Murine macrophage RAW 264.7 cell line was purchased from the American Type Culture Collection (ATCC). Human blood samples were obtained from the New York Blood Center (Long Island City, NY, USA) to harvest primary human peripheral blood mononuclear cells (PBMCs) by density gradient centrifugation as previously described (4, 15, 16). Both macrophage cultures and PBMCs were routinely incubated in DMEM media containing 1% streptomycin/penicillin and 10% fecal bovine serum or 10% human serum. When cell densities reached 80-90% confluence, adherent macrophages or PBMCs were stimulated with bacterial endotoxins (lipopolysaccharides, LPS, E. coli 0111:B4, #L4130, Sigma-Aldrich) or recombinant human or murine pCTS-L protein in the absence or in the presence of each of the 1360 compounds in the U.S. Collection of Drug (10 mM in DMSO), as well as progesterone (Cat. # P0130, Sigma-Aldrich) solubilized in ethanol (5 mg/ml) or entrapped into 2,6-dimethyl-β-cyclodextrin (DM-β-CD, Cat. #H0513, Sigma-Aldrich) nanoparticles. The progesterone-carrying DM-β-CD nanoparticles (containing 84.4 mg progesterone per 1000 mg of the PRO/DM-β-CD complex) were also purchased from Sigma-Aldrich (Cat. # P7556). The extracellular levels of various cytokines and chemokines were respectively measured by using ELISA kits or Cytokine Antibody Arrays as previously described (4, 15, 16).

Recombinant murine and human pCTS-L proteins containing an N-terminal 6×Histidine tag were respectively expressed in E. coli BL21 (DE3) pLysS cells and purified to homogeneity as previously described (4, 16). Briefly, upon sonication to disrupt the bacterial cell wall, the inclusion bodies containing pCTS-L proteins were harvested via differential centrifugation technique following sequential washings in 1% Triton X-100 dissolved in 1 × PBS buffer The purified inclusion bodies were subsequently dissolved in high concentration of urea solution (8.0 M), and then refolded by dialysis in Tris buffer (10 mM, pH 8.0). Afterward, recombinant pCTS-L proteins were purified by histidine-affinity chromatography technique and Triton X-114 extractions. The endotoxin content of recombinant pCTS-L proteins was estimated to be < 0.01 U per µg of pCTS-L protein.

We obtained the U.S. Collection of 1360 drugs ( Supplementary Table 1 , 10 mM in DMSO) from the MicroSource Discovery System Inc., and used this chemical library to search for potential pCTS-L inhibitors as previously described (16). Briefly, murine macrophage-like RAW 264.7 cells were challenged with murine pCTS-L protein in the absence or in the presence of each drug at several concentrations for 16 h, and levels of TNF in macrophage-conditioned medium was measured using specific TNF DuoSet ELISA kit (Cat# DY410, R&D Systems).

Murine Cytokine Antibody Array Kits (Cat.# AAM-CYT-3-8, RayBiotech Inc., Norcross, GA, USA) were employed to measure the relative levels of 62 cytokines/chemokines in macrophage cell culture-conditioned medium or murine serum as previously described (4, 15, 16). Similarly, human Cytokine Antibody C3 Array Kits (Cat.# AAH-CYT-3-8) were employed to measure the relative levels of 42 cytokines/chemokines in human PBMC-conditioned culture medium as previously described (4, 15, 16).

Adult male and female Balb/C mice (7-8 weeks old, 20-25 g body weight) were purchased from Charles River Laboratories (Wilmington, MA), housed in a temperature-controlled room on a 12-h light-dark cycle, and acclimated for at least 5-7 days before usage. Every attempt was made to limit the number of animals used in the present study as per the ARRIVE guidelines for reducing the number of animals in scientific research developed by the British National Centre for the Replacement, Refinement and Reduction of Animals in Research (NC3Rs). Additionally, all experiments were performed in accordance with the International Expert Consensus Initiative for Improvement of Animal Modeling in Sepsis - Minimum Quality Threshold in Pre-Clinical Sepsis Studies (MQTiPSS) (17), which includes practices such as randomization of animals in each experimental group, delayed therapeutic interventions with therapeutic agents (e.g., PRO-entrapping DM-β-CD nanoparticles) (18), establishment of specific criteria for euthanasia of moribund septic animals (e.g., labored breathing, minimized response to human touch, and immobility), as well as the administration of fluid resuscitation and antibiotics (19). This study was administratively approved by the IACUC of the Feinstein Institutes for Medical Research (FIMR, Protocol # 2017-003 Term II; Date of Approval, April 28th, 2020).

