ABSTRACT

   Pyroptosis is an inflammatory immune response of eukaryotic cells to
   bacterial lipopolysaccharide (LPS) and other pathological stimuli,
   leading to the activation of the gasdermin D (GSDMD) and secretion of
   pore-forming domain GSDMD[Nterm], facilitating the release of
   cytokines. Additionally, GSDMD[Nterm] exhibits antibacterial properties
   through interactions with bacterial outer membranes (OM). We explored
   alternative antimicrobial strategy to determine whether inducing
   natural pyroptosis via GSDMD activation by P. aeruginosa LPS could
   enhance the effectiveness of recombinant phage endopeptidase KP27
   (peptidoglycan-degrading enzyme) against P. aeruginosa, enabling
   penetration through OM and bacterial killing synergistically. Our
   findings demonstrated that recombinant GSDMD alone exhibited
   antibacterial effects against wild-type P. aeruginosa with smooth LPS,
   while recombinant GSDMD[Nterm] efficiently permeabilized both smooth
   LPS-bearing and O-chain-deficient P. aeruginosa potentially synergizing
   with endolysin KP27. Transcriptomic analyses revealed the activation of
   the immune system pathways in response to LPS, mainly in monocytic
   cells, in contrast to epithelial A549 or HeLa cell lines. LPS-induced
   pyroptosis in monocytes led to GSDMD cleavage and the release of
   interleukins, regardless of the nature/origin of the LPS used. However,
   the pyroptosis stimulation by LPS in the antibacterial assay was not
   effective enough for bacterial OM permeabilization and enhancement of
   endolysin activity. We assume that leveraging pyroptosis induction in
   monocytic cells to augment the bactericidal activity of endolysins may
   be limited.

IMPORTANCE

   Recombinant GSDMD[Nterm] protein was able to efficiently permeabilize
   P. aeruginosa outer membranes and increase endolysin activity against
   bacteria, producing either long LPS O-chains or lack them entirely. The
   obtained results suggest the limited possibility of using the natural
   process of pyroptosis occurring in monocytic cells to enhance the
   bactericidal effect of recombinant phage endolysins against
   Gram-negative bacteria infection.

   KEYWORDS: Pseudomonas aeruginosa LPS, pyroptosis, gasdermin D,
   endolysin, outer membrane permeabilization

INTRODUCTION

   Bacterial resistance to antibiotics is a major problem facing society
   today. Antibiotic-resistant (AMR) bacteria are one of the most serious
   global threats to humans, which results in the depletion of treatment
   options, a greater risk of the spread of bacteria, and a longer
   duration of infection, leading to high morbidity and mortality. AMR
   bacteria, especially Gram-negative Pseudomonas aeruginosa, are
   responsible for the majority of nosocomial infections, but the number
   of community-acquired infections that are hard to cure is also
   increasing ([40]1). The burden of AMR infections forces the search for
   new alternative treatment solutions. Several alternative antibacterials
   or virulence-arresting drugs are considered, including nanoparticles,
   probiotic organisms, bacteriophages, antimicrobial peptides,
   immunomodulators, and antibodies ([41]2[42]–[43]6). According to the
   World Health Organization Report on Antibacterial Agents In Clinical
   and Preclinical Development: An Overview and Analysis (27 May 2022;
   [44]https://www.who.int/publications/i/item/9789240047655), lytic
   bacteriophages and phage-derived enzymes have been included in five
   main groups of non-traditional antibacterials in the pipeline.

   One of the most promising is recombinant phage endolysins often used in
   combination with other drugs or proteins. Endolysins are enzymes
   hydrolyzing peptidoglycan (PG) and causing bacterial cell lysis. They
   are necessary for the release of new phage progeny from the infected
   bacteria ([45]7, [46]8). Depending on the peptidoglycan cleavage
   mechanism, we may distinguish five classes of endolysins: lysozymes
   (N-acetylmuramidases), lytic transglycosylases, glycosidases
   (N-acetyl-β-D-glucosamidases), N-acetylmuramyl-L-alanine amidases, and
   L-alanoyl-D-glutamate endopeptidases. Endolysins are a potential
   antibacterial tool for digesting the peptidoglycan layer and have the
   advantage of rarely inducing resistance due to their targeting of the
   conservative bacterial structure ([47]8, [48]9). Most endolysins
   originating from Gram-positive-specific phages are modular and composed
   of enzymatic domain/s (EAD) cleaving PG and a specific cell
   wall-binding domain/s (CBD) and can be used externally to digest the PG
   layer ([49]10). Endolysins used by phages targeting Gram-negative
   bacteria are mostly globular with EAD not effective when applied
   without ([50]11). The PG of Gram-negative bacteria is hidden beneath
   the bacterial outer cell membrane (OM), which constitutes a protective
   barrier limiting the access of endolysins. There are several reports on
   the effectiveness of endolysins, such as LysSS ([51]12), Lys68, or
   PlyPAJD-1, against Gram-negative bacteria, including P. aeruginosa,
   especially when combined with cell membrane-disrupting agents (EDTA,
   citric acid, malic acid, or nigericin), which are able to depolarize
   and permeabilize the OM ([52]13[53]–[54]15). The disturbance of the
   integrity of bacterial OM can be also achieved by the application of
   nanoparticles ([55]16, [56]17), cationic dendrimers ([57]18, [58]19),
   or antimicrobial peptides and proteins ([59]20[60]–[61]24).

   An interesting approach is to engage a natural process called
   pyroptosis, which makes pores in the OM, enabling the phage endolysin
   to reach the PG. Pyroptosis is an inflammatory type of programmed cell
   death induced by stimulation of inflammasome formation in eukaryotic
   cells such as monocytes. Inflammasome complex initiates caspase-1
   (canonical) ([62]25, [63]26) or caspase-4/-5 (non-canonical) ([64]27)
   signaling and activation of gasdermin D (GSDMD) to form the active
   N-terminal part of gasdermin D (GSDMD[Nterm]) that triggers the
   formation of pores in membranes resulting in the leakage of
   inflammatory cytokines, such as interleukin (IL)-1β and IL-18. The
   pyroptosis process may be activated by bacterial lipopolysaccharides
   (LPS), which are one of the most important compounds of bacterial OM,
   and released outer membrane vesicles interacting with the immune system
   ([65]28[66]–[67]30). The LPS consists of the core part of the
   oligosaccharide, the O-antigen, and conserved lipid A, which is
   responsible for its toxicity (endotoxin) ([68]31). LPSs may activate
   extracellular receptors Toll-like receptor 4 (TLR4) or intracellular
   receptors nucleotide-binding oligomerization domain, leucine-rich
   repeat and pyrin domain-containing protein 3 (NLRP3) inflammasome,
   which stimulate an inflammation process ([69]32, [70]33). It has been
   reported that after bacterial infection and activation of the
   pyroptotic process, GSDMD can be released outside through the pores and
   deposited in the OM leading to bacterial lysis ([71]34). This process
   could be used in conjunction with phage endolysin to enhance the
   killing effect on bacteria.

   The goal of the presented study was to investigate whether P.
   aeruginosa LPS can induce pyroptosis in selected tissue cells and
   whether it is linked with GSDMD production. If yes, that process could
   be considered as a natural permeabilization of P. aeruginosa OM and
   might enhance the transport of phage endolysin to periplasmic space for
   PG degradation, triggering/increasing bacterial lysis during the
   treatment with endolysin as an antibacterial product.

   To verify the aforementioned hypotheses, in the first step of the
   study, we analyzed the OM permeabilization efficacy of the recombinant
   GSDMD and GSDMD[Nterm] on P. aeruginosa PAO1 and the ability of
   gasdermin to increase the recombinant endolysin antibacterial activity.
   The second part focused on pyroptosis activation and GSDMD expression
   in different eukaryotic cell lines in the presence of standard
   commercially available P. aeruginosa LPS. This was done at genetic and
   protein levels to investigate the susceptibility of particular cells to
   being induced by LPS molecules.

