Lipopolysaccharide from Porphyromonas gingivalis, but Not from Porphyromonas endodontalis, Induces Macrophage M1 Profile

Apical Lesions of Endodontic Origin (ALEO) are initiated by polymicrobial endodontic canal infection. Porphyromonas gingivalis (Pg) and Porphyromonas endodontalis (Pe) lipopolysaccharides (LPS) can induce a pro-inflammatory macrophage response through their recognition by TLR2 and TLR4. However, polarization responses induced by Pg and/or Pe LPS in macrophages are not fully understood. We aimed to characterize the polarization profiles of macrophages differentiated from THP-1 cells following Pg and/or Pe LPS stimulation from reference strain and clinical isolates. A modified LPS purification protocol was implemented and the electrophoretic LPS profiles were characterized. THP-1 human monocytes differentiated to macrophages were stimulated with Pg and Pe LPS. Polarization profiles were characterized through cell surface markers and secreted cytokines levels after 24 h of stimulation. TLR2 and TLR4 cell surfaces and transcriptional levels were determined after 24 or 2 h of LPS stimulation, respectively. LPS from Pg induced a predominant M1 profile in macrophages evidenced by changes in the expression of the surface marker CD64 and pro-inflammatory cytokine profiles, TNF-α, IL-1β, IL-6, and IL-12. Pe LPS was unable to induce a significant response. TLR2 and TLR4 expressions were neither modified by Pg or Pe LPS. Pg LPS, but not Pe LPS, induced a macrophage M1 Profile.


Introduction
Apical periodontitis (AP) is a prevalent inflammatory disease caused by dental infection comprising the pulpal tissue of the root canal system following a carious lesion of the tooth, which naturally evolves to form an apical lesion of endodontic origin (ALEO), a hallmark of the disease [1]. AP represents the main cause of tooth loss and has been increasingly linked with low-grade systemic inflammation and, consequently, a higher risk of cardiovascular diseases and diabetes mellitus [2][3][4].
The development and progression of ALEO depends on the dynamic balance between the bacterial consortia, their respective virulence factors, and the host's immune response [5][6][7]. Although endodontic infections have a polymicrobial nature, black-pigmented bacteria, such as Prevotella and Porphyromonas spp., have been recognized as central agents in the development and progression of the lesion [8][9][10]. Among them, Porphyromonas endodontalis (Pe) and

Results
We characterized and visualized in silver-stained Tris-Tricine gels the electrophoretic profiles of bacterial lysates and their respective LPS extracts from reference strains P. gingivalis ATCC 33277, P. endodontalis ATCC 35406, and clinical isolates P. gingivalis (AC1Pg) and P. endodontalis (AC2Pe). The clinical isolates were previously obtained from patients with asymptomatic AP (AAP).
First, LPS from P. gingivalis ATCC 33277 and AC1Pg lysates consisted of an unsubstituted Lipid A-core region and a preferential length of lipid A-core substituted with more than eight repeating units of OAg. On the other hand, the electrophoretic profiles of the LPS obtained from both the reference strain and its respective clinical isolate (AC2Pe), showed a weaker unsubstituted lipid A-core region compared to that of P. gingivalis, and a preferential OAg substitution length of 8 to 14 repeated units (Figure 1a).