Adult male or female Balb/C mice aged 7-8 weeks and weighing 20-25 g underwent a surgical procedure referred to as “cecal ligation and puncture” (CLP) to induce experimental sepsis as previously outlined (4, 15, 16). Briefly, the cecum of Balb/C mice was surgically exposed, ligated approximately 5.0 mm from the cecal tip, and punctured once with a 22-gauge syringe needle. Prior to CLP surgery, all experimental animals received a buprenorphine injection (0.05 mg/kg, s.c.) to alleviate immediate surgical pain, because repetitive use of buprenorphine in the CLP model could paradoxically elevate sepsis surrogate markers and animal lethality (20, 21), leading to unnecessary use of more animals per experimental group. Additionally, a small dose of bupivacaine and lidocaine was locally injected around the incision site immediately after CLP surgery. Approximately 30 min post CLP surgery, all experimental animals were subcutaneously injected with imipenem/cilastatin (0.5 mg/mouse) (Primaxin, Merck & Co., Inc.), followed by resuscitation with sterile saline solution (20 ml/kg). Septic animals were only given a single dose of antibiotics before pharmacological administration of PRO-entrapping DM-β-CD nanoparticles at 24 h post CLP, worrying that subsequent antibiotics treatment may adversely affect the therapeutic efficacy of PRO-entrapping DM-β-CD nanoparticles. Before treatment, animals were randomly assigned to control vehicle and experimental groups, and PRO dissolved in sesame oil (Cat. #S3547, Sigma-Aldrich; 2.0 mg/ml) or water after complexation with DM-β-CD to form nanoparticles (containing 84.4 mg PRO per 1000 mg of PRO/DM-β-CD complex) was intraperitoneally injected to septic mice at various time points post CLP surgery. Animal survival was observed for two weeks to ensure no late death occurred. To elucidate the potential protective mechanisms of PRO-entrapping DM-β-CD nanoparticles, a separate group of Balb/C mice were subjected to CLP, and DM-β-CD vehicle (14.65 mg/kg) or PRO-entrapping DM-β-CD nanoparticles (containing 1.35 mg/kg PRO and 14.65 mg/kg DM-β-CD) were administered at 2 h and 20 h post CLP. At 24 h post CLP, animals were euthanized to collect blood and measure serum levels of various cytokines and chemokines using murine Cytokine Antibody Arrays or markers of tissue injury using specific colorimetric enzymatic assays.

Blood samples were harvested at 24 h post CLP following intraperitoneal administrations of DM-β-CD vehicle (14.65 mg/kg) or PRO-entrapping DM-β-CD nanoparticles (containing 1.35 mg/kg PRO and 14.65 mg/kg DM-β-CD) at 2 h and 20 h post CLP, and centrifuged at 3000 x g for 10 min to collect serum. Serum levels of liver injury markers such as aspartate aminotransferase (AST, Cat. No. 7561), and lactate dehydrogenase (LDH, Cat. No. 7572) were determined using specific colorimetric enzymatic assays (Pointe Scientific, Canton, MI) according to manufacturer’s instructions as previously described (15).

We employed the Nicoya Lifesciences’ gold-nanoparticle-based Open Surface Plasmon Resonance (OpenSPR) technology (Kitchener, ON, Canada) to characterize protein-drug interaction following the manufacturer’s instructions. Briefly, recombinant pCTS-L was immobilized on NTA sensor chip (Cat. # SEN-Au-100-10-NTA) as previously described (4, 22), and DMSO solution of PRO was applied as an analyte at different concentrations. The sensorgrams of the dynamic ligand-analyte interaction were recorded over time to estimate the equilibrium dissociation constant (K_D) (Nicoya Lifesciences).

This study was administratively approved by the institutional review board (IRB) of the FIMR (IRB protocol #18-0184) and consented by all patients participants who were diagnosed with sepsis or septic shock based on the Sepsis-3 criteria (23). Small volume of blood samples (5.0 ml) was obtained from eleven septic patients recruited to the Long Island Jewish Medical Center or North Shore University Hospital between 2018-2019 at three time points: Time 0 (within 24 h of the initial diagnosis); Time 24 h (24 h post the initial diagnosis); and Time 72 h (72 h post the initial diagnosis). The demographics of these eleven septic patients have been reported previously (4), but briefly described in the Supplementary Table 2 . These clinical samples were assayed for pCTS-L levels using human pCTS-L ELISA kit (Cat.# MBS7254442, MyBioSource.com) with reference to standard cure generated from using recombinant human pCTS-L. In parallel, the serum concentrations of progesterone were measured by using highly sensitive ELISA kit (Cat. # ADI-901-011, ENZO Life Sciences, Inc., Farmingdale, NY, USA).