   Further, the activation of pyroptosis was performed with previously
   immuno-characterized LPS of different P. aeruginosa strains (PAO1,
   CF832, CF217, non-CF0038) to check whether there are variations in the
   proinflammation properties of LPS derived from pathogens retrieved from
   acute versus chronic infections in CF patients. Finally, we tested the
   effect of the pyroptosis process induced by selected LPS samples on the
   antibacterial activity of Klebsiella phage KP27 endolysin.

MATERIALS AND METHODS

Bacterial strains and eukaryotic cell lines

   P. aeruginosa PAO1 (ATCC 15692) wild type and its knock-out ΔwbpL
   mutant deficient in the biosynthesis of A- and B-band O-antigens,
   provided by Andrew M. Kropinski from the Laboratory of Foodborne
   Zoonoses, Guelph, ON, Canada, were used. P. aeruginosa PAO1, P.
   aeruginosa strains isolated from patients with cystic fibrosis (CF217,
   CF532-early infection, CF832-late infection, mucoid), and P. aeruginosa
   non-CF0038 strain isolated from burn wound, were from the collection of
   the Department of Pathogen Biology and Immunology, University of
   Wroclaw, as described previously ([72]35). LPSs of P. aeruginosa O10
   and E. coli 0111:B4 were purchased from Sigma-Aldrich, St. Louis, USA.

   A549 cells (ATCC, Manassas, Virginia, CCL-185), BEAS-2B cells (ATCC,
   Manassas, Virginia, CRL-9609), HeLa cells (ATCC, Manassas, Virginia,
   CCL-2), THP1-Xblue cells (Invivogen, Toulouse, France), and THP1-Null2
   cells (Invivogen, Toulouse, France) were used in our study. The A549
   cells were maintained in F-12K Nutrient Mixture (Kaighn’s modification)
   with L-glutamine (Corning, New York, USA), HeLa cells in RPMI 1640 with
   L-glutamine and NaHCO[3] (Sigma-Aldrich, St. Louis, USA), and
   THP1-Xblue and THP1-Null2 cells in RPMI 1640 ATTC modified (Gibco,
   Thermo Fisher Scientific, Waltham, USA).

   THP-1 human monocytic cells are the most commonly used model cell line
   for the study of inflammasome activation. THP1-Null2 cells are derived
   from THP-1 and produce IL-1β upon stimulation with canonical or
   non-canonical inflammasome inducers, such as nigericin or LPS.
   Moreover, they undergo pyroptotic cell death upon inflammasome
   activation. All cell lines were supplemented with 10% heat-inactivated
   fetal bovine serum (FBS; Biowest, Nuaille, France) and cultivated at
   37°C in a humidified atmosphere of 5% CO[2].

LPS isolation and pattern analysis

   An examination of the LPS structure pattern was performed using the
   slightly modified Marolda method of the classical Westphal and Jann
   protocol, utilizing the hot aqueous phenol extraction ([73]36, [74]37).
   The overnight bacterial cultures in the TSB medium (Oxoid, Basingstoke,
   UK) were centrifuged and adjusted to OD[600] equal to 2.0 in the
   phosphate-buffered saline (PBS) buffer. The bacterial cells were
   disintegrated with lysis buffer (2% SDS, 4% β-mercaptoethanol, Tris, pH
   6.8, 1,000°C, 15 min), and the proteins were enzymatically digested
   with proteinase K (20 mg/mL, 600°C, 1 h). The protein debris was
   eliminated, and the LPS was isolated by a hot aqueous phenol solution
   (90%, 70°C, 15 min.). The LPS containing the aqueous phase was purified
   using ethyl ether to remove residual phenol. To assess the
   concentration of LPS used for electrophoresis, the purpald method of
   KDO (3-deoxy-D-manno-oct-2-ulosonic acid) measurement was performed
   ([75]38). Subsequently, the LPS samples were separated by the
   Tricine-SDS-PAGE method (14% polyacrylamide gel with 4 M urea, 80 V
   constant) ([76]37). Finally, after separation, the LPS bands were
   detected by silver staining ([77]39).

Cell line propagation for LPS assay

   The THP-1 human monocytic cell line (ATCC, TIB-202, Manassas, VA, USA)
   was cultured in RPMI-1640 medium, (Lonza, BioWhittaker, Verviers,
   Belgium) supplemented with 10% heat-inactivated human serum (HiFBS,
   EuroClone, Pero, Italy), 1:100 glutamax (Gibco, Life Technologies,
   Grand Island, NY, USA), and 1:100 antibiotic–antimycotic solution
   (Gibco, Life Technologies, Grand Island, NY, USA) at 37°C in the
   presence of 5% CO[2]. The line was propagated in flasks (SARTSTEDT,
   Numbrecht, Germany). The number of viable cells was determined by a
   standard trypan blue exclusion assay using a 0.4% trypan blue solution.

   Reporter monocytes of THP1-XBlue cells (InvivoGen) expressing various
   TLRs and a nuclear factor-kappa B/AP-1-sensitive reporter gene encoding
   secreted embryonic alkaline phosphatase were cultured at 37°C in a
   CO[2]incubator in RPMI 1640 medium supplemented with 10%
   heat-inactivated fetal bovine serum (30 min at 56°C), 1:100 Glutamax,
   and 200 µg/mL of zeocin. The cells were passaged every 3 days by
   inoculating 7 × 10^5 cells/mL.

Cytokine stimulation and detection for LPS characteristics

   The method was performed according to our protocol described previously
   ([78]40). A total of 5 × 10^5 THP1 cells in 0.5 mL of medium were
   incubated with various concentrations of isolated LPS, in duplicate for
   24 h in 24-well microplates (NUNC, Thermo Fisher Scientific, Roskilde,
   Denmark) at 37°C with 5% CO[2]. After incubation, cells were
   centrifuged at 1,200 × g for 5 min, and culture supernatants were
   harvested, centrifuged, pooled, aliquoted, and stored frozen (−80°C).
   Enzyme-linked immunosorbent assays (ELISAs) for human IL-8 and tumor
   necrosis factor alpha (TNF-α), were performed using serially diluted
   samples according to the manufacturer’s instructions (DuoSet ELISA,
   R&D, Minneapolis, USA).

Hemolytic assay for LPS characteristics

   Sheep blood stabilized in Alsever’s solution was purchased from
   Proanimali (Wroclaw, Poland) and stored at 4°C before use. Of blood
   sample, 7 mL was centrifuged at 1,500 × g for 10 min, and the resulting
   plasma fraction, together with leukocytes, was discarded from the
   samples. The pellet was washed with PBS (pH 7.4), mixed by inversion,
   and centrifuged. After five washing steps, the pellet of sheep red
   blood cells (SRBC) treated as 100% was diluted to a concentration of
   1.6% (vol/vol) in veronal buffer (pH 7.4) referring to 8 × 10^8
   cells/mL. Of 1.6% SRBC, 150 µL was mixed in a 1:1 ratio with various
   concentrations of LPS in 2-mL Eppendorf tubes and incubated for up to 2
   h in a water bath at 37°C. Then, SRBC were centrifuged (1,500 × g, 5
   min, room temperature [RT]), and 200 µL of hemolytic supernatant was
   transferred to a 96-well flat-bottom microplate. One percent Triton
   X100 and veronal buffer were used as positive and negative (diluent)
   controls, respectively. The percentage of hemolysis was calculated from
   a 10-point hemolytic standard curve constructed as follows: 0.8% SRBC
   cell lysate obtained by mixing 2.4% SRBC and 2 mL of miliQ referring to
   100% hemolysis diluted serially to obtain standards of hemolysis in the
   range 10%–100%. Each standard in duplicate was applied to the 96-well
   flat-bottom microplate, and the absorbance was measured at 541 nm on a
   Varioskan Lux multimode multi-plate reader (Thermo Fisher Scientific,
   Waltham, USA).

