Chlorhexidine Promotes Psl Expression in Pseudomonas aeruginosa That Enhances Cell Aggregation with Preserved Pathogenicity Demonstrates an Adaptation against Antiseptic

Because Pseudomonas aeruginosa is frequently in contact with Chlorhexidine (a regular antiseptic), bacterial adaptations are possible. In comparison with the parent strain, the Chlorhexidine-adapted strain formed smaller colonies with metabolic downregulation (proteomic analysis) with the cross-resistance against colistin (an antibiotic for several antibiotic-resistant bacteria), partly through the modification of L-Ara4N in the lipopolysaccharide at the outer membrane. Chlorhexidine-adapted strain formed dense liquid–solid interface biofilms with enhanced cell aggregation partly due to the Chlorhexidine-induced overexpression of psl (exopolysaccharide-encoded gene) through the LadS/GacSA pathway (c-di-GMP-independence) in 12 h biofilms and maintained the aggregation with SiaD-mediated c-di-GMP dependence in 24 h biofilms as evaluated by polymerase chain reaction (PCR). The addition of Ca2+ in the Chlorhexidine-adapted strain facilitated several Psl-associated genes, indicating an impact of Ca2+ in Psl production. The activation by Chlorhexidine-treated sessile bacteria demonstrated a lower expression of IL-6 and IL-8 on fibroblasts and macrophages than the activation by the parent strain, indicating the less inflammatory reactions from Chlorhexidine-exposed bacteria. However, the 14-day severity of the wounds in mouse caused by Chlorhexidine-treated bacteria versus the parent strain was similar, as indicated by wound diameters and bacterial burdens. In conclusion, Chlorhexidine induced psl over-expression and colistin cross-resistance that might be clinically important.


Introduction
Hospital-acquired bacterial infections caused by nosocomial pathogens are a concerning crisis due to high mortality, increased healthcare costs, and emerging antibiotic resistance [1]. As such, Pseudomonas aeruginosa has recently emerged as an important pathogen, causing chronic infections, partly due to the biofilm formation that is presented in several diseases; for example, cystic fibrosis, catheter-induced infection, and chronic  According to the proteomic study, 957 proteins were differentially expressed in C_PACL compared to PACL biofilms. Among these proteins, 496 proteins were upregulated, including 111 proteins with p < 0.05, 53 proteins with p < 0.01, and 22 proteins with p < 0.001 (Supplement Table S1), and 461 proteins were downregulated, including 116 proteins with p < 0.05, 69 proteins with p < 0.01, and 29 proteins with p < 0.001 (Supplement Table S2). The distribution of all upregulated and downregulated proteins is shown in the volcano plot (Figure 2A). The classifications of downregulated and upregulated proteins (p < 0.01) by the functional biological process are shown in Figure 2B,C. Most of the downregulated proteins were related to cellular metabolisms such as glycolysis, pyruvate metabolism, fatty acid synthesis, amide biosynthesis, and purine ribonucleotide metabolic processes, indicating the reduced metabolic rate in C_PACL ( Figure 2B). The stress-induced energy starvation of C_PACL, possibly related to the smaller colonies of SCVs, was indicated by the down-regulation of several proteins for glycolysis, pyruvate metabolic process, and the TCA (tricarboxylic acid) cycle, including enolase (Eno), glyceraldehyde-3-phosphate dehydrogenase. (Gap), pyruvate dehydrogenase E1 component (AceE), dihydrolipoyllysine-residue acetyltransferase component of pyruvate dehydrogenase complex (AceF), fumarate hydratase (FumC2), phosphoenolpyruvate carboxylase (Ppc), dihydrolipoyllysine-residue succinyltransferase component of 2-oxoglutarate dehydrogenase complex (SucB), and phosphoenolpyruvate synthase (PpsA), together with enzymes in the fatty acid synthetic process, including fatty acid biosynthetic G, H, and Z (FabG, FabH, and FabZ) ( Figure 2B and Supplement Figure S1).
Meanwhile, the upregulated proteins are associated with (i) antibiotic and Chlorhexidine resistance, including L-Ara4N (4-amino-4-L -deoxyarabinose), lipid A, and LPS biosynthetic processes and (ii) metabolic differences; such as cytochrome complex assembly and cellular polysaccharide biosynthesis ( Figure 2C,D). Because L-Ara4N obscured the lipid A from colistin binding [20], the up-regulation of L-Ara4N modification (Arn) (ArnA, ArnB, ArnC, ArnD, and ArnT) and PmrA and PmrB (Arn expression regulators) [21] might be associated with the resistance. Additionally, the up-regulated TolA and TolB (Tol-Pal system proteins) in C_PACL might strengthen the outer membrane integrity through the increased lipid A biosynthesis (during stress) [22]. The increased maturation proteins (CcmE, CcmF, and CcmH) of cytochrome C (a terminal electron receiver in the aerobic respiratory chain [23]) in C_PACL suggested an adaptation toward the reduced cell energy status, while enhanced ubiquinone biosynthesis (UbiA and UbiG), an essential electron carrier in denitrification in P. aeruginosa [24], implied the responses against reactive radical species in C_PACL. Moreover, Chlorhexidine might also reduce bacterial virulence in C_PACL because there was a downregulation of the controllers of many virulence factors, including (i) Pseudomonas quinolone signal (PQS) synthesis (PqsA, PqsB, and PqsD), (ii) host invasion promotor; LasA (protease) and LasB (elastase), and (iii) phenazine biosynthetic proteins (PhzF, PhzM, and PhzS) (Supplement Figure S1). The protein connection network study demonstrated that the upregulated proteins were related to RecA, a pathway that correlated with the significant stresses or "SOS" response (a global response to DNA damage-inducing the cell cycle arrest, DNA repair, and mutagenesis) [25] in C_PACL when compared with PACL group ( Figure 2D).   Proteomic study of the biofilms from P. aeruginosa parent strain (PACL) and Chlorhexidine (CHG)-treated strain (C_PACL). The volcano plot displayed the distribution of 957 proteins in which the protein abundance (log 2 fold change of C_PACL/PACL) plotted with the significant values (-log 10 p-value) (A). Proteins in the biological process that downregulated (B) and upregulated (C) in C_PACL compared to PACL (p < 0.01) using STRING v.10 are demonstrated. The functional protein association networks (D) for the complicated connection of upregulated proteins (p < 0.01) by STRING v.10 are also showed.

