Identification of procathepsin L (pCTS-L)–neutralizing monoclonal antibodies to treat potentially lethal sepsis

Antibody-based strategies have been attempted to antagonize early cytokines of sepsis, but not yet been tried to target inducible late-acting mediators. Here, we report that the expression and secretion of procathepsin-L (pCTS-L) was induced by serum amyloid A (SAA) in innate immune cells, contributing to its late and systemic accumulation in experimental and clinical sepsis. Recombinant pCTS-L induced interleukin-6 (IL-6), IL-8, GRO-α/KC, GRO-β/MIP-2, and MCP-1 release in innate immune cells and moderately correlated with blood concentrations of these cytokines/chemokines in clinical sepsis. Mechanistically, pCTS-L interacted with Toll-like receptor 4 (TLR4) and the receptor for advanced glycation end products (RAGE) to induce cytokines/chemokines. Pharmacological suppression of pCTS-L with neutralizing polyclonal and monoclonal antibodies attenuated pCTS-L–mediated inflammation by impairing its interaction with TLR4 and RAGE receptors, and consequently rescued animals from lethal sepsis. Our findings have suggested a possibility of developing antibody strategies to prevent dysregulated immune responses mediated by late-acting cytokines.

Table S1.Reagent sources Table S2.Demographics of two cohort of normal healthy controls and 10 septic patients Other Supplementary Materials for this manuscript include the following: Movie S1.Rotating image of the TLR4-pCTS-L complex with the lowest Gibbs free energy.

Movie S2.
Rotating image of the RAGE-pCTS-L complex with the lowest Gibbs free energy.
Data file S1.Primary data.

Western blotting
The concentrations of pCTS-L in murine macrophage-or human PBMC-conditioned culture medium, murine serum or human plasma were determined by Western blotting analysis using commercial mAb (#C0994, Sigma-Aldrich) or pAb (Cat.# SC6498, Santa Cruz), or home-made rabbit anti-murine pCTS-L pAb.The concentrations of cellular pro-Casp-11 and mature Casp-11 in pCTS-L-stimulated peritoneal macrophages were determined by Western blotting using rabbit anti-mouse caspase-11 monoclonal antibodies (Cat.# ab180673, Abcam).Equal volume of cell-conditioned culture medium or murine/human serum were resolved on sodium dodecyl sulfate (SDS)-polyacrylamide gels and transferred to polyvinylidene difluoride (PVDF) membranes.After blocking with 5% nonfat milk, the membranes were incubated with the appropriate antibodies (anti-pCTS-L, 1:1000; anti-pro-Casp-11, 1:1000) overnight.Subsequently, the membranes were incubated with the appropriate secondary antibodies, and the immune-reactive bands were visualized by chemiluminescence.The relative levels of specific proteins were determined using the UN-SCAN-IT Gel Analysis Software Version 7.1 (Silk Scientific Inc., Orem, UT, USA) with reference to appropriate controls.The liver content of fibrinogen-γ (FGG) was measured by Western blotting analysis using mouse anti-FGG monoclonal antibody following standard procedures with reference to a house-keeping protein, β-actin.Wild-type Ctsl +/+ or Ctsl -/-KO (NOD.129P2(B6)-CtsltmCptr /Rcl J) mice were intraperitoneally administered with bacterial endotoxin (8.0 mg/kg), and animals were sacrificed at 24 h post endotoxemia to harvest blood.The concentration of pCTS-L in murine serum was determined by Western blotting analysis using home-made rabbit anti-murine pCTS-L polyclonal antibodies.Equal volume of serum was resolved on sodium dodecyl sulfate (SDS)-polyacrylamide gels and transferred to polyvinylidene difluoride (PVDF) membranes.After blocking with 5% nonfat milk, the membranes were incubated with the appropriate antibodies (anti-pCTS-L, 1:1000) overnight.Subsequently, the membranes were incubated with the appropriate secondary antibodies, and the immune-reactive bands were visualized by chemiluminescence.

ELISAs
To confirm the relative cytokines levels, ELISA kits were used to quantitate the concentrations of pCTS-L in parallel with several biomarkers of experimental and clinical sepsis.An ELISA kit for human pCTS-L (Cat.#MBS7254442, MyBioSource.com)was used measure blood pCTS-L levels in normal healthy controls and septic patients.In addition, we obtain ELISA kits for other inflammatory biomarkers including GRO (Cat.# ELH-GRO-1, RayBiotech), IL-6 (Cat.# MBS8123859, MyBioSource), IL-8 (Cat.#ELH-IL8, RayBiotech), MCP-1 (Cat.#MBS7721397), and HMGB1 (Cat.#OKCD03560, Aviva Systems Biology) to measure their levels in septic patients as well as in gender-matched normal healthy controls.

