Immune Response Modulation by Pseudomonas aeruginosa Persister Cells

ABSTRACT Bacterial persister cells—a metabolically dormant subpopulation tolerant to antimicrobials—contribute to chronic infections and are thought to evade host immunity. In this work, we studied the ability of Pseudomonas aeruginosa persister cells to withstand host innate immunity. We found that persister cells resist MAC-mediated killing by the complement system despite being bound by complement protein C3b at levels similar to regular vegetative cells, in part due to reduced bound C5b, and are engulfed at a lower rate (10- to 100-fold), even following opsonization. Once engulfed, persister cells resist killing and, contrary to regular vegetative cells which induce a M1 favored (CD80+/CD86+/CD206−, high levels of CXCL-8, IL-6, and TNF-α) macrophage polarization, they initially induce a M2 favored macrophage polarization (CD80+/CD86+/CD206+, high levels of IL-10, and intermediate levels of CXCL-8, IL-6, and TNF-α), which is skewed toward M1 favored polarization (high levels of CXCL-8 and IL-6, lower levels of IL-10) by 24 h of infection, once persister cells awaken. Overall, our findings further establish the ability of persister cells to evade the innate host response and to contribute chronic infections.

appropriate growth conditions is present, persister cells can revert into an antibiotic susceptible, active state (11,12). The characteristics of persister cells render the use of most antimicrobials for the treatment of chronic infections ineffective (13).
Despite the possible contribution to chronic bacterial infections, the interactions between bacterial persister cells and the host innate immune system are still poorly understood. Typically, during an infection, macrophages are the first immune cells to recognize pathogens via their pattern recognition receptors (14). Macrophages can typically polarize toward a M1 type or M2 type response to meet the needs of the host. In an active infection, a M1 response is typically detected and is characterized by the presence of the macrophage cell membrane marker CD80, the production of reactive oxygen species (ROS) and proinflammatory cytokines such as IFN-b, IL-1, IL-6, IL-12, and TNF-A (15). M1-polarized macrophages typically promote differentiation of Th1 and Th17 T cells, which create an inflammatory feedback loop ultimately leading to the clearance of pathogens (16). However, it has previously been demonstrated that during infections with biofilms, a M2 response is mostly detected where the cells contain the cell membrane marker CD206 and secrete anti-inflammatory cytokines, such as IL-10 and CCL5 (15). M2-polarized macrophages include a subdivision known as regulatory macrophages, or M2b macrophages; these cells play a major role in modulating the inflammatory response and are distinguished by secretion of TNF-A and IL-6, high IL-10 secretion combined with low IL-12, and expression of cell marker CD86 (17).
The complement system functions as a nonspecific defense against invading bacteria, particularly a defense against Gram-negative bacteria. Several mechanisms of complement resistance by various species of Gram-negative bacteria have been previously reported, including capsular modulation which can conceal antibody epitopes (18), the decrease of complement-mediated phagocytosis (19), resistance to the insertion of the terminal complex of complement proteins C5b-C9 in the bacterial membrane resulting in an absence of the membrane attack complex (MAC) (20,21), and by sialyation of lipooligosaccharides on their surface (22,23). These findings have yet to be confirmed for persister cell populations. Although little is known regarding the innate immune response to persister cells, several studies found that persister cells can survive inside macrophages (24)(25)(26) and are engulfed at a lower rate following infections (27,28). It has also previously been found that the immune system can induce a persister state in several bacterial species such as when S. aureus is exposed to host oxidative stress (29), when Salmonella spp. are internalized by macrophages (24), upon exposure of Mycobacterium tuberculosis to cytokines (25), and when Vibrio splendidus is exposed to host-derived sea cucumber coelomic fluid (30).
Pseudomonas aeruginosa contributes to lung infection in cystic fibrosis (CF) patients (31) and to chronic obstructive pulmonary disease (COPD) (32), and has simultaneously been described as a one of the model systems for researching persister cells (33,34). Furthermore, clinical isolates of P. aeruginosa from CF patients have been identified as high-persistence strains (35). As such, we sought to determine some of the mechanisms which enable persister cells to evade or modulate the innate immune response.