Residual hemolytic activity of complement for LPS characteristics

    1. Sensitization of SRBC by hemolysin (rabbit IgG anti-SRBC): In a
       Falcon tube, 2.4% SRBC (vol/vol) in veronal buffer were mixed with
       2,500× diluted hemolysin in a ratio of 1:1 and incubated for 15 min
       at RT;
    2. Activation of serum complement by LPS: LPS at 1 mg/mL was
       preincubated with normal human serum (NHS) in a 1:9 ratio for 30
       min at 37°C in a water bath;
    3. The residual activity assay: The residual activity of
       LPS-pretreated NHS was determined by complement hemolytic assay
       (CH50). LPS-pretreated NHS was initially diluted 20× with veronal
       buffer and then used in the following reaction mixtures in
       Eppendorf tubes: (A) 100 µL of NHS, (B) 80 µL of NHS +20 µL of
       veronal buffer, (C) 60 µL of NHS +40 µL of veronal buffer, (D) 60
       µL of NHS +40 µL of veronal buffer. To each tube, 200 µL of
       IgG-sensitized SRBC was added, and the whole was incubated for 30
       min at 37°C in a water bath. Then SRBCs were centrifuged (1,500 ×
       g, 5 min, RT) and the extent of cell lysis was determined by
       transferring 200 µL of hemolytic supernatant to a 96-well
       flat-bottom microplate, and absorbance was measured at 541 nm on a
       Varioskan Lux multimode multi-plate reader (Thermo Fisher
       Scientific, Waltham, USA). Each assay contained a negative control
       (spontaneous hemolysis control of sensitized SRBC with buffer) and
       a positive control (hemolysis of sensitized SRBC in the presence of
       non-treated NHS). The percentage of hemolysis from tubes A, B, C,
       and D, was calculated from a 10-point hemolytic standard curve. The
       complement activity was expressed as activity units (AU) referring
       to the reciprocal of the serum dilution that yields 50% hemolysis
       of SRBC.

Complement serum bactericidal assay

   Active normal sheep serum (NSS) as a source of complement was purchased
   from the provider of animal biospecimens ProAnimali (Wroclaw, Poland)
   and stored at −80°C. To determine sensitivity to serum, bacteria from
   overnight liquid culture at TSB were centrifuged, washed with saline,
   and diluted to ∼5 × 10^5 CFU/mL of saline. The 100 µL of bacterial
   suspension was mixed at a 1:1 vol/vol ratio with active NSS or
   heat-inactivated NSS (56°C, 30 min) serum to obtain 50% serum
   concentration and incubated for 2 h at 37°C in a water bath. At times
   0, 1, and 2 h, bacteria were serially diluted and plated (10 µL) on
   Tryptic Soy Agar plates for CFU/mL calculation.

Recombinant endolysin preparation

   The recombinant phage-borne endolysin was prepared according to the
   method described previously by Maciejewska et al. ([79]41). Briefly,
   the coding sequence of Klebsiella phage KP27 endopeptidase (endolysin)
   was amplified using Pfu polymerase (ThermoFisher Scientific, Waltham,
   MA, United States) and cloned into the commercially available
   pEXP-5-CT/TOPOR TA expression vector (Invitrogen, Thermo Fisher
   Scientific, Waltham, USA) according to the manufacturer
   recommendations. The expression was conducted for 18 h at 20°C using E.
   coli BL21 (DE3) pLysS (Agilent Technologies, Santa Clara, CA, United
   States) and isopropyl-b-D-1-thiogalactopyranoside (final concentration
   of 0.5 mmol/L) as an inductor of the expression. The recombinant
   protein was purified from the filtered supernatant by affinity
   chromatography using NGC medium pressure chromatography systems
   (Bio-Rad, Hercules, CA, United States) combined with 5-mL nickel
   columns using Bio-Scale Mini Profinity IMAC Cartridges (Bio-Rad,
   Hercules, CA, United States) and dialyzed against a PBS buffer. The
   concentration of purified recombinant enzyme was then determined
   fluorometrically (QubitR Protein Assay Kit, Molecular Probes, Thermo
   Fisher Scientific, Waltham, MA, United States).

Permeabilization of the bacterial OM by GSDMD

   Permeabilization of P. aeruginosa strains PAO1 and its knock-out ΔwbpL
   mutant was tested after adding commercially available GSDMD protein
   (Origene, Rockville, USA; [80]NM_024736) or GSDMD[Nterm] obtained after
   incubation with 0.1 U of caspase 4, and without/with 36 µg/mL of
   Klebsiella phage KP27 endopeptidase. The experiment was performed using
   the N-phenyl-1-naphthylamine (NPN) probe uptake assay. The bacteria
   were grown at 37°C to obtain the appropriate optical density (OD[600] =
   0.5), and the bacterial pellet after centrifugation was dissolved in
   HEPES buffer (correspondingly 1 mL of the buffer per 0.5 mL of bacteria
   before centrifugation) to obtain OD[600] = 0.1. The measurement was
   performed in a sterile 96-well plate. The NPN fluorescent dye was added
   to the bacterial suspension in the well (10 µM per well). Then,
   appropriate concentrations of GSDMD and endolysin were added to the
   mixture. NPN fluorescence intensity was monitored for 30 min with a
   time interval of 1 min at an excitation wavelength of 350 nm and an
   emission wavelength of 420 nm using a TECAN Infinite 200 PRO microplate
   reader (Tecan Group Ltd., Switzerland). The results were presented in
   relative fluorescence units (RFU).

Quantitative reverse transcription PCR (RT-qPCR)

   RT-qPCR was used to test the effect of P. aeruginosa O10 LPS on GSDMD
   gene expression in THP1-Xblue cells. The cells were incubated with LPS
   at a concentration range of 0.001–100 µg/mL for 24 h in a humidified
   atmosphere of 5% CO[2]. Total RNA was extracted from THP1-Xblue and
   THP1-Null2 using SV Total RNA Isolation System (Promega, Madison, USA),
   RNeasy Plus Mini Kit (Qiagen, Hilden, Germany), or using TRI REAGENT
   (Sigma-Aldrich, St. Louis, USA) form three independent repeats. The
   amount of RNA was assessed quantitatively and qualitatively using
   Nanodrop Spectrophotometer (Thermo Scientific, Waltham, USA), where the
   RNA quality for all samples was between 1.8 and 2.0, and by RNA
   integrity number (RIN), where the value of RIN was between 8 and 10 for
   all samples using Bioanalyzer (Agilent, St. Clara, USA). Reverse
   transcription was performed using the Reverse Transcription System
   (Promega, Madison, USA) following the manufacturer’s protocol with
   random hexamer primers. One microliter of cDNA solution of each sample
   was used for qPCR in a 20-µL reaction mixture containing the following:
   10 µL of 2× SsoAdvanced Universal SYBR Green Supermix (Bio-Rad,
   Hercules, USA), 1 µL of PrimePCR SYBR Green Assay: GSDMD (Bio-Rad,
   Hercules, USA), 1 µl of PrimePCR SYBR Green Assay: ACTB (β-actin)
   (Bio-Rad, Hercules, USA), and 8 µL of dH[2]O. The relative expression
   level of β-actin was used as a control. The following conditions were
   used: 95°C for 2 min. by one cycle, 95°C for 5 s, and 60°C for 30 s by
   40 cycles using the equipment Eco Real-Time PCR (Illumina, San Diego,
   USA). Experiments were done in three independent repeats. ACTB gene was
   used as an internal reference gene, and the ΔΔCT method was used for
   relative quantification.