All data were initially evaluated for normality by the Shapiro-Wilk test before conducting appropriate statistical tests. The Student’s t test was used to compare two independent experimental groups. For comparison among multiple groups with non-normal (skewed) distribution (as assessed by the Shapiro-Wilk test), the statistical difference was evaluated with the non-parametric Kruskal-Wallis ANOVA test followed by the Dunn’s test. The Kaplan-Meier method was employed to compare the differences in mortality rates between two different groups along with the nonparametric log-rank post hoc test. Finally, the Spearman rank correlation coefficient test was used to evaluate associations between two quantitative variables that exhibited non-normal distribution. Statistical significance was defined as a P value less than 0.05.

To explore novel pCTS-L inhibitors, we adapted a macrophage cell-based bioassay that we recently developed (16) to screen a U.S. Collection of 1360 drugs supplied as 10.0 mM Dimethyl Sulfoxide (DMSO) solutions in seventeen 96-well microplates for possible activities to inhibit pCTS-L-stimulated TNF secretion ( Figure 1A ). We optimized the experimental conditions by respectively titrating the concentration of pCTS-L protein (to 1.0 μg/ml) and the confluence of macrophage cultures (to 80-90%). A complete screening of 1360 drugs of the U.S. Collection resulted in the identification of progesterone (PRO) and three analogs (i.e., dydrogesterone, exemestane, and medroxyprogesterone acetate; Figure 1B ) as inhibitors of pCTS-L-stimulated TNF production. When dissolved in DMSO, PRO dose-dependently attenuated pCTS-L-mediated TNF secretion with an estimated IC around 20.0 µM ( Figure 1B ) without affecting mitochondrial metabolic activity (MTT assay) or cell viability (Trypan blue uptake , Supplementary Figure 1 ).

To confirm PRO’s pCTS-L-inhibitory activities, we first dissolved it in ethanol before testing its effects on pCTS-L-stimulated production of 42 cytokines and chemokines in primary human PBMCs. In agreement with our earlier report (4), pCTS-L markedly stimulated the secretion of a few chemokines (e.g., ENA-78, GRO, and MCP-1) and cytokines (e.g., TNF, IL-6 and IL-10) ( Figure 2 ). However, the pCTS-L-mediated TNF production was markedly inhibited by the co-administration of PRO at a relatively high dose (40 μM, Figure 2 ), an almost most effective dose that suppressed pCTS-L-induced TNF secretion in dose-response studies using murine macrophages ( Figure 1B ) as well as primary PBMCs from several different donors. Likewise, the pCTS-L-triggered secretion of several other cytokines (e.g., IL-10) and chemokines (e.g., ENA-78 and MCP-1) was similarly suppressed by PRO ( Figure 2 ). Consistent with our earlier reports (16, 24), endotoxins effectively stimulated the secretion of several chemokines (MCP-1, ENA-78 and GRO) and cytokines (e.g., TNF, IL-6 and IL-10) in primary human PBMCs ( Figure 2 ). However, PRO was unable to affect LPS-induced production of any cytokines/chemokines ( Figure 2 ) even at the concentration that markedly inhibited pCTS-L-mediated cytokine/chemokine production, suggesting that PRO selectively inhibited pCTS-L-mediated inflammation.

The β-cyclodextrin (β-CD) is defined as cyclic oligosaccharides of seven glucopyranoses in its β-chair conformation ( Figure 3A ), thereby displaying the shape of truncated cone with an outer hydrophilic surface that render it water-soluble and an inner hydrophobic core (25) that can entrap hydrophobic molecules (e.g., PRO). To increase the water-solubility of β-CD, the hydroxyl groups of the inner hydrophobic cavity (position 2) and the outer hydrophilic surface (position 6) of β-CD can be substituted with hydrophobic methyl moieties to produce the 2,6-dimethyl-β-cyclodextrin (DM-β-CD, Figure 3A ), which still displayed the shape of cone with slightly less symmetry and smaller inner cavity ( Figure 3A ). As a lipophilic sterol, PRO could be inserted into the hydrophobic inner cavity of two DM-β-CD molecules at 1:2 molar ratio ( Figure 3A ). These PRO-carrying DM-β-CD nanoparticles dose-dependently suppressed pCTS-L-mediated production of a few cytokines (e.g., TNF, IL-6 and IL-10) and chemokines (GRO, IL-8 and MCP-3) in human PBMCs ( Figure 3B ), confirming that PRO-carrying DM-β-CD nanoparticles maintained pCTS-L-inhibiting properties of PRO under pharmacological conditions.