LPS purification with the modified TRIzol extraction protocol, consisting of an additional sonication step to ensure wall rupture, and treatments with DNase I, RNase, and proteinase K effectively removed nucleic acids and protein contaminants. The resultant P. gingivalis and P. endodontalis LPS retained their endotoxic activity (>85 EU/mL) and conserved their structural KDO, a component of the LPS core. The purified P. gingivalis LPS obtained from both the reference strain and its clinical isolate (AC1Pg), presented an electrophoretic profile similar to that observed in bacterial lysates, showing that the extraction protocol did not affect LPS integrity. In the case of the purified P. endodontalis LPS, the unsubstituted lipid A-core region was not visible (Figure 1b). However, both lysates and purified LPS exhibited a preferential length of 8 to 14 OAg repeated units. gingivalis LPS obtained from both the reference strain and its clinical isolate (AC1Pg), presented an electrophoretic profile similar to that observed in bacterial lysates, showing that the extraction protocol did not affect LPS integrity. In the case of the purified P. endodontalis LPS, the unsubstituted lipid A-core region was not visible (Figure 1b). However, both lysates and purified LPS exhibited a preferential length of 8 to 14 OAg repeated units. Subsequently, we assessed P. gingivalis and P. endodontalis LPS capacity to induce polarization on THP-1 macrophages. For this purpose, cell cultures were stimulated for 24 h with LPS (10 µg/mL) from P. gingivalis ATCC 33277, P. gingivalis AC1Pg, P. endodontalis ATCC 35406, or P. endodontalis AC2Pe or co-stimulated with P. gingivalis ATCC 33277 LPS plus P. endodontalis ATCC 35406 LPS (10 µg/mL each). E. coli LPS stimulation (10 ng/mL) was used as a positive control.
Macrophage polarization was assessed as the percentage of double-positive cells for M1 macrophages (CD64 + CD80 + ), or M2 (CD163 + CD206 + ) and the M1/M2 ratio on THP-1 macrophage. The percentages of single-positive cells and the mean fluorescence intensity (MFI) of each surface marker ( Figure 2) were also assessed.
P. gingivalis LPS, specifically LPS from AC1Pg, showed a tendency to increase the M1 population over unstimulated macrophages by 12.7%. P. gingivalis LPS from ATCC 33277 and LPS from AC1Pg showed a tendency to increase the M1/M2 ratio to 1.5 and 4.7 respectively, although the changes were not significant (p > 0.05).
On the other hand, neither varieties of forms of P. endodontalis LPS did not induce changes in surface marker percentages, nor did co-stimulus with P. gingivalis ATCC 33277 LPS and P. endodontalis ATCC 35406 LPS. None of the evaluated LPS conditions influenced the percentage of double-positive M2 macrophages (Figure 2a-c).
Regarding single positive M1 macrophage populations on unstimulated THP-1 macrophages, 71.3% of the total population were CD64 + , but only 2% were CD80 + . E. coli LPS stimulus induced a decrease in the CD64 + population down to 47.3% but increased CD80 + macrophages up to 42.4% (Figure 2d,e). P. gingivalis LPS, P. endodontalis LPS, and co-stimuli did not significantly change the percentage of CD64 + nor CD80 + cell population.
CD64 surface levels (MFI) were reduced by P. gingivalis LPS, specifically AC1Pg LPS, below unstimulated control levels (p < 0.05). P. endodontalis LPS and the co-stimulus did not promote changes in CD64 nor CD80 surface levels (Figure 2h,i).
Evaluation of single positive M2 macrophage populations showed no significant changes in any of the evaluated conditions. P. gingivalis and/or P. endodontalis LPS did not change CD163 or CD206 surface levels (MFI), although E. coli LPS induced a CD163 increase above unstimulated control (Figure 2f,g,j,k).
In summary, neither P. gingivalis LPS nor P. endodontalis LPS induced changes in the percentage of cell surface macrophage markers in vitro. Surface levels of CD64 (MFI) were significantly reduced by LPS from clinical isolates of P. gingivalis. P. gingivalis LPS, specifically LPS from AC1Pg, showed a tendency to increase the M1 population over unstimulated macrophages by 12.7%. P. gingivalis LPS from ATCC 33277 and LPS from AC1Pg showed a tendency to increase the M1/M2 ratio to 1.5 and 4.7 respectively, although the changes were not significant (p > 0.05).
On the other hand, neither varieties of forms of P. endodontalis LPS did not induce changes in surface marker percentages, nor did co-stimulus with P. gingivalis ATCC 33277 LPS and P. endodontalis ATCC 35406 LPS. None of the evaluated LPS conditions influenced the percentage of double-positive M2 macrophages (Figure 2a-c).
Regarding single positive M1 macrophage populations on unstimulated THP-1 macrophages, 71.3% of the total population were CD64 + , but only 2% were CD80 + . E. coli LPS stimulus induced a decrease in the CD64 + population down to 47.3% but increased CD80 + macrophages up to 42.4% (Figure 2d,e). P. gingivalis LPS, P. endodontalis LPS, and co-stimuli did not significantly change the percentage of CD64 + nor CD80 + cell population.