RESULTS
The cell size of the P. aeruginosa persister cell subpopulation is more homogeneous than that of the P. aeruginosa vegetative population. Persister cells are a subpopulation of the regular vegetative bacterial population. As such, we initiated this work by determining whether there is a substantial difference in cell size between the vegetative population of P. aeruginosa and isolated persister cells. Persister cells were isolated using ciprofloxacin exposure for a period of 24 h where a biphasic killing was present and cells had an exponential killing in the first 3 h followed by a plateau from that point onwards, as previously described (3,12,27,33,36,37). To ensure that no dead cells were present, we performed several centrifugation/washing steps and the resulting pellet at the end was significantly reduced in total cell counts, as expected as, the dead cells were removed (Fig. S1). In addition, we have also performed live/dead staining using SYTO9 (stains all cells) and propidium iodide (stains cells with impaired membranes) to determine whether the final pellet of persister cells and regular cells contained similar ratios of live/dead ( Fig. 1A; Fig. S1B and C). We found that no significant difference was detected between the persister, and regular cell populations and that the dead cells' control (ethanol treated cells) had more than 90% of dead cells. While overall there was no significant difference (P . 0.05) in the median forward-scattering (cell size) or side-scattering (cell granulation) between regular vegetative cells and persisters, the latter had a narrower peak, indicating a less heterogeneous population, regarding cell size ( Fig. 1B to F), consistent with the fact that cells, while in the persister state, are not dividing. However, these cells have been shown to be phenotypically different in Escherichia coli (38)(39)(40).
P. aeruginosa persister cells are tolerant to MAC-mediated killing despite being opsonized by C3b. We next determined the interaction of P. aeruginosa persister cells with human complement both for MAC-mediated killing (Fig. 2) and opsonization (Fig. 3). Due to its ability to inhibit and resist MAC-mediated killing (41,42), Staphylococcus aureus was used as a negative control ( Fig. 2A and B), while due to its susceptibility to MAC-mediated killing (43), E. coli was used as a positive control ( Fig. 2C and D). In the presence of serum with inactivated complement, persister cells of each species reverted to an active dividing state (Fig. 2  B, D, and F). As anticipated, S. aureus viability of both regular vegetative ( Fig. 2A) and persister (Fig. 2B) cells was unaffected by the presence of complement ( Fig. 2A and B) while viability of E. coli regular vegetative cells was reduced to the point of eradication (Fig. 2C). Contrary to what was previously described (43), P. aeruginosa regular vegetative cells (Fig. 2E) were eradicated, albeit at a lower rate initially when compared to E. coli regular vegetative cells (Fig. 2C). Both E. coli (Fig. 2D) and P. aeruginosa (Fig. 2F) persister cells were initially killed at an increased rate; however, by 1.5 to 3 h presented a biphasic killing trend, becoming resilient to killing by complement. Immune Response to Persister Cells mBio P. aeruginosa persister cells are opsonized by C3b similarly to regular vegetative cells but have reduced bound C5b. To establish whether the resilience of P. aeruginosa to complement killing was due to an inability of C3b (initiating protein of the alternate complement pathway) and/or C5b (initiating protein of the MAC formation) to bind persister cells, we used an anti-C3 antibody to detect and quantify the C3b binding to the cells (Fig. 3) and performed ELISA for C5b protein quantification (Fig. 4). The binding quantification was performed by FACS (Fig. 3A) and further confirmed by microscopy ( Fig. 3B). We found no significant difference (P . 0.05) of C3b binding between persister and regular vegetative cell populations albeit persister cells having a aeruginosa (E and F) were exposed to 90% complete human serum (PBS1Serum) (closed square), PBS (closed circle), or heat-inactivated serum (closed triangle) for a period of 24 h. Experiments were performed in quadruplicate. Results were analyzed according to the one-way ANOVA with Tukey's post-test (*, P , 0.05; **, P , 0.01; ***, P , 0.001) and are presented as mean 6 SD.

FIG 3
Binding of complement factor C3b to cells. Persister and regular vegetative cells were isolated and exposed to human serum containing complement for a period of 30 min. Upon ending the complement reaction with EDTA, the cells were stained with Baclight red, and complement was immunostained (green). (A) Represents the relative fluorescence of red stain and green stain in regular and persister cells upon exposure to complement. (B) Represents the ratio of cells to antibody. A total of 20 images were used in this experiment and analyzed using the Intensity Luminance Software. No statistical difference was found between the staining of regular versus persister cells, as determined by one-way ANOVA. Results presented as mean 6 SD.