To assess the PRO’s therapeutic efficacy, we first dissolved PRO in Sesame oil containing various unsaturated fatty acids that could emulsify and dissolve lipophilic PRO in the form of micelles (26–30). When dissolved in sesame oil, PRO did not show any protection against sepsis within a dose range of 1.0 mg/kg to 8.0 mg/kg ( Figure 4A , Left Panel). When given intraperitoneally at a relatively high dose (16 mg/kg) at 2 h and 24 h post CLP, PRO promoted a marked protection against sepsis, significantly increasing animal survival rates from 30% to 70% ( Figure 4A ). In an effort to enhance its pharmacological efficacy, PRO was entrapped into DM-β-CD nanoparticles to enhance its water-solubility and bioavailability, and further tested in an animal model of experimental sepsis induced by cecal ligation and puncture (CLP). The PRO-carrying DM-β-CD nanoparticles was given at 24 h after CLP, which was a time point when circulating pCTS-L plateaued and some animals started to succumb to death (4). As a vehicle control, DM-β-CD (14.65 mg/kg) itself did not affect animal survival rate when it was given at 24 h and 48 h post CLP surgery ( Figure 4A ). In contrast, the PRO-carrying DM-β-CD nanoparticles significantly rescued both male and female mice from infections even when the 1st dose was intraperitoneally administered at 24 h post CLP ( Figure 4A ). This effective protective dose of PRO entrapped in the DM-β-CD nanoparticles (even when initially given at 24 h post CLP) was almost 10-fold lower than the dosage of progesterone dissolved in sesame oil (when first given at 2 h post CLP.) The calculated molar concentration of PRO (MW = 314.47 Daltons; 1.4 mg/kg) intraperitoneally injected into septic animals was estimated to be > 4.5 μM and comparable to the effective concentrations (10 µM) of PRO in inhibiting pCTS-L-induced dysregulated inflammation in vitro ( Figure 3B ).

To elucidate the underlying protective mechanisms of PRO, we assessed its impact on CLP sepsis-triggered inflammation and tissue injury. In agreement with our previous report (4), experimental sepsis markedly elevated blood levels of G-CSF, MIP-2 and sTNFR1 ( Figure 4B ). In contrast to DM-β-CD vehicle, which did not affect sepsis-triggered systemic accumulation of any cytokines or chemokines, PRO-carrying DM-β-CD nanoparticles significantly reduced sepsis-triggered elevation of MIP-2 and sTNFRI ( Figure 4B ). Furthermore, in contrast to DM-β-CD vehicle, which did not affect sepsis-triggered release of liver enzymes (such as AST) or other general tissue injury markers such as LDH, PRO-carrying DM-β-CD nanoparticles significantly reduced the levels of AST and LDH in septic animals ( Figure 4C ), suggesting that PRO-entrapping DM-β-CD nanoparticles protect mice against infections partly by suppressing sepsis-mediated dysregulated inflammation and tissue injury.

To gain further insight into PRO’s protective mechanisms, we examined the possible interaction and relationship between systemic accumulation PRO and pCTS-L in clinical sepsis. SPR analysis revealed a strong interaction between PRO and pCTS-L, as evidenced by the relatively low equilibrium dissociation constant (KD, Figure 5A ) for PRO-pCTS-L interaction, suggesting that PRO might bind pCTS-L to inhibit its proinflammatory properties under pharmacological conditions. In agreement with previous findings of a marked (1 - 5 folds) elevation of blood PRO levels in septic animals (31) or patients with severe sepsis (32) or septic shock (33), we found a significant increase of serum PRO levels in septic mice ( Figure 5B ), as well as a parallel positive correlation with Sequential Organ Failure Assessment (SOFA, Supplementary Table 1 ) score and serum pCTS-L concentrations in patients with clinical sepsis ( Figure 5C ), confirming a possible progesterone upregulation as a potential protective mechanism against infections.