DNA microarray assay

   Total RNA was extracted from A549 and HeLa cells using the isolation
   kits Sv Total system RNA Isolation (Promega, Madison, USA) and Rneasy
   Plus Mini Kit (Qiagen, Hilden, Germany), and from THP1-Xblue cells
   using TRI REAGENT (Sigma-Aldrich, St. Louis, USA) from three
   independent replicates according to the manufacturer’s protocol. The
   quality and quantity of RNA were checked using the Nanodrop
   Spectrophotometer (Thermo Scientific, Waltham, USA) and by the RNA
   integrity number (Agilent, St. Clara, USA) as described in Quantitative
   reverse transcription PCR (RT-qPCR). Two-Color Microarray-Based Gene
   Expression Analysis assay (Agilent, St. Clara, USA) was used for the
   analysis of gene expression in A549, HeLa, and THP1-Xblue cells treated
   with 0.1 µg/mL of P. aeruginosa O10 LPS in comparison to non-treated
   control, according to the manufacturer’s instruction. The expression of
   RNA was measured using DNA chips of Agilent numbers
   US11233913_257236328396_S02_GE2_1010_Sep10,
   US11233913_257236328397_S02_GE2_1010_Sep10, and
   US11233913_257236328398_S01_GE2_1010_Sep10, which contained probes for
   approximately 56,000 human genes along with long non-coding RNA
   (lncRNA). The experiment was three times replicated as biological
   replicates. The analysis of Microarray data was made using the special
   gene expression software GeneSpring GX suitable for DNA Microarray
   instruments from Agilent.

   Total RNA was amplified and labeled using Low Input QuickAmp Labeling
   Kit Two-Color (Agilent Technologies, Santa Clara, CA, US), following
   the manufacturer’s instructions. The labeled cDNA was synthesized from
   50 mg of total RNA using 5× first-strand buffer, 0.1 M DTT, 10 mM dNTP
   mix, and Affinity Script RNase Block Mix according to the
   manufacturer’s procedures. The samples were incubated at 40°C for 2 h,
   and the reverse transcription and dsDNA synthesis were terminated by
   incubating at 70°C for 15 min. The dsDNA transcription was performed by
   adding Transcription master mix (5× Transcription buffer, 0.1 M DTT,
   NTP mix, T7-RNA polymerase, and Cyanine 3/5-CTP thereby incorporating
   Cy3 [control samples] or Cy5 [test samples]-labeled dCTPs) to the dsDNA
   reaction samples and incubating at 40°C for 2 h. Amplified and labeled
   cRNA was purified on an RNase mini column (Qiagen) according to the
   manufacturer’s protocol. After checking the labeling efficiency, the
   cyanine-3- and cyanine-5-labeled cRNA were mixed, and cRNA
   fragmentation was performed by adding 10× Gene Expression Blocking
   Agent and 25× Fragmentation Buffer and incubating at 60°C for 30 min.
   Fragmented cRNA was resuspended in 2× Hi-RPM Hybridization Buffer and
   directly pipetted onto the SurePrint G3 Human Array Microarray (Agilent
   Technologies). The array was hybridized at 65°C for 17 h using a
   hybridization oven (Agilent Technology, USA). The hybridized
   microarrays were washed as per the manufacturer’s washing protocol
   (Agilent Technologies). The hybridization images were analyzed by a DNA
   microarray scanner (Agilent Technologies).

   The analysis of microarray data was performed using GeneSpring GX
   software (Agilent, St. Clara, USA). The statistical analyses of the
   microarray assay were calculated by the gene expression software
   GeneSpring GX (Agilent, St. Clara, USA). Data were interpreted by
   comparison with a control group using a t-test with Benjamin–Hochberg
   false discovery rate (FDR) correction, and the significance level was
   set at P < 0.05. Gene ontology was analyzed using the Quick GO, and
   pathway analysis was performed based on the Reactome databases with
   Benjamin–Hochberg FDR correction and P < 0.05. The Panther databased
   was used for establishing significantly changed molecular function and
   biological processes. Additionally, genes showing statistically
   significant changes (P < 0.05) obtained by comparing the expression
   between LPS-exposed and non-exposed cell lines were analyzed using the
   PathfindR R package. The PathfindR package uses an active
   subnetwork-oriented pathway enrichment analysis to place a set of genes
   of interest in the pathway context
   ([81]https://pubmed.ncbi.nlm.nih.gov/31608109/).

ELISA for pyroptosis detection

   THP1-Xblue cells were seeded at a density of 10^6 cells per well. The
   cells were incubated with P. aeruginosa O10 LPS in the concentration
   range of 0.001–1 µg/mL. The medium was collected after 24 h. The
   supernatants were used for the ELISA test according to the manufacturer
   protocol (Human IL-1 beta ELISA Kit; Abcam, Cambridge, Great Britain).
   The concentration of IL-1β in the supernatants was calculated according
   to standard curves. The results of the amount of IL-1β are shown in
   μg/mL of supernatant. Experiments were done in three independent
   repeats.

Western blot

   In the first step, THP1-Null2 cells were pre-incubated with 1, 10, or
   100 µg/mL of E. coli 0111:B4 LPS for 3 h; incubated with nigericin at
   10 µM for 24 h; and supernatants were collected to check the
   possibility of GSDMD induction. In the second part of the study,
   THP1-Null2 cells were pre-incubated with 1 µg/mL of P. aeruginosa O10,
   PAO1, CF532, CF832, CF217, and non-CF0038 LPS; incubated with nigericin
   at 10 µM for 24 h; and the supernatants were collected. Western blot
   technique was performed according to Bio-Rad (Hercules, USA) company
   protocol, using their chemicals. Approximately 20 µL of supernatant
   with 20 µL of Laemmil protein sample buffer were loaded on Mini-PROTEAN
   TGX Gel 4%–20% SDS-PAGE for electrophoresis followed by
   electro-transfer to nitrocellulose. The membranes were blocked in 3%
   bovine serum albumin in TBST (TRIS-buffered saline with Tween 20
   buffer) and then incubated overnight with the following primary
   antibodies: anti-GSDMD (GSDMDC1, sc-81868, Santa Cruz Biotechnology,
   Dallas, USA) and anti-glyceraldehyde-3-phosphate dehydrogenase
   (anti-GAPDH; 1:1,500 dilution; Abcam, Cambridge, Great Britain; ab8245
   and ab9485) at 4°C. After washing with TBST, the membranes were
   incubated in HRP (horseradish peroxidase)-conjugated secondary antibody
   solution (1:3,000 dilution; Abcam, Cambridge, Great Britain; ab6721)
   and RBTXMS IGG (H + L) HR (1:200 dilution; Life Technologies, Carlsbad,
   USA; 31450) for 1 h at room temperature. Signals were detected with
   chemiluminescence. GAPDH was used as a control protein to normalize the
   target protein. As a positive control, 0.75 µM of GSDMD recombinant
   protein (Origene, Rockville, USA; [82]NM_024736) was used.

Antimicrobial activity assay

   The THP1-Null2 cell line was treated with a final concentration of 1
   µg/mL of LPS of P. aeruginosa (PAO1, CF532, CF832, CF217, non-CF0038)
   and of E. coli 0111:B4 and/or 10 µM nigericin (pyroptosis positive
   control). Then, the THP1-Null2 cell line supernatants, containing GSDMD
   oligomers, were used for antimicrobial activity assay. It was tested
   alone or in combination with KP27 endolysin against the P. aeruginosa
   PAO1 strain. In detail, a bacterial culture of 10^4 CFU/mL in TSB
   medium was incubated for 18 h at 37°C in the presence of 2× diluted
   cell line supernatant alone or with 36 µg/mL of KP27 endolysin. The
   effect was measured using OD[600]. The experiments were done in three
   biological repeats. One-way analysis of variance (ANOVA) with post-hoc
   Dunnett’s test was used for statistics.