Immune Response to Persister Cells mBio clear bi-modal pattern of binding (Fig. 3B). We also found that after 1.5 h and 3 h of incubation in human serum, significantly less C5b was deposited on the viable persister cells' membranes relative to viable regular cells, while after 24 h no regular cells were viable (Fig. 2E), resulting in much more C5b deposition per viable cell in the persister population (Fig. 4). No significant change in C5b deposition on persister cells was detected from 3 h onwards (P . 0.05) (Fig. 4).
Macrophages can engulf P. aeruginosa persister cells, albeit at a lower rate, but do not kill them. Upon determining that persister cells were not killed by the membrane attack complex function of complement (Fig. 2), were opsonized with C3b similarly to regular vegetative cells (Fig. 3), and had lower C5b bound (Fig. 4), we quantified the macrophages' ability to engulf P. aeruginosa persister cells with and without prioropsonization. THP-1 macrophages were exposed to the same inoculum of bacteria, whether regular or persister, for 30, 60, 90, and 180 min and engulfment was evaluated based on intracellular bacterial viability (Fig. 5). Persister cells of P. aeruginosa were engulfed significantly less (P , 0.001) by THP-1 macrophages compared to regular vegetative cells, with an overall 100-fold decrease (Fig. 5). As anticipated, opsonization of regular vegetative cells did not result in a significant change in engulfment (44). However, we expected a change to occur for persister cells; however, although the engulfment was slightly higher following opsonization of persister cells, it was neither significant (P . 0.05) nor to the level of regular vegetative cells (Fig. 5).
P. aeruginosa persister cells are resilient to killing by macrophages. Typically following engulfment, the phagosome fuses with a lysosome and the engulfed pathogenic organisms are eliminated. Thus, once it was established that persister cells were resilient to killing by complement (Fig. 2) and were engulfed at a lower rate (Fig. 5) when compared to regular vegetative cells, we decided to further explore their resilience to killing once inside the macrophages. Clearance of P. aeruginosa persister cells was quantified by infecting macrophages for 1.5 h, subsequently removing all external bacteria, and then allowing the macrophages to kill the intracellular bacteria for a period of 24 h. The number of viable intracellular regular vegetative cells present was significantly reduced (P , 0.05) by a total of 1.4 Logs (Fig. 6A), while no change of cell viability was detected for the intracellular persister cells (Fig. 6A), indicating a lack of killing by macrophages. The viable cell count at 24 h postinfection was similar for infections with both regular vegetative and persister cells (Fig. 6A). FACS analysis of internalized cells within macrophages, at 90 min of infection and FIG 4 Binding of complement factor C5b to cells. Persister and regular cells were isolated and exposed to human serum containing complement for a period of 1.5, 3, and 24 h. Upon ending the complement reaction with EDTA, the cells were harvested and the presence of C5b on the cell envelope was quantified using ELISA. Results presented as mean 6 SEM. Statistical significance was determined by ANOVA with Tukey's post-test, *, P , 0.05; **, P , 0.005; ***, P , 0.001; ****, P , 0.0001.
Immune Response to Persister Cells mBio 24 h postinfection, where bacteria were stained with Baclight red, also showed fewer regular vegetative cells (with a shift of the fluorescence peak) but not fewer persister cells (Fig. 6B). In addition, we also quantified the 16s rRNA gene expression, to determine whether the P. aeruginosa cells were active in the macrophages upon engulfment and found that it was decreased at time zero by 5.6-fold 6 1.7 and no change at time 24 (1.  (Fig. 6), while also being engulfed at a lower rate (Fig. 5). Therefore, we questioned whether upon FIG 5 Macrophage infections. Differentiated THP-1 monocytes were exposed to P. aeruginosa regular and persister cells for 30, 60, 90, and 180 min. Regular vegetative cells were diluted to match viable persister cell concentrations before inoculation. Bacterial viability was quantified at the various time points of the infection. Results were analyzed according to the one-way ANOVA with Tukey's posttest and are presented as mean 6 SEM. Experiments were performed in quadruplicate. Engulfment of persister cells was significantly lower comparing to regular vegetative cells (P , 0.001) but no significant difference (P , 0.05) was observed when cells were exposed to complement.  (46), using flow cytometry (Fig. 8). When quantifying the cell membrane markers, we found that after 1.5 h of infection, macrophages infected with vegetative P. aeruginosa cells expressed high levels of CD80/CD86, but not CD206 (Fig. 8). In contrast, macrophages infected with persister cells expressed both high levels of CD80/CD86 and CD206 (Fig. 8). These results suggest that infections with persister cells elicit macrophage polarization toward a M2 response while still retaining M1-associated cell surface proteins. Regarding cytokine secretion, we found that in the first 1.5 h, all inflammatory cytokines were secreted at lower levels by macrophages infected with persisters compared to infections by regular vegetative cells, but higher than unchallenged macrophages ( Fig. 7 A to C). Similarly, the anti-inflammatory IL-10 also presented that pattern at 0.5 h; however, at 1.5 h minutes of infection, IL-10 secretion was significantly higher in persisterinfected macrophages than in infections with regular vegetative cells (Fig. 7D). This high anti-inflammatory response coincides with the plateau of engulfment established between 0.5 and 1.5 h for persister cells (Fig. 4), and the consistently lower engulfment of persister cells. However, at 24 h of infection with persister cells, a bi-modal trend was present in IL-6 ( Fig. 7A) and CXCL-8 ( Fig. 8C) with an overall increase in secretion, compared to 1.5 h, while IL-10 was at levels identical to uninfected macrophages (Fig. 7D). These changes were anticipated as, similarly to what occurs in infections in Immune Response to Persister Cells mBio vivo (27), a percentage of the persister population reverted into an active metabolic state due to the presence of nutrients in the medium following 7 h of incubation, albeit at significantly lower levels than regular vegetative cells (Fig. S2), and as such, should activate the proinflammatory response while tampering the anti-inflammatory response as demonstrated by the decrease of IL-10. In infections with regular vegetative cells, the IL-6 ( Fig. 7A) and TNF-a (Fig. 7B) inflammatory response continued to increase in the first 1.5 h, remaining constant at 24 h, while CXCL-8 decreased by 24 h (Fig. 7C). IL-10 remained constant in the first 1.5 h, decreasing to uninfected control levels by 24 h (Fig. 7D).
To determine whether the variation of cytokine secretion of macrophages upon infections with persister and regular cell infections (Fig. 7) was transcriptionally regulated, we quantified the relative expression of the genes related to the several cytokines (Fig. S3). We found that except for the 24 h time point of IL-6, no significant difference in the gene expression quantification was present, indicating that the macrophage's transcription of cytokine genes is equally activated when exposed to both cell types. As such, posttranscriptional regulation must be occurring as the cytokine secretion is significantly different between infections with regular and persister cells (Fig. 7). In infections with both bacterial populations cytokine mRNA is still transcribed, but due to the persister cells' low metabolic status, there is a reduction/absence of microbial products which triggers the macrophage response into an event similar to the clearance of microbial products in vivo, where it is known that the mRNA coding for cytokines becomes unstable resulting in a reduction of translation (47) and an absence of bacterial elimination. Thus, this links the absence of microbial products to posttranscriptional control, which is normally used to prevent unwarranted cytokine production, explaining the intermediate cytokine secretion (Fig. 7) in infections with persister cells and the lack of elimination once engulfed (Fig. 6).

DISCUSSION
The innate immune response to bacterial persister cells remains ambiguous, despite their hypothesized role in chronic and resilient infections. Previously it has been found that a persister state can be induced by the immune system in several bacterial species (24,25,29,30), and that persister cells are engulfed at a lower rate following infections (27,28) and can survive once inside macrophages (24)(25)(26).