Preparation of recombinant human and murine pCTS-L proteins
The cDNA encoding for human (residue 17-333) and murine (residue 18-334) pCTS-L was cloned into a pReceiver expression vector downstream of a T7 promoter with an N-histidine tag, and recombinant pCTS-L protein was expressed in E. coli BL21 (DE3) pLysS as previously described ( 49).The inclusion body-associated recombinant pCTS-L protein was isolated by differential centrifugation and urea solubilization before refolding in Tris buffer (pH 8.0) containing N-lauroylsarcosine.The recombinant pCTS-L protein with N-His Tag was then further purified by histidine-affinity chromatography, followed by extensive Triton X-114 extractions to remove contaminating endotoxins.Recombinant pCTS-L protein was tested for LPS content by the chromogenic Limulus amebocyte lysate assay (Endochrome; Charles River), and the endotoxin content was less than 0.01 U per microgram of recombinant protein.For comparison, we also obtained bacterial products-free pCTS-L expressed in human HEK293 cells (Cat.#.CT1-H5222, Acro Biosystems) as an additional control for recombinant protein expressed in E. coli.

Generation of anti-pCTS-L polyclonal and monoclonal antibodies
Polyclonal antibodies were generated in female New Zealand white rabbits by the Covance Inc.
(Princeton, NJ, USA) using recombinant murine and human pCTS-L in combination with Freund's complete adjuvant following standard procedures.Blood samples were collected in 3week cycles of immunization and bleeding, and the antibody titers were determined by direct pCTS-L ELISA.Total IgGs and pCTS-L antigen-binding IgGs were purified from anti-pCTS-L rabbit serum using Protein A and pCTS-L-affinity chromatography as described in the Supplemental Materials.
The monoclonal antibodies were generated in Balb/C and C57BL/6 mice by the GenScript (Piscataway, NJ, USA) using highly purified human or murine pCTS-L following standard procedures.Blood samples were collected every two weeks, and serum titers were assessed by indirect ELISA and Western blotting analysis.After four immunizations, mouse splenocytes were harvested, fused with mouse Sp2/0 myeloma cell line, and screened for antibody-producing hybridomas by indirect ELISA, dot blotting, and Western blotting analysis.After limiting dilution, purified hybridoma clones were generated to produce mAbs following standard procedures.

Affinity purification of polyclonal antibodies
Total IgGs and pCTS-L antigen-binding IgGs were purified from anti-pCTS-L rabbit serum using Protein A and pCTS-L-affinity column chromatography, respectively.Briefly, rabbit serum was pre-buffered with PBS and slowly loaded onto the Protein A/G Sepharose (Cat.# ab193262) column to allow sufficient binding of IgGs.After washing with 1xPBS to remove unbound serum components, the IgGs were eluted with acidic buffer (0.1 M glycine-HCl, pH 2.8), and then immediately dialyzed into 1×PBS buffer at 4°C overnight.For pCTS-L antigen-affinity purification, recombinant murine pCTS-L was conjugated to cyanogen bromide (CNBr)activated Sepharose4 agarose beads (Cat.# 17098101, GE Healthcare), and the pCTS-Lconjugated Sepharose beads were then loaded onto columns.Following repetitive washings with acid buffer (0.1 M Acetic/Sodium Acetate, 0.5 M NaCl, pH4.0) and alkali buffer (0.1 M Tris-HCl, 0.5 NaCl, pH 8.0), anti-pCTS-L total IgGs were slowly loaded onto the column, and the flow-through fractions were collected.Following repetitive washing with 1×PBS buffer, the pCTS-L-binding antibodies were eluted with acidic elution buffer, and immediately neutralized in 1×PBS buffer.

Peptide dot blotting
A library of 24 synthetic peptides corresponding to different regions of human or murine pCTS-L sequence were synthesized at the Genscript, and spotted (0. 1 μg in 2.5 μl) onto nitrocellulose membrane (Thermo Scientific, Cat No. 88013).Subsequently, the membrane was probed with anti-pCTS-L rabbit or murine serum, or IgGs isolated from anti-pCTS-L rabbit serum or murine hybridoma cultures following a standard protocol.