Alteration of P. Aeruginosa Extracellular Polysaccharides Leading to Cell Aggregation Possibly through Psl-Predominance after the Exposure to Chlorhexidine
Because SCVs play a critical role in the biofilms of P. aeruginosa during chronic infection [26], we characterized the biofilm structure of C_PACL ( Figure 3A-F). Indeed, C_PACL formed denser crystal violet-stained biofilms at the solid-liquid interface at the lateral surface of the well in polystyrene plates, indicating the increased strength of the interface biofilm when compared with PACL strain ( Figure 3A,C). Moreover, the liquid-air interface biofilms, called pellicles, in PACL produced rough cracked pellicles, while C_PACL produced large fine pellicles (intact sheets), indicating the higher cell aggregate in the C_PACL pellicles ( Figure 3B). With the confocal microscopy, PACL formed enormous extracellular matrix (ECM) biofilms (stained with red fluorescence, Figure 3D,F), indicating robust biofilm mass but a similar level of bacterial cells (stained with green fluorescence, Figure 3E,F). Conversely, C_PACL showed significantly less fluorescence-stained biofilm mass ( Figure 3D,F), but there were densely packed bacterial cells in the biomass ( Figure 3E,F), indicating an aggregate biofilm of C_PACL. Hence, Chlorhexidine exposure in C_PACL established a vigorous solid-liquid interface, prominent cell aggregation, and well-formed pellicle biofilms but not robust ECM production.
Among three extracellular polysaccharides (EPSs) produced by P. aeruginosa ( Figure 4A), alginate is a major component of slime biofilms, while Psl and Pel maintain the strength of solid-liquid and the air-liquid interface, respectively [7]. Both PACL and C_PACL isolates formed a similar wrinkling colony biofilm but at the edge of wrinkling of C_PACL colonies exhibited more Psl and Pel than PACL as indicated by the stained color of Congo red (a color stained for Pel or Psl but not alginate) [27,28], particularly in 14-day biofilms ( Figure 4B). These data suggested a possible higher binding capacity of C_PACL than PACL. In addition, a high Congo red binding capacity of C_PACL was confirmed in Luria-Bertani (LB) broth ( Figure 4C,D). There was an upregulation of pslB and pelA (the genes encoding for Psl and Pel, respectively) in the 24-h biofilm of C_PACL than PACL without the difference in algD (an alginate-encoding gene) ( Figure 4E-G), possibly associated with Psl and Pel-mediated denser cells [9]. Because the different expression of pslB between C_PACL and PACL was more prominent than the alteration in pelA, the enhanced aggregation during C_PACL biofilm-forming might be mostly due to psl than pel expression.   were performed in independent triplicate. Mean ± SEM are presented with the one-way ANOVA followed by Tukey's analysis (*, p ˂ 0.05; Φ , p ˂ 0.05; δ , p ˂ 0.05; ε , p < 0.05; and κ , p < 0.05).

Chlorhexidine Induced Psl-Predominated Biofilms in P. aeruginosa via SiaD and LadS/GacSA Regulatory System
Several "Surface attachments defective diguanylate cyclases" (such as SadC and SiaD; the control of SiaABCD complex for psl operon regulation [29,30]) through bis-(3 -5 )-cyclic dimeric guanosine monophosphate (c-di-GMP) [29,31], while LadS/GacSA (the hybrid histidine kinase recognizing several signals, including Ca 2+ ) induces small RNA (rsmZ and rsmY) [32,33] to sequester RsmA (negative psl regulator through psl mRNA degradation [34]) in c-di-GMP-independent Psl production ( Figure 5A). Hence, several genes of these pathways were explored ( Figure 5B-J). At 12 h biofilms, the c-di-GMP-independent Psl biosynthesis [35] in C_PACL was more prominent than PACL as indicated by upregulation of gacS, gacA, ladS, rsmZ, rsmY, and rsmA ( Figure 5B-G). Likewise, there was less motility and upregulation of pilA and oprF (the encoding genes for bacterial appendages used for cell attachment) ( Figure 6A-D) with lower expression of sadC and siaD (the c-di-GMP-dependent Psl production) in C_PACL than PACL ( Figure 5H-I). But the activity of c-di-GMP-independent Psl biosynthesis by these parameters of C_PACL decreased in 24 and 48 h biofilms ( Figure 5B-G). Then, the c-di-GMP-dependent Psl biosynthesis of C_PACL was more prominent than PACL at 24 h biofilms as indicated by upregulation of siaA and siaD (but not sadC) ( Figure 5H-J), implying the SiaD-mediated c-di-GMP-dependent of Psl-production at 24 h ( Figure 5A). Notably, the upregulation of rsmA (a negative regulator of Psl production) was demonstrated only in 12 h of C_PACL biofilms, but neither in 24 nor 48 h of biofilm production ( Figure 5G). These results suggested that Chlorhexidine enhanced P. aeruginosa Psl-predominated aggregate initially through c-di-GMP-independent LadS/GacSA at 12 h and the Psl production was sustained via SiaD-mediated c-di-GMP dependent pathway at 24 h and 48 h.
Because (i) the activation of c-di-GMP-independent LadS/GacSA is Ca 2+ dependence [33], (ii) Ca 2+ is one of the divalent ions stabilizers for the negative charge of LPS in Gram-negative bacteria [36], and (iii) Chlorhexidine replaces the divalent ions [13] that might release Ca 2+ from the bacterial membrane and activate the inner membrane sensor (LadS and GacSA system) ( Figure 5A), the importance of Ca 2+ in Psl production is tested ( Figure 7A). Indeed, Chlorhexidine with CaCl 2 in combination showed a greater expression of gacS, ladS, and pslB than either Chlorhexidine or CaCl 2 alone at 12 h biofilm production ( Figure 7B-D), indicating a Chlorhexidine-Ca 2+ synergy on the GacSA activation that subsequently enhanced psl upregulation. Perhaps, the free Ca 2+ from the bacterial cell membrane, after Chlorhexidine cationic replacement, might turn on the LadS/GacSA-mediated psl operon, at least in part, through SiaD-induced c-di-GMP pathway ( Figure 5A).    . Mean values of the relative mRNA expression were plotted with error bars representing the SEM. The p-values were calculated using one-way ANOVA followed by Tukey's analysis (*, p < 0.05 and # , p < 0.05 considered statistically significant).