CD64 surface levels (MFI) were reduced by P. gingivalis LPS, specifically AC1Pg LPS, below unstimulated control levels (p < 0.05). P. endodontalis LPS and the co-stimulus did not promote changes in CD64 nor CD80 surface levels (Figure 2h,i).
Evaluation of single positive M2 macrophage populations showed no significant changes in any of the evaluated conditions. P. gingivalis and/or P. endodontalis LPS did not change CD163 or CD206 surface levels (MFI), although E. coli LPS induced a CD163 increase above unstimulated control (Figure 2f,g,j,k). Next, the cytokine profiles associated with M1 (TNF-α, IL-1β, IL-6, and IL-12) and antiinflammatory M2 (IL-10 and IL-1RA) were evaluated in the supernatant of the stimulated THP-1 macrophages ( Figure 3). P. gingivalis LPS stimulated the secretion levels of proinflammatory cytokines similarly to E. coli LPS, but in a minor amount, compared with the unstimulated control. Specifically, P. gingivalis ATCC 33277 LPS induced increases in TNF-α and IL-1β. Moreover, LPS from AC1Pg induced increases in TNF-α, IL-1β, IL-6, and IL-12 compared to unstimulated THP-1 macrophages (p < 0.05) ( Figure 3). In the case of the anti-inflammatory cytokines, LPS from AC1Pg induced an increase of IL-10 compared to unstimulated cells (p < 0.05).
P. endodontalis LPS stimulation and the co-stimulus with P. gingivalis ATCC 33277/ P. endodontalis ATCC 35406 LPS did not induce changes in the secretion of M1 or M2 cytokines levels compared with P. gingivalis ATCC 33277 LPS ( Figure 3).
To evaluate the capability of P. gingivalis and P. endodontalis LPS to induce changes in the expression levels of TLR2 and TLR4, we assessed the percentage of positive macrophage populations for TLR2 + and TLR4 + and the MFI of TLR2 and TLR4 under the conditions previously described and, in addition, we included PAM 3 CSK 4 stimulation, a TLR2 ligand, as a positive control. P. gingivalis and P. endodontalis LPS did not show significant differences in the percentages of TLR2 + and TLR4 + cells nor induced significant changes in TLR2 or TLR4 MFI compared with non-stimulated control ( Figure 4). Although, both positive controls induced significant decrease of TLR2 + and TLR4 + populations, and TLR2 and TLR4 MFI decreased compared to the unstimulated control.
P. endodontalis LPS stimulation and the co-stimulus with P. gingivalis ATCC 33277/P. endodontalis ATCC 35406 LPS did not induce changes in the secretion of M1 or M2 cytokines levels compared with P. gingivalis ATCC 33277 LPS ( Figure 3). To evaluate the capability of P. gingivalis and P. endodontalis LPS to induce changes in the expression levels of TLR2 and TLR4, we assessed the percentage of positive macrophage populations for TLR2 + and TLR4 + and the MFI of TLR2 and TLR4 under the conditions previously described and, in addition, we included PAM3CSK4 stimulation, a TLR2 ligand, as a positive control. P. gingivalis and P. endodontalis LPS did not show significant differences in the percentages of TLR2 + and TLR4 + cells nor induced significant changes in TLR2 or TLR4 MFI compared with non-stimulated control ( Figure 4). Although, both positive controls induced significant decrease of TLR2 + and TLR4 + populations, and TLR2 and TLR4 MFI decreased compared to the unstimulated control.  Correlation analysis between the macrophage surface markers, TLR2, TLR4, and the measured secreted cytokines revealed a strong positive correlation between CD64 and TLRs and a strong negative correlation between CD64 and pro-inflammatory cytokines TNF-α, IL-1β, and IL-6 ( Table 1). Correlation analysis between the macrophage surface markers, TLR2, TLR4, and the measured secreted cytokines revealed a strong positive correlation between CD64 and TLRs and a strong negative correlation between CD64 and pro-inflammatory cytokines TNF-α, IL-1β, and IL-6 ( Table 1). Table 1. Spearman's correlation matrix of M1 and M2 surface markers, TLR2, TLR4, and secreted cytokines levels.
