involving TSZ treatment, the medium was replaced after 1.25 h, followed by the addition of SYTOX Green and 1 μg/mL Hoechst. For 384-well plates, SYTOX Green signals were measured every 1–3 h using a microplate reader. For 24-well plates, SYTOX Green-positive signals were monitored every 30 min using a live cell High-Content Screening (HCS). For 384-well plates: cells treated with 0.1–0.3% Triton X-100 served as positive controls, and untreated cells served as negative controls. The relative cell death ratio was calculated as: (experimental group signal – negative control signal) / (positive control signal – negative control signal). The percentage of dead cells was determined by dividing the number of SYTOX Green-positive cells by the total number of Hoechst-labeled cells. All experiments were performed with 3–4 biological replicates per condition for 384-well plates and 3 biological replicates per condition for 24-well plates.
Mouse models of unilateral AKI and hepatic ischemia-reperfusion
IRI-induced AKI model: unilateral kidney IRI was performed as previously described
52. Eight-week-old adult male mice were anesthetized with isoflurane and placed in a right lateral decubitus position. A small incision was made below the ribs to expose the left kidney. The kidney was carefully dissected to remove surrounding fat and connective tissues around the renal artery. A vascular clamp was applied to occlude the renal artery for 70 min. During ischemia, the kidney was repositioned into the abdominal cavity, and 60 mg/kg pentobarbital sodium was administered intraperitoneally. The incision was covered with sterile cotton moistened with normal saline to prevent dehydration.
HIRI model: mice were placed in a supine position, and a midline abdominal incision was made to expose the liver. The hepatic artery beneath the liver was identified and clamped for 25–35 min. During ischemia, the liver was repositioned, and 60 mg/kg pentobarbital sodium was administered intraperitoneally. The incision was covered with sterile cotton moistened with normal saline. All surgical procedures were performed with mice maintained on a heating pad to ensure stable body temperature.
After the designated ischemia period, the vascular clamp was removed to restore blood flow to the kidney or liver. The incision was sutured. Before the final suture closure, 1 mL of sterile saline was administered intraperitoneally to maintain fluid balance. Mice were kept on the heating pad to keep warm until fully awake. Sham-operated mice received identical surgical procedures, except that vascular clamps were not applied. After the designated reperfusion period, cardiac perfusion was performed to collect the left kidney for the AKI model and the largest lobes of the liver for the HIRI model for subsequent experiments.
Sample preparation for neo-N-terminomic analysis
Cells were lysed using 8 M guanidine hydrochloride in 100 mM HEPES, pH 7.8, followed by sonication at 4 °C for 1–2 min per sample. Protein concentration was determined using a BCA assay (ThermoFisher Scientific, 23225), and samples were diluted to 1 mg/mL. For each 1 mL of sample, 10 mM DTT (Beyotime, ST043) was added and incubated at 55 °C for 30 min, followed by 25 mM iodoacetamide (IAA, MCE, HY-34477) incubation in the dark at room temperature for 2 h. Finally, 5 mM DTT was added and incubated at room temperature for 30 min. Exposed N-terminal amino groups were labeled by adding 30 mM formaldehyde (Sigma, F8775) and 30 mM sodium cyanoborohydride (NaCNBH3, Sigma, 156159), followed by incubation at 37 °C with shaking for 12 h. An additional 15 mM formaldehyde and 20 mM NaCNBH3 were supplemented, and the reaction continued for another 12 h under the same conditions. The reaction was quenched by adding 1 M Tris-HCl, pH 6.8 to a final concentration of 100 mM and incubating at 37 °C for 4 h. Approximately 1.1 mg of protein (about 1,200 μL sample) was mixed with 3,600 μL of ice-cold ddH2O, 4800 μL of ice-cold methanol, and 1,200 μL of ice-cold chloroform. After vigorous vortexing, the mixture was centrifuged at 3,200× g at 4 °C for 8 min. The protein precipitate, formed at the interface between organic and aqueous phases, was carefully transferred using a pipette tip to a 1.5 mL EP tube containing ice-cold methanol. The precipitate was washed twice with 1 mL of ice-cold methanol and air-dried. The protein precipitate was dissolved in 70 μL of 8 M guanidine hydrochloride in 100 mM HEPES, pH 7.8, diluted with 800 μL of 100 mM HEPES, pH 7.6, and supplemented with 1.8 mM CaCl2. Then 70 μL of trypsin (≈ 28 μg) was added, and the mixture was incubated at 37 °C for 14 h. After centrifugation at 21,600× g for 10 min, the supernatant was collected. The pellet was resuspended in 20 μL of 8 M guanidine hydrochloride in 100 mM HEPES, pH 7.8, mixed with 150 μL of 100 mM HEPES, pH 7.6 containing 1.8 mM CaCl2, and digested with 35 μL of trypsin (≈ 14 μg) at 37 °C for 12 h. This trypsin-digestion step for protein pellet was repeated to ensure complete proteolytic digestion of all labeled proteins. The supernatants from digestions were combined. 1 mg protein digest was mixed with 100 μL HPG-ALD (49 mg/mL, University of British Columbia) and 37 mM NaCNBH3, followed by incubation at 37 °C for 12 h. An additional 23 μL HPG-ALD and 37 mM NaCNBH3 were supplemented, and the reaction continued for another 12 h. The sample was centrifuged at 21,600× g for 15 min, and the supernatant was collected. The supernatant was filtered through a 10 kDa MWCO centrifugal filter at 13,500× g for 30 min, and the flow-through was retained. The filter was washed with 490 μL ddH2O and centrifuged at 13,500× g for 50 min. The flow-through from the filtration and wash steps was pooled into a 1.5 mL EP tube and desalted for mass spectrometry analysis.