The Responses of Fibroblasts and Macrophages against Pseudomonas Parent Strain (PACL) and Chlorhexidine-Activated Strain (C_PACL)
Because the psl-mediated aggregation in P. aeruginosa plays a critical role in chronic wound infection [37,38] which consists of several cell types, including fibroblasts and macrophages, the responses of these cells against both strains of P. aeruginosa are interesting. The difference between the fibroblast responses against sessile bacteria (bacteria with biofilms) versus planktonic form (bacteria without biofilms) within PACL parent strain or C_PACL was not obviously different. Accordingly, there was the upregulated TLR-4 but downregulated IL-6 without the difference in other genes nor cytokine responses in sessile PACL compared with planktonic PACL (Figure 8A-M). Meanwhile, there was a higher TLR-4 and TNF-α expression but lower IL-6, IL-8, TGF-β, and GM-CSF in sessile C_PACL when compared with the activation by planktonic C_PACL (Figure 8A-M). Thus, the difference of the activation by sessile versus planktonic PACL or sessile versus planktonic C_PACL toward fibroblasts were not obvious. In comparison with planktonic PACL, the planktonic C_PACL upregulated IL-8, TGF-β, GM-CSF, TLR-2, TLR-5, and TLR-6 but downregulated IL-6 without the alteration in other parameters in fibroblasts ( Figure 8A-M) implied a mildly prominent pro-inflammation of planktonic C_PACL over planktonic PACL. In contrast, the sessile C_PACL down-regulated IL-6, IL-8, TGF-β, and GM-CSF when compared with sessile PACL (Figure 8A-M) suggested a less inflammatory impact of C_PACL biofilms on fibroblasts possibly due to the reduced biofilm biomass in sessile C_PACL compared with sessile PACL (Figure 3A-F). These data indicated that Chlorhexidine decreased the pro-inflammatory properties of the sessile P. aeruginosa (but not the planktonic form) toward fibroblasts. In THP-1-derived macrophages, the sessile form PACL induced more inflammatory responses than the planktonic PACL as indicated by the higher expression of TLR-2, TLR-4, and iNOS (the pro-inflammatory molecules) with lower expression of Arg-1 and IL-10 (the anti-inflammatory molecules) without the alteration in other parameters ( Figure 9A-O). The more prominent proinflammation of sessile PACL over planktonic PACL on macrophages supports a pro-inflammatory impact of biofilm structures on macrophages [39]. Meanwhile, the sessile C_PACL induced less inflammation than the planktonic C_PACL as indicated by the lower expression of TNF-α, IL-8, and iNOS (but lower IL-10) ( Figure 9A-O) suggesting that the biofilm of C_PACL induced some antiinflammatory effects on macrophages, possibly Psl or Pel-induced anti-inflammation [39,40]. Additionally, the planktonic C_PACL induced lower macrophage inflammatory responses than planktonic PACL as indicated by lower expression of TLR-2, TLR-4, TLR-5, TLR-6, IL-6, and iNOS with higher IL-10 expression ( Figure 9A-K) supporting the less prominence pro-inflammation of C_PACL bacteria without the biofilm molecules. Likewise, the sessile C_PACL induced lower macrophage inflammatory responses than sessile PACL as indicated by lower expression of TLR-6, TNF-a, IL-6, IL-8, and iNOS; however, with the lower Arg-1 and IL-10 expression ( Figure 9A-K). These data indicated that Chlorhexidine decreased the pro-inflammatory properties of both sessile and planktonic P. aeruginosa toward macrophages.