We finally evaluated the P. gingivalis and P. endodontalis LPS effects on TLR2 and TLR4 mRNA expression in THP-1 macrophages, using the DNA demethylating agent decitabine.
We showed that TLR2 or TLR4 mRNA levels were not affected by P. gingivalis and P. endodontalis LPS, nor did previous decitabine treatment ( Figure 5).

Discussion
Black-pigmented bacteria, such as P. gingivalis and P. endodontalis, are recognized as central agents in the etiopathogenesis and progression of AP, promoting proinflammatory and osteolytic responses due to their diversity of virulence factors [9,[13][14][15][16]20]. LPS is a distinctive component of the outer membrane of Gram-negative bacteria and a potent inducer of cytokine secretion. In the case of P. gingivalis LPS, it has been described to induce an immune response by ligation of both TLR2 and TLR4 because of the heterogeneity of its lipid A structure [36].
In the present work, we purified P. gingivalis and P. endodontalis LPS from reference strains (ATCC) and clinical isolates and then we evaluated the polarization profiles in macrophages differentiated from THP-1 after stimulation with these LPSs. We observed that P. gingivalis LPS induced an M1 profile evidenced by reduction of CD64 + cells and increased pro-inflammatory cytokines secretion (TNF-α, IL-1β, IL-6, and IL-12), whereas LPS from P. endodontalis failed to induce a significant response.

Discussion
Black-pigmented bacteria, such as P. gingivalis and P. endodontalis, are recognized as central agents in the etiopathogenesis and progression of AP, promoting pro-inflammatory and osteolytic responses due to their diversity of virulence factors [9,[13][14][15][16]20]. LPS is a distinctive component of the outer membrane of Gram-negative bacteria and a potent inducer of cytokine secretion. In the case of P. gingivalis LPS, it has been described to induce an immune response by ligation of both TLR2 and TLR4 because of the heterogeneity of its lipid A structure [36].
In the present work, we purified P. gingivalis and P. endodontalis LPS from reference strains (ATCC) and clinical isolates and then we evaluated the polarization profiles in macrophages differentiated from THP-1 after stimulation with these LPSs. We observed that P. gingivalis LPS induced an M1 profile evidenced by reduction of CD64 + cells and increased pro-inflammatory cytokines secretion (TNF-α, IL-1β, IL-6, and IL-12), whereas LPS from P. endodontalis failed to induce a significant response.
The LPS purification was done using the extraction protocol described by Yi and Hackett, which is a fast, convenient, and low-scale extraction method that uses TRIzol as the main reagent [37]. Nevertheless, we found proteins and nucleic acid traces in our LPS extracts. For this reason, we incorporated enzymatic treatments with DNase, RNase, and proteinase K into the bacterial pellet as a previous step [38,39]. This modification allowed us to remove the contaminants, except for DNA/RNA traces visualized in GelRed-stained agarose gel as an electrophoretic band with a molecular weight smaller than 100 base pairs previously undescribed. Bacterial nucleic acids have immunogenic potential by ligating endosomal TLR7 and TLR9, triggering a response similar to TLR2. To eliminate any trace of nucleic acids, we incorporated the mechanical breakdown step by sonication of the bacterial pellet before the enzymatic treatment [38]. With these modifications to the Yi and Hackett protocol, we obtained P. gingivalis and P. endodontalis LPSs free of any detectable traces of proteins and nucleic acids. However, adding extra steps to the original extraction process diminished the performance extraction, fluctuating between 50-90% less dry weight LPS than initially obtained. Despite the latter, the presence of endotoxic activity was confirmed, which was greater than 85 EU/mL, well above that reported in commercial P. gingivalis LPS (~2 EU/mL) [40]. KDO, a structural component of the LPS inner-core, was also detected in the purified P. gingivalis and P. endodontalis LPS, showing that the structure of LPS was not affected by the extraction protocol.