Statistical analysis

   STATISTICA version 13.3 (StatSoft, Poland) was used for statistical
   data analyses of RT-qPCR, ELISA, and antibacterial studies. The normal
   distribution of variables was verified using the Shapiro–Wilk test,
   whereas the homogenicity of variance was tested by the Brown–Forsyth
   test. Data from RT-qPCR and antibacterial studies were established
   using ANOVA with post-hoc Dunnett’s test. Results from ELISA were
   analyzed using ANOVA with post-hoc Tukey test. The results are
   expressed as the mean ± standard deviation (SD). The significance level
   was considered for P < 0.05 (*).

RESULTS

   The study was designed in four steps to investigate the natural
   pyroptosis process as an anti-pseudomonal immune response potentially
   supporting the antibacterial activity of PG-degrading endolysin derived
   from phages ([83]Fig. 1).

Fig 1.

   [84]Flowchart outlines four experimental steps to study gasdermin D,
   bacterial outer membrane permeabilization, LPS-induced pyroptosis, and
   inflammatory properties.
   [85]Open in a new tab

   The experimental pipeline of the study ([86]1). Recombinant gasdermin D
   (GSDMD) or its N-terminal region for membrane permeability of P.
   aeruginosa PAO1 and its LPS O-chain-deficient mutant (ΔwbpL) was
   monitored using NPN uptake assay. Membrane permeability experiments
   were performed for GSDMDs alone and in the presence of KP27 endolysin
   ([87]2). Mimicking bacterial infection—gasdermin D activation by P.
   aeruginosa O10 LPS in A549, HeLa, or THP1 cells was measured by
   microarray (gene expression) and ELISA (IL-1β) assays ([88]3). The
   inflammatory properties of LPS isolated from E.coli (control) and
   tested P. aeruginosa strains were evaluated using immune-serological
   tests. The patterns of isolated LPS samples (the O-chain presence or
   absence) were characterized using the SDS-PAGE method ([89]4).
   LPS-induced pyroptosis alone or in combination with nigericin was
   tested on the THP1 cell line to obtain active gasdermin D for enhanced
   KP27 endolysin activity against P. aeruginosa PAO1 strain. LPS was
   isolated from tested P. aeruginosa strains and E. coli (control).
   Gasdermin (GSDMD) production was monitored using Western blot, and the
   antibacterial activity of endolysin was evaluated using culture optical
   density (OD[600]).

Gasdermin is able to permeabilize P. aeruginosa OM, but only GSDMD[Nterm]
increases endolysin activity

   To study whether GSDMD affects phage endolysin activity against P.
   aeruginosa, the fluorescent NPN probe uptake method was used. This
   assay is based on the assumption that intact OM is a permeability
   barrier blocking the influx of hydrophobic substances such as NPN. Once
   the OM is damaged, the NPN enters the phospholipid layer resulting in
   prominent fluorescence ([90]42). This step verified the
   membrane-permeabilizing activity of commercially available recombinant
   GSDMD and GSDMD[Nterm] obtained after cleavage by caspase-4, while
   combined or not with phage KP27 endolysin. PAO1 and its ΔwbpL mutant
   were used to check whether the O-chain presence or absence affects
   GSDMD interaction with bacterial membranes and the formation of pores.
   Our results showed that both GSDMD and GSDMD[Nterm] used alone were
   able to permeabilize P. aeruginosa PAO1 wild-type cells with smooth LPS
   ([91]Fig. 2). This effect was potentiated for GSDMD[Nterm] and weakened
   for GSDMD in the presence of KP27 endolysin. For the ΔwbpL mutant cells
   lacking the O-chain, only GSDMD[Nterm] was able to permeabilize
   bacterial membrane ([92]Fig. 2), and this effect was enhanced in
   combination with KP27 endolysin. It suggested that GSDMD and
   GSDMD[Nterm] might have different mechanisms for membrane
   permeabilization. GSDMD can increase the membrane’s permeability via
   electrostatic/hydrophilic interactions promoted by LPS O-chain, which
   is disturbed in the presence of endolysin. Moreover, the lack of
   membrane permeabilization in the ΔwbpL mutant (R form) confirmed the
   role of O antigen in GSDMD activity. In contrast, GSDMD[Nterm] could
   permeabilize the membrane of both S and R bacteria and this effect was
   potentiated in the presence of endolysin.

Fig 2.

   [93]Line graph plots of relative fluorescence units over time to
   compare P. aeruginosa wild-type (indicated by triangles) and mutant
   (indicated by circles) strains treated with GSDMD, endolysin, and
   combinations indicated by circles.
   [94]Open in a new tab

   OM permeabilization of P. aeruginosa PAO1 wild type with LPS (S) and
   its ΔwbpL mutant lacking LPS O-chain (R) by 1.5 µM GSDMD or
   GSDMD[Nterm]recombinant proteins alone or in combination with endolysin
   (36 µg/mL) calculated as the relative fluorescence units (RFU). The
   control consisted of untreated P. aeruginosa cells.

P. aeruginosa LPS induces various responses in different eukaryotic cell
lines with the strongest upregulation in monocytic cells

   After the confirmation that extracellularly supplied GSDMD and
   GSDMD[Nterm] can form pores in P. aeruginosa OM, we next verified
   whether GSDMD activation is possible in a natural pyroptosis process
   via P. aeruginosa LPS. For this purpose, the following three cell types
   that potentially could be exposed to P. aeruginosa in the infection
   process were chosen: the well-defined monocytic cell line (THP1-Xblue),
   epithelial human lung cells A549, and epithelial human uterus/cervix
   cells HeLa ([95]43[96]–[97]46). All cell lines were treated with 0.1
   µg/mL of standardized commercial LPS of P. aeruginosa O10 serotype, and
   the gene expression profile was studied by microarray assay as well as
   RT-qPCR technique.

   Using a Two-Color Microarray-Based Gene Expression Analysis, 50
   differentially expressed genes (DEGs) were identified with significant
   fold changes (P ≤ 0.05) in A549 cells ([98]Fig. 3A and B), and the
   pathway enrichment analysis (PahthfindeR, R package) based on comparing
   expression between LPS-exposed and non-exposed A549 cells line is
   presented in [99]Fig. 3C. The analysis using the Reactome software
   showed only the SSBP3-AS1 gene with an upregulated expression, which
   encodes a putative uncharacterized protein with an unknown function.
   The gene ontology (GO) terms of downregulated genes were associated
   with the G protein-coupled receptor signaling pathway, positive
   regulation of calcium ion import, and positive regulation of MAP kinase
   activity (kappaB kinase/NF-kappaB signaling positive regulation of ERK1
   and ERK2 cascade chemokine-mediated signaling pathway, cellular
   response to interferon-gamma, immune response, and protein binding
   ([100]Table S1, Supporting Data). Generally, the response of A549 to
   LPS was low because epithelial cells do not express the CD14 receptor
   engaged in the LPS signal transduction pathway ([101]47, [102]48). In
   contrast, the enrichment of cancer-associated pathways (melanoma,
   glioma, prostate cancer, central carbon metabolism in cancer, and
   choline metabolism in cancer) and viral infection (viral life cycle-
   HIV-1) was observed in the A549 cell line ([103]Fig. 3C).

Fig 3.

   [104]Venn diagrams compare overlapping and unique gene expression
   patterns among the cell lines. Bar chart visualizes enriched pathways
   across the cell lines. Heatmaps display changes in specific genes
   across conditions.
   [105]Open in a new tab

   Two-color microarray-based gene expression analysis of cell lines
   treated with P. aeruginosa O10 LPS. The number of up- (A) and
   downregulated (B) genes in tested cell lines. Heatmaps of KEGG pathway
   enrichment analysis (PahthfindeR, R package), based on comparison of
   expression between LPS-exposed and non-exposed cell lines, are
   presented for A549 (C), HeLa (D), and THP (E).