Sample preparation for proteomic analysis of secreted proteases
Serum sample preparation
To deplete immunoglobulins, 200 μL of Protein G beads were added to 1 mL of mouse or FBS and incubated overnight at 4 °C. The supernatant was carefully collected after centrifugation. The collected supernatants were then subjected to two rounds of albumin depletion using the Invent Albumin Removal Kit (Invent, WA-013) according to the manufacturer’s instructions. The resulting protein precipitate was resuspended in 3 mL cold PBS and concentrated by centrifugation using a 10-kDa molecular weight cut-off (MWCO) ultrafiltration device at 4 °C until minimal volume was achieved. This washing process was repeated once with 3 mL PBS. Finally, the sample was recovered in 1 mL PBS. Protein concentrations were determined by BCA assay: 3.2 mg/mL for mouse serum and 5.3 mg/mL for FBS measured. A 200 μL aliquot of each sample was digested with trypsin and subjected to mass spectrometry analysis.
Cell culture supernatant preparation
HEK293T cells, HT29 cells, and BMDMs were seeded in 15-cm dishes and cultured until 80–90% confluency. Cells were washed twice with PBS and subsequently cultured in serum-free DMEM for 12 h. Supernatants were collected and centrifuged at 1,000× g for 5 min to remove cellular debris. The clarified supernatant was concentrated using a 3-kDa MWCO ultrafiltration device to approximately 320 μL. Protein concentrations were determined as follows: HEK293T, 1.05 mg/mL; HT29, 1.56 mg/mL; BMDM, 1.50 mg/mL. Samples were digested with trypsin and subjected to mass spectrometry analysis.
Mass spectrometry and data analysis
The peptides were desalted and analyzed using a nanoElute LC system coupled to a timsTOF Pro mass spectrometer (Bruker, Bremen, Germany) or a Q Exactive HF-X mass spectrometer (Thermo Fisher Scientific). Peptide identification, quantification and protein inference were performed using FragPipe v21.1
53. Tandem mass spectra were searched against the UniProt human or mouse protein database with a precursor and fragment mass tolerance of 20 ppm. Cysteine carbamidomethylation (+57.02146 Da) was set as a static modification, while methionine oxidation (+15.9949 Da), as well as dimethylation of peptide N-termini and lysine residues (+28.031 Da) for neo-N-terminomic analysis were considered variable modifications. The false discovery rate (FDR) at the peptide-spectrum match level was controlled to < 1%. Data processing and statistical analysis was performed using online platform FragPipe-Analyst. PCA was performed using the online platform bioinformatics.com.cn (last accessed 10 Dec 2024)
54. GO analysis was performed by DAVID Bioinformatics
55,56 (National Institutes of Health) and was visualized by bioinformatics.com.cn (last accessed on 10 Dec 2024)
54. Cleavage motif analysis was conducted online using BioLadder (bioladder.cn)
57.
Flow cytometric sorting of SYTOX Green positive and negative cells
HT22 cells, HT29 cells, or BMDMs were subjected to indicated treatments. SYTOX Green (Invitrogen, S7020) was added 30 min before collection. After treatment, both adherent and floating cells were collected. Adherent cells were gently detached, combined with the supernatant, and centrifuged at 750× g. The cell pellet was resuspended in PBS containing 10% FBS, gently pipetted to achieve a single-cell suspension, and filtered through a 70 μm strainer prior to sorting. A portion of the pre-sorted cell suspension was retained and lysed directly with 4% SDS lysis buffer to serve as a control. Cells were sorted based on SYTOX Green fluorescence (positive vs negative) directly into tubes containing 10% SDS lysis buffer. The protein concentrations of both the pre-sorted control and sorted samples were determined using a BCA assay. Samples were normalized to the same concentration and subjected to SDS-PAGE and subsequent analysis.