In this study, we describe a mechanism of the effect of persister cells on the immune response, while describing their resilience to other aspects of innate immunity namely, MAC-mediated complement killing and macrophage killing. Our study provides evidence Immune Response to Persister Cells mBio that P. aeruginosa persister cells (i) resist both MAC-mediated complement killing and macrophage killing albeit being opsonized by C3b, and (ii) elicit the polarization of macrophages toward a M2 response, which switches to a M1 response upon persister awakening. To our knowledge, this is the first comprehensive study of the overall innate immune response to P. aeruginosa persister cells. We found that there is a decreased susceptibility to MAC-mediated killing in P. aeruginosa and E. coli persister cells compared to regular-metabolically active-bacterial cells (Fig. 2). This resistance to MAC-mediated killing was due to a decrease of C5b binding (Fig. 4) but not due to a reduction of C3b deposition on the bacterial surface (Fig. 3) as previously hypothesized for the evader phenotype (43), which was most likely a subset of the persister cell phenotype as the persister cell population in P. aeruginosa PA14 consists of 0.1% to 0.001% of the regular vegetative population (12), consistent with evader phenotype values (43). Functional C5 convertases have been observed on C3b-opsonized P. aeruginosa (48); therefore, it is possible that these convertases are less functional on the surface of persister cells, leading to decreased deposition of C5b. Furthermore, this is in accordance with previous findings where, in E. coli, exposure to human serum has resulted in the induction of both the persister and the viable but nonculturable (VBNC) state (49) which has previously been shown, in E. coli, to describe the same bacterial stress state (50). The resilience to complement killing in both E. coli and P. aeruginosa could be due to changes in the cell membrane and outer membrane, as previously it was reported that several outer membrane proteins (OprF, OprB, OprD, and OprM) and the chaperone protein SurA were present in higher abundance in P. aeruginosa cells reverting from a persister state (12). In the absence of SurA, bacterial cells are highly susceptible to complement (51) while OprF is a complement C3 binding acceptor molecule (52), and its absence reduces the bacterial escape from phagosome vacuoles (53). A similar process could be occurring in E. coli as OmpA, the homolog to OprF, has been implicated in C3 convertase inhibition (52) and the inhibition of the classical complement pathway (54). It has also been established that P. aeruginosa can cleave complement protein C3 through binding complement Factor H via cell surface-associated proteins Tuf (55) and LpD (56). Persister-like cells in P. aeruginosa, E. coli, and four other relevant human pathogens show tolerance to eradication by complement-mediated lysis; however, these cells require a level of metabolic activity to persist in blood (43).
Similar to previous work that reported phagocytosis of S. aureus persister cells (27) and Mycobacterium tuberculosis (28) dormant cells was significantly lower than active/ regular vegetative cells, persister cells of P. aeruginosa were engulfed at a lower rate when compared to regular vegetative cells. However, this was independent of bacterial cell opsonization (Fig. 3). Once engulfed, P. aeruginosa persister cells numbers remained constant, indicating a lack of killing, contrary to regular vegetative cells where the viable cell level was reduced to persister cell levels (Fig. 6). As such, it seems that P. aeruginosa switches to a persister state once inside the macrophages, similarly to the intracellular pathogen Listeria monocytogenes which switches to a persistence phenotype when found in vacuoles (26). However, the fate of persister cells after engulfment remains mostly unclear, and further studies need to be performed, as both S. typhimurium and M. tuberculosis persister cells are metabolically active following engulfment by macrophages (24,25). Furthermore, P. aeruginosa uses type III secretion system (T3SS) proteins to attack host phagocytes (57), and these proteins have been shown to accumulate in P. aeruginosa persister cells, killing host immune cells (58). We did not, however, detect changes in the macrophage numbers postinfection (data not shown).
When examining transcription of several cytokines characterized in M1 and M2 polarizations (Fig. S3), we found that infections with both persister and regular vegetative cells result in a similar gene transcription supporting the hypothesis that both cell populations activate the immune system, albeit with posttranscription or translation modifications. The killing of the regular vegetative cells of P. aeruginosa (Fig. 6) indicates that macrophages are activated upon infection, as further evidenced by them being CD80 1 /CD86 1 CD206 2 (Fig. 8) together with the secretion of high levels of CXCL-8, IL-6, and TNF-a, and were not deterred by the initial high concentration of IL-10 (Fig. 7D), compared to uninfected cells. In contrast, macrophages infected with persister cells were CD80 1 /CD86 1 /CD206 1 (Fig. 8) and were initially intermediately activated when exposed to persister cells, as shown by their secretion levels of IL-6, CXCL-8, and TNF-a, compared to uninfected and regular-infected macrophages, followed by a tampering down of the proinflammatory response, due to the high IL-10, resulting in a lack of elimination of the intracellular bacteria when the infection is stopped at 1.5 h (Fig. 6), previous to persister cell reversion to an active state in RPMI medium (Fig. S2), as demonstrated to occur in when exposed to heatinactivated serum (Fig. 2F), and known to occur upon the removal of stress (1,2,45). These results are supported by previous findings described for S. typhimurium persisters (45) where it was found that S. typhimurium persisters induced anti-inflammatory polarization of macrophages and extended the survival of the bacteria within the host (45). Additionally, when M. tuberculosis chronically infects the lungs of wild-type mice, a subpopulation of dormant cells is present, whereas mice lacking in interferon-g lack this subpopulation (25), suggesting that the presence of host cytokines is important to the persistence of M. tuberculosis during infection.