   For HeLa cells, 4,740 genes (351 upregulated, 4,389 downregulated) were
   identified with significant fold changes (P ≤ 0.05) ([106]Fig. 3A and
   B). The analysis of GO terms for HeLa cells using the Reactome software
   revealed upregulated expression of the G protein-coupled receptor
   signaling pathway, G protein-coupled receptor activity, DNA binding,
   regulation of transcription, ion transport and binding, and negative
   regulation of the apoptotic process. Only two genes were related to GO
   terms such as innate immune response, antibacterial humoral response or
   immunoglobulin production involved in immunoglobulin-mediated immune
   response ([107]Table S2, Supporting Data). The GO terms of
   downregulated genes were associated with ion binding and transport,
   negative regulation of inflammatory response, response to transforming
   growth factor beta, and protein binding ([108]Table S3, Supporting
   Data). The level of gene downregulation was 10 times higher than its
   upregulation in HeLa cells evoked by LPS. The analysis using
   PahthfindeR (R package) of the pathway enrichment shows that mainly
   immune system response mechanisms are downregulated in the presence of
   P. aeruginosa O10 LPS ([109]Fig. 3D). It seems possible that LPS
   decreases the production of cellular components involved in the
   immunity of HeLa cells. Moreover, enrichment of pathways related to the
   immune system (B-cell receptor signaling pathway, neutrophil
   extracellular trap formation, T-cell receptor signaling pathway,
   cytokine–cytokine receptor interaction, complement, and coagulation
   cascades) and intercellular junctions (focal adhesion, tight junctions)
   were blocked in the HeLa line.

   In the case of THP1-Xblue cells, 493 genes (413 upregulated, 80
   downregulated) were identified with significant fold changes (P ≤ 0.05)
   ([110]Fig. 3A and B). The top 15 upregulated genes in THP1-Xblue with
   the highest fold change are listed in Table S4 (Supporting Data), which
   include CXCL13, CXCL3, CCL2, IL1β, RGS1, MIR146A, TNFAIP6, CCL8, CXCL8,
   MPP9, CCL7, IL6, CXCL1, SOCS3, and CCL20. These genes are generally
   associated with a cellular and inflammatory response to LPS, including
   IL-1β specific to pyroptosis. The top 15 downregulated genes in
   THP1-Xblue with the highest fold change are listed in [111]Table S5
   (Supporting Data), which include MS4A3, SERPINB10, GPA33, CLEC12A,
   CNR1, DEPDC4, SLC40A1, TMOD1, DSC3, NMUR1, RET, CLRN1, F13A1, SERPINI2,
   and MS4A6A. These genes encode integral components of
   membrane-associated proteins generally involved with signal
   transduction, signaling, disease, metabolism of protein, metabolism of
   RNA, hemostasis, immune system, and metabolism. Moreover, gene lists
   obtained by comparing gene expression of THP1-Xblue cells exposed to
   LPS compared with that of the same cell line without this exposure were
   analyzed using the Pathfinder R package ([112]Fig. 3E). For the
   THP1-Xblue cell line, enrichment of pathways related to the immune
   system activity (NF-kappa B signaling pathway, TNF signaling pathway,
   chemokine signaling pathway, and cytokine–cytokine receptor
   interaction) and viral or protozoan infection (hepatitis B,
   leishmaniasis, Kaposi sarcoma-associated herpes virus infection, viral
   carcinogenesis, and viral protein interaction with cytokine and
   cytokine receptor) was observed.

   Taking into account both analyses by the Reactome software and
   Pathfinder R package, the activation of pathways associated with the
   action of the immune system in response to LPS can be observed mainly
   in monocytic cells. The heatmaps with the general gene expression
   profile, including all upregulated and downregulated genes in
   THP1-Xblue cells generated from DNA microarray data and GeneSpring GX
   software, are presented in [113]Fig. S1 (Supporting Data).

   The DNA microarray analyses showed that the tested LPS induced the
   strongest upregulated immune response in THP1 cells, including
   inflammatory response to LPS (IL-1β) specific to pyroptosis, compared
   to epithelial A549 or HeLa cell lines. Nevertheless, the gene
   expression profile from the DNA microarray assay did not show
   statistically significant GSDMD gene overexpression relative to the
   control. We hypothesized that there is a specific mechanism for
   limiting GSDMD transcript expression or GDSMD degradation to protect
   the cell against self-lysis caused by multiple pore formation
   regardless of LPS concentration. To double check that phenomenon, the
   RT-qPCR method was utilized for the GSDMD gene expression in THP1-Xblue
   cells treated with serial concentrations of P. aeruginosa O10 LPS
   ([114]Fig. 4A). The RT-qPCR confirmed the increased GSDMD gene
   expression level in monocytic cells but only for ≥0.1-µg/mL
   concentrations of P. aeruginosa O10 LPS. It might indicate that a
   detectable level of GSDMD transcript is an effect of limited mRNA
   expression or its degradation to protect cells from self-lysis via
   pyroptosis.

Fig 4.

   [115]Bar graphs display effect of P. aeruginosa O10 LPS on GSDMD
   expression and IL-1 beta production at various concentrations. Error
   bars in both graphs represent standard deviations, indicating
   variability in measurements.
   [116]Open in a new tab

   The induction of pyroptosis in monocytic cells by P. aeruginosa O10
   LPS. The GSDMD expression level analysis by RT-qPCR in THP1-Xblue cells
   after treatment with serial concentrations of O10 LPS. The ACTB gene
   was used as an internal reference gene, and the ΔΔCT method was applied
   for relative quantification (*P < 0.05) (A). The IL-1β production by
   THP1-Xblue cells in the presence of P. aeruginosa O10 LPS measured
   using ELISA; *P < 0.05 (B). One-way ANOVA with Dunnett’s post-hoc test
   was used to test the statistically significant differences.

   Simultaneously as a pyroptosis marker, the level of IL-1β was
   determined by ELISA test in the post-culture medium of THP1-Xblue cells
   treated with LPS. It turned out that the induction of IL-1β production
   in THP1-Xblue cells was seen for ≥0.001 µg/mL of P. aeruginosa O10 LPS
   ([117]Fig. 4B). It seems important that LPS at lower concentrations was
   able to induce IL-1β production in contrast to GSDMD mRNA expression in
   THP1-Xblue cells. This effect might also be explained by the
   limitations of ELISA assay, which, in fact, may detect biologically
   inactive precursors (pro-IL-1β) ([118]49).

Inflammatory properties of LPS derived from clinical strains

   In this step, we checked whether LPS derived from various P. aeruginosa
   strains, including those from clinical sources (acute infections and
   cystic fibrosis), could differ in the inflammatory properties and thus
   potentially also in the ability to induce GSDMD synthesis by monocytic
   cells via induction of pyroptosis. Several LPS features impacting
   immunoreactivity were tested for selected strains (PAO1, non-CF0038,
   CF217, CF532, and CF832). Their LPS differed in terms of the following:
   (i) the presence of long O chain and most strains had smooth type (S),
   while CF217 strain lacked O antigen—rough type (R) ([119]Fig. S2), (ii)
   hemolytic properties with LPS from CF217 as the strongest inducer of
   hemolysis ([120]Fig. S3), and (iii) sequestration of complement
   components with LPS from strain CF217 as the weakest complement
   consumer ([121]Table S6). The full activity of serum complement against
   CF217 was confirmed in a 50% serum-dependent bactericidal assay showing
   its complement-sensitive phenotype ([122]Table S7).

   Likewise, the common proinflammatory properties of the LPS tested,
   including the induction of TNF and IL-8 release by THP-1, were
   dependent on the LPS nature/origin ([123]Fig. 5).

Fig 5.