Fig. S1. Western blotting analysis of LPS-and SAA-induced pCTS-L release and expression in human peripheral blood mononuclear cells (PBMCs).
Human PBMCs were stimulated with LPS or SAA at indicated concentrations for 16 hours, and proteins in the PBMC-conditioned culture medium (Panel A) or whole-cell lysate (Panel B) were immunoblotted with different anti-CTS-L monoclonal antibodies.For cellular proteins, a house-keeping protein, β-actin was used to confirm equivalent loading of different samples.In the culture medium conditioned by LPS-or SAA-stimulated human PBMCs, two different mAbs similarly recognized a single band of 40 kDa, confirming that LPS and SAA induced pCTS-L release in human PBMCs.In PBMC whole-cell lysate, the Santa Cruz mAb SC32801 not only recognized a 40-kDa band that was up-regulated by SAA or LPS stimulation, but also a 25-35 kDa band that matched the molecular weight of mature CTS-L.pCTS-L is folded with the assistance of three disulfide bonds (not shown) and the pro-region (green), which can then be cleaved in the endosome to release the cathepsin L (CTS-L).Upon delivery to the lysosome, the CTS-L is further proteolytically processed to produce the active enzyme consisting of a heavy (dark blue) and light chain (light blue).The amino acid sequence homology among mouse, human, and a parasite (liver fluke, Fasciola hepatica) CTS-L is also shown.

B)
Expression and purification of recombinant human and murine pCTS-L proteins.Recombinant human ("H") and murine ("M") pCTS-L corresponding to residue 17-333 or 18-334 of respective precathepsin L with a N-histidine tag were expressed in E. coli BL21 (DE3) pLysS cells as insoluble inclusion bodies.After sonication to disrupt the bacteria, the inclusion bodies were isolated by differential centrifugation following extensive washing in 1 × PBS containing 1% Triton X-100.The inclusion bodies were then solubilized in 8 M urea, and refolded by dialysis in 10 mM Tris buffer (pH 8.0) containing N-lauroylsarcosine.Subsequently, the recombinant proteins were subjected to extensive Triton X-114 extractions to remove contaminating endotoxins.

Fig. S4. Critical role of TLR4 and RAGE in pCTS-L-induced production of KC/RGO-α and MIP-2/GRO-β in peritoneal macrophages.
Thioglycollate-elicited primary macrophages were isolated from wild-type (WT) C57BL/6 or mutant C57BL/6 mice deficient in both TLR4 and RAGE (T/R dKO).Following stimulation with pCTS-L (1.0 or 3.0 mg/ml) for 6 or 24 h, the extracellular levels of MIP-2 and KC were determined by ELISA assays and expressed as mean ± SEM of three experiments (n = 3).*, P < 0.05 versus "-pCTS-L" negative control of respective genotype (WT or T/R dKO); #, P < 0.05 versus WT " + PCTS-L" positive control treated with pCTS-L at the same concentrations (1.0 or 3.0 μg/ml).Note that the disruption of both TLR4 and RAGE completely abrogated pCTS-L (1.0 mg/ml)-induced production of both MIP-2 (homolog of human GRO-β) and KC (homolog of human GRO-α) at both 6 and 24 h post pCTS-L stimulation.Wildtype C57BL/6 (WT) or mutant C57BL/6 mice deficient in both TLR4 and RAGE (T/R dKO) were intraperitoneally administered with recombinant pCTS-L (40 mg/kg), and animals were euthanized to harvest blood to measure serum levels of 62 cytokines by Cytokine Antibody Arrays.The relative cytokine levels were expressed as mean ±SEM of three animals (n = 3) in arbitrary units.*, P < 0.05 versus "-pCTS-L" negative control or "+ pCTS-L" positive control of TLR4/RAGE double KO mice (T/R dKO).Note that pCTS-L significantly elevated blood levels of BLC, G-CSF, MIP-1γ and sTNF-RI in the wildtype, but not in mutant mice deficient in TLR4 and RAGE, confirming an important role for these receptors in pCTS-L-mediated dysregulated inflammation.For comparison, serum samples from normal healthy Balb/C mice or Balb/C mice subjected to CLP surgery were used as additional controls.Note that CLP surgery induced a marked elevation of blood concentrations of three cytokines/chemokines (i.e., BLC, G-CSF, MIP-1γ) and sTNF-RI in the WT, but not in T/R dKO, some of which (i.e., IL-6 and KC) were known as surrogate markers of experimental sepsis.A) Scheme for purifying pCTS-L antigen-binding IgGs from rabbit total IgGs.Total IgGs were purified from anti-murine pCTS-L rabbit serum using Protein A affinity chromatography, and pCTS-L-binding IgGs were then purified by pCTS-L antigen-affinity chromatography.The non-pCTS-L-binding control immunoglobulins ("C-IgG") were collected in the washout fractions before the pCTS-L antigen-bound antibodies ("A-IgGs") were eluted from the column by an acidic buffer into solution with physiological pH to prevent acid-catalyzed antibody denaturation.
B) pCTS-L antigen-affinity purified IgGs abrogated the pCTS-L-induced cytokines and chemokines.Thioglycollate-elicited peritoneal macrophages were isolated from Balb/C mice, and stimulated with recombinant pCTS-L either alone or in the presence of control IgGs (C-IgGs) or antigen-affinity purified IgGs ("A-IgGs") for 16 h, and the extracellular levels of 62 different cytokines and chemokines were measured by Cytokine Antibody Arrays.The relative cytokine levels were expressed as mean ±SEM of two experiments in duplicates (n = 4) in arbitrary units.*, P < 0.05 versus "-pCTS-L" negative control; #, P < 0.05 versus "+ pCTS-L" positive control.Note that antigen-affinity purified IgGs effectively abrogated the pCTS-L-induced secretion of IL-6, sTNF RII and six different chemokines such as RANTES, MCP-1, MIP-1γ, LIX, MIP-2 and KC.