From our findings, we propose that the mechanism of infection and immune system modulation between regular vegetative cells and persister cells is distinct (Fig. 9). Regular vegetative cells induce a macrophage favored polarization toward M1 (CD80 1 / CD86 1 /CD206 2 , high levels of CXCL-8, IL-6, and TNF-a), while persister cells initially induce a polarization favoring M2, more specifically M2b (CD80 1 /CD86 1 /CD206 1 , high levels of IL-10, and intermediate levels of CXCL-8, IL-6, and TNF-a), which is then skewed toward M1 polarization, by 24 h of infection, once the internalized persister cells revert into an awakened metabolically active state.
In conclusion, we found that in addition to being tolerant to antibiotics, persister cells are also resilient to an immune system attack/response where they are not eliminated by MAC-mediated killing, as demonstrated by the decrease of bound C5b, despite being bound by C3b, and elicit an intermediate anti-inflammatory response by triggering macrophage M2b favored polarization. This study sheds further light as to how persister cells modulate the immune response and survive in the host during an infection. By escaping/ resisting the immune response, persister cells can later become active dividing cells and reinfect the host, further confirming that these cells are involved in chronic and recurrent Immune Response to Persister Cells mBio infections. Despite these advances, there remain many unknowns relating to persister cell behavior when infecting a host, and how the immune response to persister cells occurs in other bacterial species.
Isolation of persister cells. Persister cells were isolated as described previously (1)(2)(3)(4)(5)(6). Briefly, isolation streak plates of S. aureus, P. aeruginosa, and E. coli were prepared on 100% LB agar and incubated at 37°C for 24 h. Planktonic overnight cultures were prepared by removing an isolated colony from the streak plate and inoculating it into 100% LB broth and grown at 37°C with agitation (220 rpm) for a period of 24 h. Cells were then collected (16,000 Â g for 5 min at 4°C), washed twice with saline (16,000 Â g for 5 min at 4°C), and subsequently resuspended in either saline (0.85% NaCl) or ciprofloxacin (20Â MIC) in saline to a final OD600 of 1.6. Ciprofloxacin was used as means to induce oxidative DNA damage which results in an accumulation of persister cells (39). Cultures were subsequently incubated at 37°C with agitation (220 rpm) for a period of 24 h. Bacterial cells were collected via centrifugation (16,000 Â g for 5 min at 4°C). The cells were then resuspended in saline and washed two times by centrifugation (16,000 Â g for 5 min at 4°C). These 2 washes were performed to remove lysed dead cells, as ciprofloxacin was previously demonstrated to lyse cells of P. aeruginosa, E. coli, and Enterobacter cloacae (59)(60)(61). Once the first wash is performed, a large amount of biomass-the dead lysed cells-was removed (Fig. S1) and only live cells were present in the final resuspension. Ciprofloxacin concentrations used were 20Â the MIC and consisted of 50 mg/L for S. aureus cells and 20 mg/L for E. coli and P. aeruginosa (3,12,24,30,33,34). Viability of persister and regular vegetative cells was determined at 0, 1, 3, 6, and 24 h, by serial dilution and drop plating on 1:2 plate count agar (PCA) with 1% MgCl 2 Á7H 2 O for the inactivation of ciprofloxacin. Cell viability was also determined by staining persister and regular cells with propidium iodide and SYTO9 (Thermo Fisher Scientific), where after a 15-min incubation, the cells were washed (to remove excess of stain) with PBS and resuspended. Bacterial fluorescence (from the stains) was measured using a SpectraMax I3x Multi-Mode plate reader, Molecular Devices. We also used dead cells-cells exposed to 70% ethanol for 30 min-as a control.
Growth of Pseudomonas aeruginosa in RPMI 1640. P. aeruginosa regular and persister cells were isolated as above, collected by centrifugation, and washed in 0.85% saline three times by centrifugation (16,000 Â g for 5 min at 4°C). Each population was then resuspended in RPMI at a final OD 600 of 1.5. Each population was added to a 96-well plate and changes in absorbance (595 nm) was monitored every hour, for 24 h in a microtiter plate reader (Beckman DTX880) at 37°C.
Activation of metabolism in Pseudomonas aeruginosa in RPMI 1640. Regular and persister cells of P. aeruginosa MPAO1 attTn7::P(A1/04/03)::GFPmut (62), constitutively expressing GFP, were isolated as above, collected by centrifugation, and washed in 0.85% saline three times by centrifugation (16,000 Â g for 5 min at 4°C). Each population was then resuspended in RPMI to a final OD 600 of 1.5. Each population was added to a 96-well plate and changes in fluorescence (excitation 488 nm, emission 509 nm) were monitored for 24 h, at 30 min intervals (SpectraMax I3x Multi-Mode plate reader, Molecular Devices).