Live-cell fluorescent protein signal imaging via HCS
BMDMs expressing specific fluorescent proteins were seeded in 24-well plates with glass bottoms at a density of 3–4 × 106 cells per well. After pretreatment with compounds (e.g., SM164, zVAD, or LPS), the culture medium was replaced with phenol red-free medium. Indicated treatment agents were then added, followed by the addition of 1 μg/mL Hoechst stain for nuclear labeling. The plates were immediately transferred to a HCS system for live-cell imaging. Fluorescence channels were set according to the target fluorescent protein. The laser intensity was optimized to minimize photo toxicity. Time intervals were defined for sequential image acquisition, and z-stack settings (number of layers and spacing between slices) were adjusted to capture volumetric data. Multiple fields of view were selected for each well to ensure statistical robustness. Time-lapse imaging was initiated to monitor dynamic changes in fluorescent protein signals under the specified treatments.
Screening for proteases mediating necrotic protein cleavage
HEK293T cells cultured in 6-well plates or 10 cm dishes were transfected with 2 μg or 13 μg of plasmid DNA encoding specific proteases. 12–20 h post-transfection, the culture medium was replaced with serum-free DMEM. After 8–12 h of incubation, the conditioned medium was collected and centrifuged at 4,000× g for 5 min to remove cellular debris. BMDMs seeded in 6-well plates were pretreated with specified compounds. The medium was then replaced with either fresh normal medium or the conditioned medium from protease-expressing HEK293T cells, followed by the addition of indicated compound treatments. After treatment for the designated time, cells were harvested and processed for western blotting analysis to evaluate proteolytic events.
Genomic DNA extraction
Genomic DNA was extracted from HT22 cells using the FastPure Cell/Tissue DNA Isolation Mini Kit (Vazyme, DC102) according to the manufacturer’s instructions with minor adaptations. HT22 cells were seeded in 6-well plates at a density of 3 × 105 cells per well. A total of 27 wells were prepared, representing 9 treatment conditions with 3 biological replicates each. The DNA concentration and elution volume were measured using a spectrophotometer or fluorometer. Genomic DNA from HT22 cells and BMDMs for DNA gel analysis was extracted using the DNA Ladder Extraction Kit (Beyotime, C0008). The extracted genomic DNA was mixed with 10× DNA loading buffer and separated on a 2% agarose gel.
In vitro phagocytosis assay
The donor cells (BMDMs-Sec61β-RFP, BMDMs-Histone4-RFP) were treated with indicated agents to induce necroptosis under standard culture conditions or in the presence of leupeptin combined with heat-inactivated serum (60 °C). Approximately 30 min before sample collection, Hoechst was added to the culture at 2 μg/mL. BMDMs expressing RFP-tagged markers were gently dislodged from the plate using a pipette, centrifuged at 1,200× g for 1 min, washed once with phenol red-free DMEM (supplemented with 10% FBS or 10% heat-inactivated FBS), and centrifuged again at 1,200× g for 1 min. The pellet was resuspended in an appropriate volume (100–400 μL) of the corresponding medium. For necroptotic cells, 200 μM zVAD was added during resuspension; for apoptotic cells, 50 μM GSK872 was added. BMDMs-GFP were detached using trypsin digestion. The cells were resuspended in phenol red-free, serum-free DMEM for subsequent use. In phagocytosis experiments, the phagocytes (BMDMs-GFP) and necrotic cells (BMDMs expressing RFP-tagged markers) were mixed at a ratio of 3:1 or 2:1. Typically, BMDM-GFP phagocytes were seeded at 3–5 × 106 cells. After mixing, serum or leupeptin was applied as required by the experimental conditions. Immediately after mixing, the coculture was transferred to a HCS system. The time-lapse imaging was performed every 10–20 min during the early phagocytosis phase (0–4 or 0-6 h) and every 30 min intervals thereafter.
Intraperitoneal injection of necrotic cells and subsequent analyses
BMDMs or MEFs expressing Histone4-RFP and Sec61β-mBaojin were treated with TSZ (in 9% FBS + 1% mouse serum) or TSZ + leupeptin (in heat-inactivated 9% FBS + 1% mouse serum) for 4 h to induce necroptosis. The cells were gently detached, collected into 50 mL conical tubes, and centrifuged at 1,600× g for 2 min. The supernatant was aspirated, and the pellet was resuspended in PBS (1 mL per 10 cm dish for BMDMs; 1 mL per 15 cm dish for MEFs). A suspension was prepared by mixing 500 μL of cell suspension with 150 μL of 100 mM leupeptin (or 150 μL PBS for controls) and 350 μL PBS. 600 μL of this mixture was administered via intraperitoneal injection per 20 g body weight. This injection regimen was repeated for three consecutive days per week, over a total period of 2–3 weeks.