The Responses of Fibroblasts and Macrophages against Pseudomonas Parent Strain (PACL) and Chlorhexidine-Activated Strain (C_PACL)
Because the psl-mediated aggregation in P. aeruginosa plays a critical role in chronic wound infection [37,38] which consists of several cell types, including fibroblasts and macrophages, the responses of these cells against both strains of P. aeruginosa are interesting. The difference between the fibroblast responses against sessile bacteria (bacteria with biofilms) versus planktonic form (bacteria without biofilms) within PACL parent strain or C_PACL was not obviously different. Accordingly, there was the upregulated TLR-4 but downregulated IL-6 without the difference in other genes nor cytokine responses in sessile PACL compared with planktonic PACL (Figure 8A-M). Meanwhile, there was a higher TLR-4 and TNF-α expression but lower IL-6, IL-8, TGF-β, and GM-CSF in sessile C_PACL when compared with the activation by planktonic C_PACL (Figure 8A-M). Thus, the difference of the activation by sessile versus planktonic PACL or sessile versus planktonic C_PACL toward fibroblasts were not obvious. In comparison with planktonic PACL, the planktonic C_PACL upregulated IL-8, TGF-β, GM-CSF, TLR-2, TLR-5, and TLR-6 but

Psl-Predominated Chlorhexidine-Activated Strain (C_PACL) Similarly Induced Wound Infection in Mice When Compared with the Parent Strain (PACL)
From the in vitro experiments, Chlorhexidine reduced Pseudomonas biofilm abundance ( Figure 3) with increased Psl-predominance (Figure 4), partly through Ca 2+ -dependent operon activation (Figures 5 and 7) through several signals (Figures 5 and 7), resulting in a decrease in the responses of fibroblasts and macrophages (Figures 8 and 9). Then, we tested both strains on a wound model to explore bacterial pathogenicity ( Figure 10A). As such, the severity of infection as determined by wound size and bacterial burdens using C_PACL was lower than PACL at 7 days of experiments, but not at 3 and 14 days, without the difference in serum cytokine levels at 14 days of experiments ( Figure 10B-H). The infection by a single type of organism was supported by the similar colony morphology (flat, opaque, and green-pigment colonies with irregular margins) with Gram-negative bacilli from the culture of wound fluid (Supplement Figure S2A) and the bacterial colonies from the wound fluid of C_PACL infection were smaller than the PACL colonies, supporting Chlorhexidineinduced small variant colonies. Despite no bacteremia in all groups (data not shown), serum IL-1β, IL-6, and TNF-α (but not IL-10) were increased in the infected mice from both bacterial strains ( Figure 10E-H). On the 14th-day experiment, the wound severity score (abundance of infiltrated immune cells beneath the wound beds, depth, and wellness) of C_PACL infection was similar to the PACL group using H&E staining (Supplement Figure S3B-E). Our data indicated that Chlorhexidine decreased Pseudomonas virulence as indicated by biofilm biomass, host cell responses (fibroblasts and macrophages), and the virulence in mice (at 7 days of the wounds), despite an elevated Psl production. Although Chlorhexidine did not enhance bacterial virulence, the induction of antibiotic resistance should be concerned. With the exposure of Chlorhexidine on the other 12 clinical-isolated strains of P. aeruginosa, the increased MICs against Chlorhexidine or colistin was detected in 8 and 10 strains, respectively, MICs against both agents increased in 6 strains, and small colonies (SCVs) was demonstrated in 6 strains ( Figure 11A-C and Supplement Table S3). Hence, our study strongly indicated that the use of Chlorhexidine could induce P. aeruginosa with pathogenicity and antibiotic cross-resistance (especially against colistin) which might be clinically important.

Chlorhexidine Use and Colistin Cross-Resistance
Although Chlorhexidine, a frequently used antiseptic in hospital, effectively control bacteria by the chemical properties [41] and antiseptic reservoir (sponge effect of the killed bacteria that destroys other bacterial generations referred to as the zombie effect of Chlorhexidine remnants) [42], the sub-lethal Chlorhexidine concentrations after each use will naturally be lower than the bactericidal levels and possibly induce bacterial adaptations [41]. Indeed, Chlorhexidine remnant replaces Ca 2+ and penetrates the phospholipid layer of the P. aeruginosa membrane [13] with several evasion mechanisms to increase bacterial membrane integrity against the antiseptic through various proteins (OprF, TolA, and TolB) [15] supported the TolA and TolB upregulation in our Chlorhexidine-stimulated bacteria. Due to similar properties and actions of Chlorhexidine and colistin, colistin crossresistance is frequently demonstrated following Chlorhexidine exposure [15,18], possibly through L-Ara4N modification in lipid A (a binding target of colistin [17]) of the bacterial membrane [43]. In colistin-resistant P. aeruginosa, the colistin-modified lipid A can be reformed by adding the positively charged L-Ara4N [43] or enhanced lipid A biosynthesis or increased efflux pump functions [15,18,44]. Here, increased L-Ara4N was demonstrated in proteomic analysis of Chlorhexidine-treated P. aeruginosa.

Chlorhexidine Use Induced Small Colony Variants (SCVs) with Psl-Predominated Biofilms through c-di-GMP-Independence and -Dependence at 12 and 24 h of Biofilms, Respectively
The P. aeruginosa SCVs (a small colony with 1-3 mm in diameter firstly isolated from patients with cystic fibrosis at 24-48 h of 37 °C incubation [45]) demonstrated metabolic arrest as indicated by the reduction of several proteins in proteomic analysis. Other characteristics of SCVs, including wrinkle, auto-aggregation, hyper-adherence, and biofilm formation as previously described [46], were also demonstrated in Chlorhexidine-treated strain (partly associated with Psl, a neutral pentasaccharide for biofilm initiation, attachment, and microcolony formation [4]). The cell aggregation through Psl-mediated polymer bridging manifests a less order alignment that intensely packed cells at the center of biofilms [9] similar to our Chlorhexidine-treated strain ( Figure 3F). Indeed, the psl operon is composed of 15 co-transcribed genes (pslA to pslO) encoding essential proteins for Psl