Our P. gingivalis LPS electrophoretic profile exhibited a high mobility region of unsubstituted lipid A-core clearly defined in the gel and a preferential length of ≥8 OAg repeated units, with a molecular weight between 43 and 100 KDa, consistent with previous descriptions in the literature in other P. gingivalis strains [41][42][43][44]. On the other hand, compared to P. gingivalis LPS, P. endodontalis LPS exhibited a faint lipid A-core region not substituted by OAg and a preferential length of 8 to 14 OAg substitutions repeating units.
It has previously been reported that P. endodontalis LPS electrophoretic profile in Tris-Glycine gels exhibits a preferential chain length greater than 8 OAg repeated units [45]. In this study, we report that in Tris-Tricine gels, the P. endodontalis LPS exhibited a preferential chain length greater than 8 OAg repeated units but smaller than the preferential length of E. coli LPS. Although the electrophoretic profiles of P. endodontalis LPS previously presented by Park's group are slightly different from ours, it is due to methodological differences, since Tris-Glycine gel provides higher resolution to bands with lower electrophoretic mobility while Tris-Tricine gel is suited for a higher resolution to bands with greater electrophoretic mobility [46].
During AP, macrophages exhibit a polarization switch towards M1, whereas disease exacerbation during symptomatic AP is associated with a reduced M2 differentiation profile, based on the reduced surface expression of CD163 along with higher IL-6 and IL-23 levels, which supports the role of the macrophages in the severity of apical lesions that can be triggered by endodontic bacterial antigens [32]. It has been previously reported that P. gingivalis LPS induces M1 cytokine secretion in bone marrow-derived macrophages once polarized to M1 or M2 [47], as well as M1 cytokine secretion in THP-1 differentiated to macrophages while being polarized with IFN-γ or IL-4 [48]. However, despite the above, the capacity of P. gingivalis and P. endodontalis LPS to induce macrophage polarization by analyzing their surface markers or cytokine secretion have not yet been described. In the present study, we phenotypically characterized M1 and M2 macrophage subpopulations based on the expression of their cell surface markers CD64 and CD80 for M1, as well as CD163 and CD206 for M2 [32,[49][50][51].
Our positive control, E. coli LPS, induced an M1 macrophage phenotype with lower CD64 and higher CD80 surface levels in THP-1 macrophages. In the same way, P. gingivalis LPS induced a reduction of CD64 surface levels, although not on the same level. CD64 is an IgG Fc-γ receptor constitutively expressed on monocytes and macrophages and participates in cellular cytotoxicity by triggering phagocytosis, pro-inflammatory cytokine secretion, ROS production, and antigen presentation to T cells. A decrease in CD64 could suggest a decline in the phagocytic capacity of those macrophages [52][53][54]. In contrast, CD80 is expressed at low basal levels and, like CD86, is a co-stimulatory molecule that, combined with MHC/antigen complex, induces activation, proliferation, and differentiation of T cells [55,56]. For that reason, it might be plausible that macrophages differentiated from THP-1 and stimulated by LPS for 24 h decrease their phagocytic capacity but increase their ability to activate, proliferate and differentiate T cells.
Specifically, P. gingivalis LPS from our clinical isolate (AC1Pg) induced a decrease in CD64 surface levels and an increase in TNF-α, IL-1β, IL-6, and IL-12 cytokine secretion; while P. gingivalis LPS from ATCC 33277 strain only induced increased secretion of TNFα and IL-1β in a statistically significant manner. The difference in magnitude response between E. coli LPS and P. gingivalis LPS is probably because P. gingivalis LPS is less immunologically active than E. coli LPS [57][58][59][60], which could be attributed to its lipid A heterogeneity. P. gingivalis LPS expresses two varieties of lipid A, a penta-acylated variety, capable of activating TLR4 in a similar way to E. coli LPS, and a tetra-acylated variety that antagonizes TLR4 by binding it but that does not trigger a response [36,57,61,62].