   [124]Bar graphs display IL-8 and TNF release from THP-1 cell line in
   response to LPS from different P. aeruginosa strains at 1- and 10-μg/mL
   concentrations. Error bars represent SEM, highlighting variability in
   the data.
   [125]Open in a new tab

   Induction of inflammatory response of THP-1 monocytes by P. aeruginosa
   LPS. Supernates from 24-h THP-1 cultures stimulated with various
   concentrations of LPS were evaluated using ELISA for the presence of
   proinflammatory cytokines (A) IL-8 and (B) TNF. Data are expressed as
   mean ± SEM from at least two independent experiments performed in two
   plates at four technical repeats each (two technical repeats per
   plate). Supernates from duplicates from each plate were pooled and
   assessed using ELISA. Dashed lines refer to non-stimulated control.
   One-way ANOVA with post-hoc Tukey test was used to test the
   differences. *P < 0.05, **P < 0.01, ***P < 0.0005. As a positive
   control, LPS from O10 strains was used.

   A comparison of LPS-dependent IL-8 release showed significant
   differences depending on the source of LPS used, but only at the higher
   (10 µg/mL) concentration of endotoxin ([126]Fig. 5A). In this case, the
   strongest stimulator was LPS from the strain CF217, and the weakest was
   LPS from the non-mucoid CF532 strain. Analogous one-way ANOVA analyses
   of TNF also showed significant differences for higher concentrations of
   LPS used; however, the results indicated that, of the clinical
   isolates, only the rough LPS CF217 had the strongest pro-inflammatory
   properties, whereas proinflammatory potential of other LPS was at a
   similar level ([127]Fig. 5B). At lower 1-µg/mL concentrations of LPS,
   no differences were observed between the LPS tested in the induction of
   pro-inflammatory cytokines. Compared to the unstimulated control, the
   presence of TNF in the culture supernatant was sevenfold higher, while
   IL-8 was 100-fold higher.

   The highest inflammatory response was shown for commercial control O10
   LPS. Summing up, LPS from various clinical strains in higher
   concentrations have different immunoreactive and proinflammatory
   properties as reflected by IL-8 release.

LPS-induced pyroptosis for enhanced endolysin activity

   To study the potential variation in propyroptotic activity of the
   aforementioned LPS and influence of LPS-induced GSDMD release on
   enhanced endolysin activity, we investigated the antibacterial effect
   of recombinant Klebsiella phage KP27 endopeptidase against P.
   aeruginosa PAO1 in the presence of post-LPS-induced THP1 culture media.
   P. aeruginosa PAO1 culture reduction was measured by the optical
   density at 600 nm ([128]Fig. 6). KP27 endolysin in post-culture THP1
   medium without LPS stimulation showed moderate but significant
   antibacterial effect against PAO1 bacterial cells (P = 0.021). It
   turned out that THP1-Null2 post-culture media after stimulation by
   tested LPS samples, alone or in combination with nigericin, did not
   significantly increase the antibacterial effect of endolysin in
   comparison to the relevant controls indicating at least an insufficient
   level of pore-forming GSDMD forms.

Fig 6.

   [129]Bar graph along with statistical table display absorbance at 600
   nm for bacterial strains treated with LPS, nigericin, and endolysin
   combinations.
   [130]Open in a new tab

   The OD[600] of P. aeruginosa PAO1 culture growth was measured in TSB
   supplemented with 2× diluted THP1-Null2 cell line supernatant, either
   alone or with 36 µg/mL of KP27 endolysin (E) after 18 h of incubation
   at 37°C. The THP1-Null2 line was first pre-treated with 1 µg/mL of LPS
   (S) and/or 10 µM nigericin (N). C, bacterial culture treated with
   THP1-Null2 supernatant; C[E], bacterial culture treated with THP1-Null2
   supernatant and endolysin; C[N], bacterial culture treated with
   THP1-Null2 supernatant after nigericin stimulation; C[EN], bacterial
   culture treated with THP1-Null2 supernatant after nigericin stimulation
   and supplemented with endolysin; S, bacterial culture treated with
   THP1-Null2 supernatant after LPS stimulation; S[E], bacterial culture
   treated with THP1-Null2 supernatant after LPS stimulation and
   supplemented with endolysin; S[N], bacterial culture treated with
   THP1-Null2 supernatant after LPS and nigericin stimulation; S[EN],
   bacterial culture treated with THP1-Null2 supernatant after LPS and
   nigericin stimulation and supplemented with endolysin. Color columns
   show particular LPS samples used for pyroptosis induction. The
   experiments were done in three biological repeats, and statistical data
   were established by one-way ANOVA with post-hoc Dunnett’s test. The
   P-values are presented in the bottom table.

   To verify the presence of pore-forming GSDMD in tested supernatants,
   the Western blot technique was applied to detect the GSDMD protein and
   its cleavage products in a post-culture medium after incubation of
   THP1-Null2 with the LPS samples ([131]Fig. S4). The GSDMD was detected
   in the control growth medium, which could suggest that GSDMD might be
   produced constitutively by THP1-Null2 cells as an inactive pro-protein
   crucial for the biochemical "ready to react" system in response to
   bacterial infection (LPS) via pyroptosis. Moreover, the full GSDMD
   protein was also observed after treatment with E. coli 0111:B4 LPS at
   1–100 μg/mL ([132]Fig. S4A) and different P. aeruginosa LPS at 1 µg/mL
   ([133]Fig. S4B). Theoretically, GSDMD (484 amino acids residues; 53
   kDa) in response to LPS should be cleaved by inflammatory caspases
   (caspase-1 and -4/-5/-11) to the GSDMD C-terminus domain (GSDMD[Cterm],
   276–484 amino acids residues; 22 kDa) and the GSDMD N-terminus domain
   (GSDMD[Nterm,] 1–275 amino acids residues; 31 kDa) forming pores in the
   membrane to enable the release of inflammatory cytokines IL-1β and
   IL-18 ([134]50, [135]51).

   By testing only the supernatants in our assay, we were able to detect
   only the soluble fractions of GSDMD and GSDMD[Cterm], with the latter
   one as the indirect proof for the formation of membrane interacting
   GSDMD[Nterm] likely associated with the lipid fraction of the cells not
   tested here.

   Interestingly, in all samples and controls, GSDMD (53 kDa) and
   GSDMD[Cterm] (22 kDa) were detected. An additional band (~30 kDa) was
   noted after administration of E. coli 0111:B4 LPS alone at a higher
   concentration (100 µg/mL), and a lower concentration (10 µg/mL) post
   nigericin stimulation ([136]Fig. S4A). Moreover, ⁓40 kDa was also seen
   for E. coli 0111:B4 LPS at 100 µg/mL in combination with nigericin.
   Those bands might represent the GSDMD[Cterm] aggregates formed in
   response to increased stimulation of THP1 cells by concentrated LPS
   combined with nigericin.

   In general, the results suggest that GSDMD cleavage is observed in
   THP1-Null2 overexpressing the inflammasome system and that 1 µg/mL of
   LPS derived from E. coli or different P. aeruginosa strains did not
   stimulate monocytes to increase the formation of the pore-forming GSDMD
   domain.

DISCUSSION

   Bacteria and their components (LPS) are capable of activating the
   pyroptosis process (an inflammatory form of cell death) in the
   eukaryotic immune response ([137]52). In this process, the gasdermin D
   (GSDMD) protein is activated and cleaved to N- and C-terminal regions
   followed by GSDMD[Nterm] accumulation and pore formation in the
   membrane required for the release of cytokines such as interleukin
   IL-1β and IL-18 ([138]52[139]–[140]54). Simultaneously, a formed
   GSDMD[Nterm] serves as an antibacterial response thanks to the
   interactions with bacterial OM, pore formation, and bacterial lysis
   ([141]51, [142]55). Assuming that process, the pyroptosis leading to
   bacterial OM permeabilization may help the externally applied endolysin
   to reach the periplasmic space and peptidoglycan; thus, the combination
   of gasdermin (GSDMD) and endolysin could act synergistically in
   phage-based therapy.