Chlorhexidine Use and Colistin Cross-Resistance
Although Chlorhexidine, a frequently used antiseptic in hospital, effectively control bacteria by the chemical properties [41] and antiseptic reservoir (sponge effect of the killed bacteria that destroys other bacterial generations referred to as the zombie effect of Chlorhexidine remnants) [42], the sub-lethal Chlorhexidine concentrations after each use will naturally be lower than the bactericidal levels and possibly induce bacterial adaptations [41]. Indeed, Chlorhexidine remnant replaces Ca 2+ and penetrates the phospholipid layer of the P. aeruginosa membrane [13] with several evasion mechanisms to increase bacterial membrane integrity against the antiseptic through various proteins (OprF, TolA, and TolB) [15] supported the TolA and TolB upregulation in our Chlorhexidine-stimulated bacteria. Due to similar properties and actions of Chlorhexidine and colistin, colistin crossresistance is frequently demonstrated following Chlorhexidine exposure [15,18], possibly through L-Ara4N modification in lipid A (a binding target of colistin [17]) of the bacterial membrane [43]. In colistin-resistant P. aeruginosa, the colistin-modified lipid A can be reformed by adding the positively charged L-Ara4N [43] or enhanced lipid A biosynthesis or increased efflux pump functions [15,18,44]. Here, increased L-Ara4N was demonstrated in proteomic analysis of Chlorhexidine-treated P. aeruginosa.

Chlorhexidine Use Induced Small Colony Variants (SCVs) with Psl-Predominated Biofilms through c-di-GMP-Independence and -Dependence at 12 and 24 h of Biofilms, Respectively
The P. aeruginosa SCVs (a small colony with 1-3 mm in diameter firstly isolated from patients with cystic fibrosis at 24-48 h of 37 • C incubation [45]) demonstrated metabolic arrest as indicated by the reduction of several proteins in proteomic analysis. Other characteristics of SCVs, including wrinkle, auto-aggregation, hyper-adherence, and biofilm formation as previously described [46], were also demonstrated in Chlorhexidine-treated strain (partly associated with Psl, a neutral pentasaccharide for biofilm initiation, attachment, and microcolony formation [4]). The cell aggregation through Psl-mediated polymer bridging manifests a less order alignment that intensely packed cells at the center of biofilms [9] similar to our Chlorhexidine-treated strain ( Figure 3F). Indeed, the psl operon is composed of 15 co-transcribed genes (pslA to pslO) encoding essential proteins for Psl biosynthesis [7] that are regulated by c-di-GMP [29] through many P. aeruginosa diguanylate cyclases (SadC and SiaD) [30,47] or non-c-di-GMP pathways [34]. In c-di-GMP-independent Psl production, RsmA (a repressor protein that switches off psl [34]) is sequestered by rsmY and rsmZ [48], and controlled by GacSA (a positive regulator activating by the phosphorylation using GacS and LadS kinases [32] that is partly stimulated by Ca 2+ [33,35], thiamine, and other metabolites [49]). Although the role of c-di-GMP on the psl promoter is still unclear due to the non-predicted Psl-binding site found in SadC and SiaD [30], the addition of exogenous Psl in a previous publication promotes psl expression implying Psl positive feedback on SadC and SiaD [30]. In our study, C_PACL biofilms overexpressed pslB (Figure 4F) related to siaD ( Figure 5I) with an increased binding affinity of Congo red (the color stained for Psl or Pel) suggesting that Psl expression was partially maintained by c-di-GMP, especially in 24-h biofilms. On the other hand, LadS/GacSA (the c-di-GMP-independent Psl synthesis) was important at 12 h biofilms of C_PACL which could be facilitated by Ca 2+ (a universal signal used for cellular responses [50]) possibly as the switching from an acute to a chronic adaptation [33]. Due to the Ca 2+ displacement and the loss of membrane integrity by Chlorhexidine, Ca 2+ might be released from the membrane and activates LadS/GacSA ( Figure 5A) which is responsible for the initial adaptation against Chlorhexidine. More studies on this topic are interesting.

Chlorhexidine-Treated Pseudomonas Caused Chronic Wound in Mice, despite a Mildly
Reduced Disease Severity, with a Possibility of Antibiotic Cross-Resistance P. aeruginosa biofilm with alginate and Psl predominance has been demonstrated in cystic fibrosis (a chronic infection) [2,38] and Psl causes treatment failure partly through antibiotic tolerance from intense cell aggregation [9,38]. In contrast, there are only a few studies of P. aeruginosa biofilms in wound infections. In the infected wounds [51], fibroblasts mediate tissue homeostasis [52] and recognize pathogen-associated molecular patterns (PAMPs) by several TLRs with other mediators such as IL-6, IL-8, and MMP-1 (matrix metalloproteinase-1) [53]. Meanwhile, macrophages professionally recognize PAMPs and produce several cytokines (such as IL-6) that also promote wound healing through the proliferation of several cell types (fibroblasts, keratinocytes, and myofibroblasts) [54]. Although the expression of some inflammatory genes of fibroblasts and macrophages induced by the Chlorhexidine-activated strain was lower than the parent strain, the cytokine production after activation by both strains was not different. Because some molecules of P. aeruginosa biofilms (such as alginate) directly induce skin inflammation that might be a chemical interference of the wound model [55], the planktonic P. aeruginosa are applied in mouse wounds. Here, bacterial burdens in the wounds of C_PACL at 7 days were less than the PACL parent strain indicating a possible milder disease severity from Chlorhexidineexposed bacteria, despite the non-difference at 3 and 14 days of the model. The Psldominant biofilms resist several immune responses (phagocytosis, neutrophil oxidative burst, and complement deposition) [56], and the Psl-overproduced P. aeruginosa PAO1 induces a higher bacterial load than the regular strain in 35 days porcine wounds [26]. In contrast, the 14 days mouse wounds from Psl-producing and Psl-depleted Pseudomonas PAO1 are not different [40]. Then, impacts of Psl on the wounds are still inconclusive. Despite the non-difference of Chlorhexidine-exposed Pseudomonas to the parent strain in our wound model, the Psl-predominated Pseudomonas after Chlorhexidine use might induce more severe infections in other models, as determined from several Psl-mediated bacterial evasion mechanisms [56]. Moreover, the cross-resistance against colistin and imipenem after Chlorhexidine use might lead to further clinical problems because of the importance of both agents against the current antibiotic-resistant bacteria [57]. Because chronic exposure to low-dose Chlorhexidine is necessary for the development of Psl-mediated bacterial adaptation, the adequate doses with frequent use of Chlorhexidine might be beneficial for the clinical translation. More studies on this topic are interesting.