Effect of human serum on regular and persister cells. Regular vegetative and persister cells of P. aeruginosa, S. aureus, and E. coli were collected by centrifugation, washed in 0.85% saline three times by centrifugation (16,000 Â g for 5 min at 4°C). Samples were then resuspended in 0.85% saline at a final concentration of 10 7 cells/mL, and subsequently in either 90% complete human serum, PBS, or heat-inactivated serum. Cells were then incubated at 37°C, and cell viability was determined by adding 10 mL of 10 mM EDTA to stop the reaction at 0, 0.75, 1.5, 3, and 24 h of incubation followed by serial dilutions and plating of bacteria on 1:2 PCA with MgCl 2 Á7H 2 O for 48 h at 37°C. Controls consisted of cells exposed to heat-inactivated human serum (adapted from 63). Experiments were performed in quadruplicate.
Complement binding. P. aeruginosa regular and persister cells were collected by centrifugation, resuspended in 0.85% saline to a concentration of 10 7 cells/mL. The saline was supplemented with 100 mL of PBS or 100 mL of complete human serum to a final concentration of 10%. Cultures were incubated for 30 min to allow for binding of complement proteins to the bacterial cells, after which, the complement activity was stopped with 10 mM EDTA. Cells were then washed by centrifugation (16,000 Â g for 4 min at 4°C), unbound proteins from the serum were decanted, and cells were resuspended in PBS. Serial dilutions of each population were performed, and each dilution was stained. Bacterial cells were stained with BacLight Red (6 ng, Thermo Fischer Scientific, Waltham, MA, USA), and complement protein C3b was labeled with fluorescent Anti-C3 antibody (35 mg MP Bio). Samples were imaged using an epifluorescence microscope (Olympus BX46) at Â100 magnification. Experiments were performed in quadruplicate with five images being taken per sample. Images were analyzed using Intensity Luminance V1 software (64). To assess the binding of C5b, regular cells and persister cells of P. aeruginosa were collected by centrifugation, washed in 0.85% saline three times by centrifugation (16,000 Â g for 5 min at 4°C) then resuspended in 0.85% saline at a final concentration of 10 7 cells/mL, and subsequently incubated in either 90% complete human serum or PBS for a period of 1.5, 3, or 24 h. The complement reaction was then stopped with 10 mM EDTA, and the cells were washed in 0.85% saline three times by centrifugation (16,000 Â g for 5 min at 4°C). The cell pellets were resuspended in PBS and then ELISAs were executed per manufacturer's instructions to measure membrane bound C5b (Novus Biologicals).

Immune Response to Persister Cells mBio
Maintenance and differentiation of THP-1 macrophages. THP-1 monocytes were cultured in suspension on RPMI 1640 complemented with 10% fetal bovine serum, with media changes every 2 to 3 days. Cultures were split once they reached 8 Â 10 5 cells/mL. THP-1 monocytes were seeded at a concentration of 5 Â 10 5 cells/mL onto 24-well plates and differentiated into M0 Macrophages via the introduction of 100 nM phorbol 12-myristate 13-acetate (PMA) for 3 days at 37°C 5% CO 2 , after which they were ready for the experimental procedures (27,65).
Infection of THP-1 macrophages. P. aeruginosa persister and regular vegetative cells were isolated and resuspended in infection media (27,65) and standardized to 5 Â 10 6 cells/mL. Infections were initiated with an MOI of 10:1 and incubated for different time periods, including, 30, 60, 90, and 180 min. Following the infection period, THP-1 macrophages were washed twice with PBS and subsequently exposed to gentamicin (40 mg/L) for 1 h to remove any remaining extracellular bacteria (27,65). Macrophages were then lysed with 10% Triton X-100 for 45 min and intracellular bacteria viability was quantified as described above. Experiments were performed in quadruplicate.
Opsonization and engulfment of P. aeruginosa persister cells. P. aeruginosa regular vegetative cells and persister cells were incubated in a solution of 10% human serum or PBS for a period of 30 min (63). The complement reaction was stopped with 10 mM EDTA, cells were washed via centrifugation (16,000 Â g for 5 min at 4°C), followed by resuspension in PBS, and subsequently used to infect THP-1 macrophages as described above. Experiments were performed in quadruplicate.