Fig. S8. Epitope mapping and antigen affinities of human and murine pCTS-L-reactive monoclonal antibodies.
A) Use of eight pairs of homologous peptides corresponding to different regions of murine (M) human (H) pCTS-L to determine the epitope profile of monoclonal antibodies raised against murine pCTS-L.The isotype, relative binding affinity (KD) to human pCTS-L (Hp) or murine pCTS-L (Mp), as well as protective efficacy in CLP sepsis were also noted.

B)
Use of 24 peptides corresponding to different region of human pCTS-L to characterize the epitope profile of monoclonal antibodies raised against human pCTS-L.Note that three P13reactive mAbs recognized both recombinant human pCTS-L (Hp) and murine pCTS-L (Mp) on the dot blots.Pathogen-associated molecular pattern molecules (PAMPs such as bacterial lipopolysaccharide, LPS) relies on cell surface pattern recognition receptors (PRR, TLR4) to activate innate immune cells to immediately release "early" proinflammatory mediators (such as TNF, IL-1β, and IFNγ), which then stimulate hepatocytes and innate immune cells to synthesize and secrete a proinflammatory mediator, serum amyloid A (SAA).SAA then activate innate immune cells to upregulate and secrete procathepsin-L (pCTS-L), which binds to cell surface PRRs such as TLR4 and RAGE to induce: i) the expression of cytokines/chemokines to trigger dysregulated inflammation; and ii) the expression of pro-Casp-11 to activate inflammasome and pyroptosis.The pCTS-L-mediated dysregulation inflammation and pyroptosis-associated immunosuppression may adversely contribute to the pathogenesis of lethal sepsis.A panel of P13-reactive mAbs, such as mAb20, could bind to pCTS-L to interrupt its interaction with TLR4 and RAGE, thereby impairing pCTS-L-mediated dysregulated inflammation to confer protection against lethal sepsis.Other Supplementary Materials for this manuscript include the following: Movie S1.Rotating image of the TLR4-pCTS-L complex with the lowest Gibbs free energy.

Movie S2.
Rotating image of the RAGE-pCTS-L complex with the lowest Gibbs free energy.
Data file S1.Primary data.

Fig. S2 .
Fig. S2.Comparison of blood pCTS-L concentrations between age-matched healthy controls and septic patients.

Fig. S3 .
Fig. S3.Expression and purification of recombinant human and murine pCTS-L.

Fig. S6 .
Fig. S6.Sequence of 24 synthetic peptides of human and murine pCTS-L and epitope mapping of anti-pCTS-L rabbit and murine serum.

Fig. S9 .Fig. S10 .
Fig. S9.Recombinant pCTS-L expressed in human kidney cells similarly activated human PBMCs.Fig. S10.Top ten models of ClusPro Protein-Protein Docking of pCTS-L interaction with TLR4 or RAGE Fig. S11.Proposed model for the pCTS-L-neutralizing mAbs against lethal sepsis.
Fig. S2.Western blotting analysis of serum pCTS-L in a cohort of older healthy controls.
Fig. S5.Requirement of TLR4 and RAGE for pCTS-L-induced cytokine/chemokine production in vivo.

Fig. S6 .Fig
Fig. S6.Sequence of 24 synthetic peptides of human and murine pCTS-L and epitope mapping of anti-pCTS-L rabbit and murine serum.

Fig
Fig. S9.Recombinant pCTS-L expressed in human kidney cells similarly activated humanPBMCs.
, P = 0.50 for the mean age between Healthy Control II (H-II) and Septic Patients (S-I).*, P = 0.00014 for mean blood pCTS-L concentration between age-matched Healthy Control II (H-II) and Septic Patients (S-I).