Animals and Animal Model
The Institutional Animal Care and Use Committee of the Faculty of Medicine, Chulalongkorn University has approved the Animal care and use protocol (SST012/2564) following the US National Institutes of Health guidelines using specific pathogen free mouse facility with a 12:12 h light/dark cycle at 22 ± 2 • C and 50% relative humidity. As such, 8-week-old male C57BL/6 mice (Nomura Siam International, Pathumwan, Bangkok, Thailand) were used with free access to water and food at the Faculty of Medicine, Chulalongkorn University for a mouse model of wound infection as previously described [58]. Under the isoflurane anesthesia (Baxter, Puerto Rico, USA), the skin at the back of mice (left and right sides) was shaved, cleaned with 10% povidone-iodine, and induced an open wound with an aseptic procedure using a skin puncture (a punch biopsy with diameter 8 mm) (Oem factory, Guangdong, China) before the inoculation of 1 × 10 6 colony-forming unit (CFU) of P. aeruginosa (parent and Chlorhexidine-exposed strains; details below) in normal saline solution (NSS) with the total volume of 50 µL per wound or NSS alone. Then, the open wounds were covered with 3M Tegaderm films (1622W) (3M Science, USA) and the films were daily changed until sacrifice. After the wound induction, a single subcutaneous injection (25 mg/kg) of tramadol (Sigma-Aldrich, St. Louis, MO, USA) was performed for postoperative analgesia. Fluid material on the wound was collected on the 3rd, 7th, and 14th days of experiments for P. aeruginosa culture (representing bacterial burden in the wounds). Because the wound fluid was not enough for the direct pipetting, NSS (50 µL) was put on the wound before stirring with the pipette tip, and collection of 20 µL wound fluid for the culture in LB (Luria-Bertani) agar plates (the non-selective media) before colony enumeration after 24 h incubation at 37 • C. The infection by the single organism was supported by the staining of Gram-negative bacilli in all colonies. On day 14, all mice were sacrificed with cardiac puncture under isoflurane anesthesia before aseptic sample collection. The serum cytokine levels were determined by ELISA kit (Invitrogen, Carlsbad, CA, USA). After sacrifice, wound tissues were carefully collected and preserved in 10% formalin. The tissues were processed, cut into sections, and placed on a glass slide, followed by hematoxylin and eosin staining. The histopathological analysis was evaluated based on inflammatory cells infiltration, wound diameter, wellness, and redness (inflammatory degree evaluation) with a score (0-3) according to a previous publication [59].  [60]. Briefly, serial dilutions of these agents were prepared in 96-well culture plates, followed by inoculation of P. aeruginosa isolates (final concentration 1 × 10 5 CFU/well) and incubation at 37 • C for 18 h. The MICs, the lowest concentration that can inhibit bacterial growth, and the susceptibility test were visually evaluated following the CLSI guidelines (Supplementary Table S4).

Chlorhexidine Induction Protocol
For Chlorhexidine induction, all 13 P. aeruginosa isolates were grown in LB broth (Difco, Rockville, MD, USA) containing Chlorhexidine at the concentrations 0.25-fold (0.25×) of the MIC at 37 • C with shaking at 200 rpm for 48 h. After incubation, 1:100 of bacterial culture was grown in LB broth supplemented with a serially increasing concentration of Chlorhexidine (0.25×, 0.5×, 1×, and 2×) at 37 • C with shaking at 200 rpm for 48 h. After induction, P. aeruginosa was sub-cultured daily on the Chlorhexidine-free LB agar plate for ten days.