   In this work, we studied whether the induction of the natural
   pyroptosis process through the activation of GSDMD could, first, be
   achieved by P. aeruginosa LPS, and, second, enhance the
   anti-pseudomonal effect of recombinant phage endopeptidase KP27.

   Initially, commercially available GSDMD and GSDMD[Nterm] were tested
   together with endolysin KP27 on two P. aeruginosa strains PAO1 and
   ΔwbpL to determine the possibility for enhancing enzyme activity
   ([143]8, [144]56). Our study showed that GSDMD alone has an
   antibacterial effect only on the wild-type PAO1 possessing the smooth
   LPS (S), but not on the O-chain-deficient mutant. The antibacterial
   activity of GSDMD might be associated with its protein structure
   containing two different antibacterial peptides. Kuand and coworkers
   showed that 80–90 amino acid residues (peptide 1) and mainly 70–90
   amino acids of peptide 2, with a flexible loop present in the GSDMD,
   can inhibit the growth of E. coli and Mycobacterium smegmatis
   ([145]50). In our study, GSDMD could permeabilize P. aeruginosa cells
   by the interaction of those peptide regions with bacterial surface
   structures. Interestingly, GSDMD-mediated bacterial cell membrane
   permeabilization in the P. aeruginosa PAO1 wild type was reduced in the
   presence of KP27 endolysin, probably by the enzyme interactions either
   with the LPS O-chain masking the available moieties for GSDMD peptides
   or directly with the GSDMD molecules.

   In contrast, we proved that the GSDMD[Nterm] domain was able to
   permeabilize both P. aeruginosa with smooth LPS as well as O-chain
   deleted mutant. GSDMD[Nterm], as a full protein (GSDMD)-cleaving
   product by caspase-1 (1–275 amino acids residues), can form pores in
   the membrane in contrast to full GSDMD protein ([146]51). The presence
   of these additional regions in GSDMD[Nterm] close to peptides 1 and 2
   might promote higher interaction with a membrane (phosphatidylserine
   and cardiolipin), which could support the activity of endolysin in
   effective membrane permeabilization. Indeed, such an effect was
   observed for both smooth WT PAO1 and ΔwbpL mutant lacking an LPS
   O-chain.

   The first line of humoral innate immune defense against Gram-negative
   bacteria regards usually the serum complement activity; therefore, most
   P. aeruginosa produces long O-chains to escape permeabilization by the
   membrane attacking complex (MAC) of an activated complement. The
   results obtained in the above experiments suggest hypothetically that
   the alternative or additional defense mechanism effective against
   bacteria with all types of LPS (long or short) is based on gasdermin D
   activity induced during pyroptosis.

   Another aspect elucidated in our study concerned the verification of
   whether GSDMD activation takes place in a pyroptosis process induced by
   P. aeruginosa LPS. Three cell types stimulated by 0.1 µg/mL of P.
   aeruginosa O10 LPS were used for transcriptomic analyses, including
   epithelial human lung cells (A549), epithelial human uterus/cervix
   cells (HeLa), and monocytic cells (THP1), to mimic bacterial infection.
   The same concentration of E. coli LPS was significant in the studies of
   Eder et al. ([147]57).

   Transcriptomic analyses revealed that the activation of pathways
   associated with the action of the immune system in response to LPS can
   be observed mainly in monocytic cells in contrast to epithelial A549 or
   HeLa cell lines. It seems reasonable that epithelial tissues are not
   intensively proinflammatory stimulated by the presence of Gram-negative
   bacteria while having constant contact with the natural microbiome.

   The analyses showed that upregulated genes of THP1 were associated with
   several significant GO terms associated with cellular response to
   lipopolysaccharides ([148]26, [149]28). Nevertheless, no statistically
   significant GSDMD gene overexpression relative to the control was
   detected in monocytic cells, which we hypothesized to be related to the
   protection against cell self-lysis caused by multiple pore formation.
   Indeed, the RT-qPCR confirmed the increased GSDMD gene expression level
   in monocytic cells but only for ≥0.1-µg/mL concentrations of P.
   aeruginosa O10 LPS used in microarray assay. It might indicate an
   effect of limited mRNA expression or its degradation to protect cells
   from self-lysis via pyroptosis.

   In the next step, we checked whether the immune response, including
   pyroptosis ([150]31, [151]45), can be achieved in the human monocytic
   cells after induction by different LPS derived from P. aeruginosa
   strains (O10, PAO1, non-CF0038, CF217, CF532, CF832) and E. coli LPS
   (0111: B4). Additionally, other immunoreactivity of tested LPSs were
   also examined to elucidate variations in proinflammatory properties. A
   comparison of LPS-dependent IL-8 release showed significant differences
   depending on the source of LPS used, but only at a higher concentration
   of endotoxin indicating that the strongest stimulator was LPS from
   strain CF217, and the weakest was LPS from the non-mucoid CF532 strain.
   In the case of TNF, only the rough LPS CF217 had the strongest
   pro-inflammatory properties, whereas the proinflammatory potential of
   other LPS was on a similar level. Interestingly, P. aeruginosa
   undergoes lipid A structural modification during CF chronic infection,
   and this altered lipid A structure induces stronger signaling through
   TLR4, which is reflected in the production of proinflammatory cytokines
   ([152]58). Excluding the rough LPS from CF217, the other smooth LPSs
   showed only subtle differences in complement sequestration that were
   not correlated with the sensitivity of particular strains in
   complement-mediated bactericidal tests. The reason for that could be
   not only the different lengths and density of O-antigen but also
   O-antigen composition and chemistry within the studied LPS ([153]59,
   [154]60). It is logical, therefore, that a similar activation of the
   complement does not necessarily result in its lytic effect if, for
   example, the overall LPS structure block MAC (C5b-9) insertion or
   surface mucoid properties shields the outer membrane from lytic
   complement activity (non-mucoid CF532 versus mucoid CF832).

   The above-described well-defined LPS samples were then used to
   stimulate pyroptosis in THP1 cells to enhance the antibacterial
   activity of phage endolysin KP27. It turned out that although the
   endolysin alone exhibited moderate activity against PAO1 strain, no
   synergistic effect was noted for the combination with supernatants of
   LPS-induced monocytic cells, which indicates at least an insufficient
   level of pyroptosis event and bacterial OM permeabilization by GSDMD.

   The Western blot experiments confirmed that GSDMD protein is
   synthesized on a constant level in THP1 cells, and its soluble domain
   GSDMD[Cterm] can be found even in non-stimulated cells. The increased
   production of gasdermin D domains was noted when monocytic cells were
   treated with high concentrations of LPS (10–100 µg/mL). It means that
   the basic level of GSDMD is stable when cleaved by caspase 1 in
   response to LPS presence and sufficient for GSDMD[Nterm]-dependent
   membrane pore formation needed for cytokine release during pyroptosis.
   In our antibacterial assay, the LPS stimulation was too low for
   effective bacterial OM permeabilization and enhancement of endolysin
   activity. This seems to correlate with the role of GSDMD in the immune
   response to bacterial presence, whereas its overexpression could lead
   to unfavorable self-cell lysis of monocytic cells.

   In conclusion, only recombinant GSDMD[Nterm] protein was able to
   efficiently permeabilize P. aeruginosa OM and increase endolysin
   activity against bacteria producing either long or lacking LPS
   O-chains. The obtained results suggest the limited possibility of using
   the natural process of pyroptosis occurring in monocytic cells to
   enhance the bactericidal effect of recombinant phage endolysins against
   Gram-negative bacteria infection.

ACKNOWLEDGMENTS