Elimination assays of intracellular bacteria. To assess the elimination of intracellular bacteria, infections were performed for a period of 90 min as described above. Once infections were stopped, and following the gentamicin exposure, infection media was replenished, and the cultures were incubated for further 24 h. After incubation, macrophages were lysed with 10% Triton X-100 for 45 min and intracellular bacteria viability was quantified as described above (66). Experiments were performed in quadruplicate.
Flow cytometry. To assess the size of persister cells relative to regular vegetative cells, persister cells were isolated as above, stained with BacLight Red (6 ng, Thermo Fischer Scientific, Waltham, MA, USA) and then fixed with paraformaldehyde. Flow cytometry was performed using the BD Accuri C6 Plus system and the data were analyzed with the flow cytometry software FlowJo. Similarly, to assess the elimination of persister cells, THP-1 macrophages were infected as above with prestained bacteria for 90 min or 24 h, followed by fixation with paraformaldehyde and flow cytometry analysis as above. To determine macrophage polarization, THP-1 macrophages were infected as above with either regular or persister cells for 90 min followed by immunostaining for M1/M2 cell-surface marker proteins CD80 (Phycoerythrin [PE] ANTI-HU CD80, from Biolegend), CD86 (PE ANTI-HU CD86, from Biolegend), and CD206 (Allophycocyanin [APC] ANTI-HU CD206, from Biolegend), and DAPI (Thermo Fisher Scientific) DNA staining for 30 min. The cells were then fixed with paraformaldehyde and flow cytometry analysis was performed with the Bio-Rad ZE5 Cell Analyzer and FlowJo. Experiments were performed in quadruplicate.
Quantitative reverse transcriptase PCR. Relative transcriptional levels of THP-1 innate immune genes and engulfed P. aeruginosa 16s rRNA were quantified. To quantify the relative expression, infections were performed as described above. At the end of the incubation with gentamicin, TRIzol reagent was added to macrophages, the contents of each well were collected, and RNA was isolated using the Zymo RNA purification kit (Zymo Research). A total of 0.5 mg of RNA was used for cDNA synthesis, and cDNA was generated using QScript cDNA Synthesis kit. Quantitative reverse transcriptase PCR (qRT-PCR) was performed with an Eppendorf Mastercycler Ep Realplex instrument (Eppendorf AG, Hamburg, Germany) and the Kapa SYBR Fast qPCR kit (Kapa Biosystems, Woburn, MA) with the oligonucleotides for THP-1 cells (obtained from Qiagen) and P. aeruginosa 16s rRNA (12). No template controls (NTC) and no reverse transcriptase (NRT) reactions were executed to confirm the lack of DNA contaminants during sample and mastermix preparation. Relative transcript quantitation was accomplished using Ep Realplex software (Eppendorf AG), with the transcript abundance (based on the threshold cycle [CT] value) being normalized to the housekeepers gadph for THP-1 (67) and cysD (FW: CTGGACATCTGGCAATACAT; RV: TCTCTTCGTCAGAGAGATGC) for P. aeruginosa (68) before the determination of transcript abundance ratios. Single-product amplification verification was accomplished through analysis of the melting curves. Experiments were performed at least in quadruplicate.
ELISAs of secreted cytokine. Macrophages were infected with P. aeruginosa regular and persister cells and the supernatant was collected at 0.5, 1.5, and 24 h of infection. The supernatant was centrifuged for 5 min at 16,000 Â g to remove bacterial cells, and the resulting solution was assessed for the presence of cytokines. Samples were diluted up to 1:100, and ELISAs were performed to quantify protein concentration, per manufacturer's instructions using the following kits (Invitrogen, Carlsbad, CA, USA): IL-10 (BMS2152), TNF-a (BMS223HS), IL-6 (BMS213HS), and CXCL-8 (KHC0081). At the end of each assay, the absorbance of each sample was determined at 450 nm (DTX880 multimode detector, Beckman Coulter, CA). Cytokine concentrations were determined using standard curves generated in each assay, then accounting for the dilution factor. Experiments were performed at least in quadruplicate.
Statistical analysis. All data were analyzed using GraphPad Prism 9.3.1. One-way ANOVA was performed for multivariant analysis followed by Tukey's or Dunnett's multiple-comparison tests.

SUPPLEMENTAL MATERIAL
Supplemental material is available online only. No conflict of interest declared.