Biofilm Formation Study
Biofilms forming of P. aeruginosa parent strain (PACL) and Chlorhexidine (CHG)treated strain (C_PACL) in polystyrene 96-well culture plates, on glass coverslips, or in glass flasks were evaluated. The turbidity of P. aeruginosa that were cultured overnight in LB broth was adjusted to 0.5 McFarland standard. In 96-well plates, bacteria were cultured for 12, 24, and 48 h of static incubation at 37 • C. before biofilm identification with crystal violet color following a previous publication [61]. Briefly, the planktonic bacteria were removed, and biofilms were gently washed with distilled water twice before staining with 0.1% crystal violet for 15 min and washed twice with distilled water. Similarly, biofilms on glass coverslips were prepared in 6-well culture plates containing bacterial suspension with a glass coverslip at the bottom before removing the suspension, washed with phosphate buffer saline (PBS), fixed with 10% formaldehyde for 10 min, and stained for bacterial DNA with SYTO 9 green fluorescence (Invitrogen, Carlsbad, CA, USA) for 30 min. Then, the biofilm matrix was stained with Alexa Fluor 647 conjugated of Concanavalin A (AF647) (Invitrogen, Carlsbad, CA, USA) for 30 min. The biofilm images were photographed by the confocal laser scanning microscope, LSM 800 with Airyscan (Carl Zeiss Microscopy GmbH, Jena, Germany). Bacterial DNA (green) and biofilm matrix (red) fluorescent stains were visualized with a 63×/1.4 oil immersion lens. A total of 6 Z-stack images of the biofilms were acquired and analyzed using Zen microscopy software (Carl Zeiss Microscopy GmbH, Germany). The fluorescent intensities were also calculated using Zeiss microscopy software.

Congo Red Binding Assay
The non-alginate components of polysaccharide (Psl or Pel) in P. aeruginosa biofilms were stained by Congo red in LB agar. Briefly, P. aeruginosa colony biofilms of P. aeruginosa parent strain (PACL) and Chlorhexidine (CHG)-treated strain (C_PACL) were formed on LB agar plus 1% Congo red for 14 days at 25 • C and visually observed (red color on the biofilms is Pslor Pel-stained polysaccharide). In parallel, the Congo red binding (an ability of bacteria to bind with the color) was quantified using P. aeruginosa growing in LB broth supplemented with 40, 80, and 160 µg/mL of Congo red for 24 h at 37 • C with shaking. The bacterial cells were pelleted and the remaining Congo red color in the culture supernatant was measured using OD 495nm (Congo red detection). The Congo red binding was calculated by the initial OD 495nm of the culture supernatant (the color intensity before capturing by bacteria) subtracting with OD 495nm of the culture supernatant after 24 h of bacterial growth (the color intensity after binding by bacteria).

Bacterial Gene Expressions
The expression of genes encoding for biofilm-related proteins of P. aeruginosa parent strain (PACL) and Chlorhexidine (CHG)-treated strain (C_PACL) was determined by quantitative real-time reverse transcription-polymerase chain reaction (qRT-PCR) using primers listed in Supplementary Table S5. Total RNA of planktonic (bacteria without biofilms) or sessile (bacteria with biofilms) forms were extracted using TRIzol ® Reagent (Invitrogen, Carlsbad, CA, USA), and cDNA was synthesized using RevertAid First Strand cDNA Synthesis Kit (Thermo Scientific, Vilnius, Lithuania). The number of biofilm-related transcripts was quantified by the QuantStudio6 (Applied Biosystems, Foster City, CA, USA) using the SYBR ® Green PCR Master Mix (Applied Biosystems). The relative number of transcripts was normalized with 16S rRNA and calculated using the 2 −∆∆ct method.

Bacterial Motility Testing
The motile activity of P. aeruginosa parent strain (PACL) and Chlorhexidine (CHG)treated strain (C_PACL) was performed on 0.5% LB agar. Briefly, overnight culture of P. aeruginosa in LB broth was spotted at the center of the 0.5% LB agar plate. After incubation at 37 • C for 18 h, the bacterial growth wider than the initial spot indicated the ability to motile on the agar surface. In contrast, the bacterial growth at the initial spot demonstrated the motility defect.

Proteomic Analysis
The 24-h biofilms of the parent P. aeruginosa (PACL) and Chlorhexidine-activated P. aeruginosa (C_PACL) were prepared in glass flasks and removed by cell scrapers. The biofilm pellets were collected by centrifugation at 4500 rpm at 4 • C for 30 min, mixed with 8 M urea, sonication, and determining protein concentrations (BCA assay). Then, the protein mixtures were digested with trypsin and the peptides from C_PACL and PACL were labeled with light (CH 2 O and NaBH 3 CN) and medium (CD 2 O and NaBH 3 CN) dimethyl labeling reagents, respectively, before quenching the reaction with ammonia solution and formic acid. The labeled peptides were separated into 9 fractions using the Pierce High pH Reversed-Phase Peptide Fractionation Kit. Eluates of each fraction was dried in a SpeedVac centrifuge before LC-MS/MS analysis. The fractionated samples were resuspended in 0.1% formic acid and then analyzed via an EASY-nLC1000 system (Thermo Fisher Scientific, Waltham, MA, USA) coupled to a Q-Exactive Orbitrap Plus mass spectrometer (Thermo Fisher Scientific) equipped with a nano-electrospray ion source (Thermo Fisher Scientific). The MS methods included a full MS scan at a resolution of 70,000 followed by 10 datadependent MS2 scans at a resolution of 17,500. An MS scan range of 350 to 1400 m/z was selected and precursor ions with unassigned charge states, a charge state +1, or a charge state of greater than +8 were excluded. The peak list-generating software used in this study was Proteome Discoverer™ Software 2.1 (Thermo Fisher Scientific). The SEQUEST-HT search engine was employed in the data processing. The MS raw data files were searched against P. aeruginosa (strain ATCC 15692/DSM 22644/CIP 104116/JCM 14847/LMG 12228/1C/PRS 101/PAO1) Database (1435 proteins, March 2022) and a list of common protein contaminants (www.thegpm.org/crap/) (accessed on 27 May 2022). The Proteome Discoverer decoy database together with the Percolator algorithm was used to calculate the false-positive discovery rate of the identified peptides based on Q-values which were set to 1%. The Precursor Ions Quantifier node in Proteome Discoverer™ Software was employed to quantify the relative MS signal intensities of dimethyl labeled peptides. The control channels were used as denominators to generate abundance ratios of C_PACL/PACL. Log2 of the normalized ratio was used to calculate the mean and standard deviation of fold change across all three biological replicates. The proteomic data were submitted to the ProteomeXchange consortium (PXD034530).