Two weeks after injection with necrotic BMDMs, the mice were euthanized and spleens were collected, weighed, and photographed. Alternatively, the mice were euthanized three weeks after the injection. Blood was collected via cardiac puncture, allowed to clot, and serum was isolated by centrifugation. The serum was filtered through a 0.22 μm filter and stored for autoantibody analysis.
IF for autoimmunity analysis
MEFs were seeded and allowed to attach on coverslips in 24-well plates for 12 h, and then fixed with 4% PFA for 15 min, and washed three times with PBS. The cells were permeabilized with 0.2% PBST (containing Triton X-100) for 10–15 min, blocked with 5% goat serum for 1 h, and incubated at 4 °C overnight with mouse serum derived from different treatment conditions, filtered through a 0.22 μm filter and diluted 1:50. After washing with PBS, the cells were incubated with goat anti-mouse Alexa Fluor 488 secondary antibody (1:1,000 dilution) at room temperature for 75 min, washed, and stained with 1 μg/mL Hoechst for 10 min. Coverslips were mounted with ProLong antifade reagent.
Peritoneal cell harvest and phagocytosis assay
At 4 or 17 h after injection of necrotic MEFs expressing Histone4-RFP or Sec61β-mBaojin (MHS), mice were euthanized by cervical dislocation, surface-sterilized with 75% ethanol for 3–5 min, and the abdominal skin was incised. 5–6 mL of RPMI-1640 medium containing 2% FBS was injected intraperitoneally for collecting cells from peritoneal cavity. The peritoneal lavage fluid was collected, centrifuged at 400× g for 3 min, and resuspended in RPMI-1640 containing 10% FBS. Cells were seeded in 24-well glass-bottom plates (sample of one mouse per well), allowed to adhere for 2–4 h, and non-adherent cells were removed by washing. Adherent cells were stained with CoraLite® Plus 405 Anti-Mouse CD11b (M1/70) antibody (CL405-65055) from Proteintech (1:250 dilution) in RPMI-1640 containing 1% FBS for 15–30 min at room temperature, washed twice with RPMI-1640 containing 10% FBS, and imaged using HCS to quantify phagocytosis.
Ethics statement
All research involving human participants, human biological materials, and animals complied with the relevant ethical regulations. All animal procedures were conducted according to the protocols approved by the Standing Animal Care Committee at the Interdisciplinary Research Center on Biology and Chemistry. The clinical study was approved by the ethical review board of Renji Hospital of Shanghai Jiao Tong University (2023-047, 2023-048) and was performed according to the Declaration of Helsinki. The clinical samples used in this study were medical waste in nature. The clinical and histological characteristics of the human samples are not publicly available due to data privacy, but that anonymized data can be obtained from the corresponding author. Informed consent was obtained from all individual participants included in this study. Participants were fully informed about the purpose of the research, procedures, potential risks, and benefits, and voluntarily agreed to participate. Since the samples were medical waste, no compensation or payment was provided to participants.
Clinical samples
Human kidney samples were obtained from excess materials from diagnostic renal biopsies performed at Renji Hospital, Shanghai Jiao Tong University School of Medicine, which would otherwise have been discarded as medical waste. Samples were collected from patients with diabetic nephropathy (n = 2; ages 67 and 69 years; both male) and from non-diabetic nephropathy controls (n = 2; ages 46 and 52 years; both male). HOMA-IR higher than 2.69 were considered as patients with insulin resistant. Patients with cancer or auto-immune diseases or under treatment with glucocorticoid or immune suppressive agents were excluded.
Statistics and reproducibility
Curve fitting and statistical analyses were performed using GraphPad Prism 9.0 or Microsoft Excel 2020 software. Statistical significance was assessed using two-tailed paired or unpaired Student’s t-tests, one-way ANOVA with Sidak’s multiple comparisons test, or two-way ANOVA with Dunnett’s or Tukey’s multiple comparisons test, as appropriate. Two-way ANOVA was used for analyses involving multiple time points or two independent variables. Details of the statistical tests used for each experiment are provided in the corresponding figure legends. Statistical significance was defined as *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001. Data are expressed as the mean ± standard deviation (SD).
No statistical method was used to pre-determine sample sizes. We aimed for at least 3 biological replicates per group to allow basic statistical analysis. The exact sample size was indicated in each figure legends. Data collection and analysis were not performed blind to the conditions of the experiments. No animals or data points were excluded from the analysis.
DATA AVAILABILITY
The mass spectrometry proteomics data have been deposited to the ProteomeXchange Consortium (
proteomecentral.proteomexchange.org) via the iProX partner repository
58,59, with the dataset identifier PXD069280.
The Author(s) 2026. Published by Higher Education Press. This is an Open Access article distributed under the terms of the CC BY license (https://creativecommons.org/licenses/by/4.0/).