Fibroblast and Macrophage Experiments
Responsiveness of human fibroblasts and macrophages against P. aeruginosa parent strain (PACL) and Chlorhexidine (CHG)-treated strain (C_PACL) planktonic and sessile forms was evaluated. Primary human dermal fibroblasts (isolated from healthy male neonatal foreskin tissues; Protocol number IRB 120/63 approved by the Institutional Review Board of the Faculty of Medicine, Chulalongkorn University) were grown in Dulbecco's Modified Eagle Medium (DMEM) (Gibco, Life Technologies, Ltd., Paisley, Scotland, UK) supplemented with 10% fetal bovine serum (FBS) (Gibco, Life Technologies, Ltd.) in a humidified atmosphere at 37 • C, 5% CO 2 . The primary human dermal fibroblasts (1 × 10 5 cells/well) in 12-well culture plates were incubated with either the Pseudomonas planktonic bacteria (1 × 10 4 CFU) or the sessile form, prepared by 1 × 10 4 CFU of bacteria on the polystyrene plates, at 37 • C in the CO 2 incubator for 6 h. The protocol of primary fibroblasts isolation from a publication was followed [62]. Briefly, the foreskin tissues from healthy volunteers (IRB 120/63) were collected in sterile containers supplemented with DMEM and 10% FBS with 5 µg/mL of Plasmocin (the mycoplasma elimination reagent) (InvivoGen, San Diego, CA, USA). Then, the tissues were minced and incubated on DMEM with 20% PBS plus gentamicin (Sigma-Aldrich) at 37 • C in the CO 2 incubator. The skin remnants were discarded after 7 days of incubation with the presence of dermal fibroblasts (observed by light microscopy analysis). At 21 days of incubation, the dermal fibroblasts were harvested with 0.25% trypsin in EDTA (Ethylenediaminetetraacetic acid) before centrifugation. The pellets of fibroblasts were resuspended in DMEM with 10% PBS plus gentamicin (Sigma-Aldrich) for quantification using a hemocytometer. In parallel, THP-1 (ATCC TIB202), a human monocytic cell line, was cultured in RPMI 1640 (Gibco Life Technologies, Ltd.) supplemented with 10% FBS before the differentiation into macrophages using 100 nM phorbol myristate acetate (Sigma-Aldrich). The THP-1-derived macrophages, seeded at 1 × 10 5 cells/well in 12-well plates, were treated with planktonic or sessile forms of P. aeruginosa as described in the fibroblast experiments. After treatment, the supernatant was collected and determined for cytokine concentration using ELISA assays. The cytokine and Toll-like receptor gene expressions were determined by qRT-PCR using primers listed in Supplementary Table S5 with the previously mentioned protocol (in the section on bacterial gene expression). Briefly, total RNA was extracted and converted to cDNA before determination of the expression of genes for cytokines and Toll-like receptors (normalized with β-actin) using the 2 −∆∆ct method.

Statistical Analysis
All data were analyzed using the Statistical Package for Social Sciences software (SPSS 22.0, SPSS Inc., Chicago, IL, USA) and Graph Pad Prism version 7.0 software (La Jolla, CA, USA). Results were presented as mean ± standard error (SE). The differences between two groups and multiple groups were examined for statistical significance by an unpaired t-test or one-way analysis of variance (ANOVA) with Tukey's analysis, respectively. The survival analysis and time-point data were determined by the Log-rank test and repeated measures ANOVA, respectively. A p-value < 0.05 was considered statistically significant.

Conclusions
With prolonged exposure to sub-lethal Chlorhexidine, P. aeruginosa developed a crossresistance to several antibiotics, especially colistin, possibly through the release of Ca 2+ from bacterial membrane due to Chlorhexidine replacement, which initially triggered LadS/GacSA (c-di-GMP-independent) followed by SiaD-induced-c-di-GMP-dependent Psl-mediated aggregation causing small-colony variants, metabolic arrests, and reduced inflammatory activation (on fibroblasts and macrophages). However, the Chlorhexidinemediated Psl-predominant P. aeruginosa demonstrated a similar wound infection to the parent strain with the cross-resistance to some antibiotics with a possible significance. Hence, we propose adequate doses and frequency of Chlorhexidine use.

Supplementary Materials:
The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/ijms23158308/s1, References [47,60,61,63]   Institutional Review Board Statement: The study was conducted in accordance with the Declaration of Helsinki, and approved by the Institutional Review Board, the Faculty of Medicine, Chulalongkorn University, Bangkok, Thailand (610/2564). The study was approved by MDCU-IBC in accordance with the levels of risk in pathogens and animal toxins list in the Risk Group 2 of pathogens by the Institutional Biosafety Committee, the Faculty of Medicine, Chulalongkorn University, Bangkok, Thailand (MDCU-IBC001/2022). The animal study protocol was approved by the Institutional Animal Care and Use Committee of the Faculty of Medicine, Chulalongkorn University, Bangkok, Thailand, following the National Institutes of Health (NIH), USA (SST 012/2564).

Informed Consent Statement: Not applicable.
Data Availability Statement: The data presented in this study are available on request from the corresponding author.