When comparing P. gingivalis LPS from different sources, the LPS from clinical isolate induced a more significant response than LPS from ATCC 33277 strain. It has been previously described that clinical isolates of P. gingivalis with low passages are more virulent than reference strains [63,64], presumably because P. gingivalis gradually loses the properties necessary to survive straight in vivo, in order to adapt to the in vitro culture conditions [65]. Therefore, since P. gingivalis from our clinical isolate was obtained from the root canal of a patient with AAP, we can attribute the different results of AC1Pg LPS compared to LPS of ATCC 33277 as being due to a higher virulence of P. gingivalis LPS from clinical isolates over the reference strains.
We also evaluated the immune response triggered by P. endodontalis LPS, and we observed that it did not induce changes in the secretion of the evaluated cytokines nor the surface levels of M1 or M2 markers. It would be plausible that P. endodontalis LPS possesses a lipid A heterogeneity similar to that described in P. gingivalis, but capable of inducing even lower immune responses, It cannot be ruled out that differences in the OAg region between P. endodontalis and P. gingivalis LPS may play a role in the low immune response induced in THP-1 macrophages [44,66,67].
Since AP is caused by a polymicrobial infection of the root canal system, we also studied the immune response of THP-1 macrophages co-stimulated with P. gingivalis and P. endodontalis LPS. Unexpectedly, LPS co-stimulation did not result in an increased immune response, suggesting that P. endodontalis LPS could attenuate the immune response induced by P. gingivalis LPS, by competing and blocking TLR4 and TLR2 [22,36,57], However, whether or not P. endodontalis LPS can attenuate the immune response in macrophages induced by P. gingivalis LPS or LPS from other Gram-negative bacteria needs further research. Altogether, these data suggest that macrophage polarization towards M1 during AP progression might be associated with the LPS burden of specific pathogens, such as P gingivalis.
Finally, we analyzed the TLR2 and TLR4 surface membrane levels on THP-1 cells differentiated into macrophages after stimulation 24 h with LPS. Surprisingly, both positive controls for ligands for TLR2 and TLR4 induced a reduction in TLR2 + and TLR4 + population in addition to reduced TLR2 and TLR4 surface levels. These results are difficult to interpret. Although increases in TLR2 and TLR4 levels have been described when stimulated with their respective ligands, other studies have reported a decrease. Specifically, an increase in surface expression of TLR2 and TLR4 was observed in THP-1 cells differentiated for 72 h with PMA (Phorbol 12-myristate 13-acetate) 17 nM and stimulated with P. gingivalis LPS or E. coli LPS 1 µg/mL for 24 h [68]. Similarly, THP-1 monocytes stimulated with P. gingivalis LPS or E. coli LPS 1 µg/mL for 24 h increased their TLR2 + or TLR4 + population respectively [69]. On the other hand, a decrease in surface expression of both TLR2 and TLR4, but not of TLR2 and TLR4 mRNA, was described after 24 h of stimulation with P. gingivalis LPS (1 µg/mL), PAM 3 CSK 4 (1 µg/mL) or E. coli LPS (0.1 µg/mL) in human gingival fibroblasts primary cultures [70]. This result could suggest that the protein expression of TLR2 and TLR4 may depend on the concentration of LPS, the cell model, and the conditions of differentiation to macrophages (concentration and time of exposition to PMA). Additionally, the reduction of CD64 was directly correlated with TLR2 and TLR4. In this regard, a directly proportional expression between CD64, TLR2, and TLR4 has been described in monocytes differentiated into macrophages from peripheral blood [71] and neutrophils from tuberculosis patients [72], but not in vitro macrophages differentiated to a pro-inflammatory profile. These antecedents suggest that there might be a functional relationship between the two, which would help explain the reduction in TLRs observed on the surface of LPS-stimulated macrophages in our experimental model.
When E. coli LPS bonds with TLR4, a hexameric complex is formed, (TLR4/MD-2/LPS) 2 , which activates downstream signaling that induces endocytosis of the complex and, subsequently, transcriptional factor activation [73][74][75][76][77]. It has been reported that bone marrowderived macrophages (BMDM) internalize TLR4 after 30 min of E. coli LPS stimulation [76]. In THP-1 differentiated to macrophages and stimulated with P. gingivalis or P. endodontalis LPS, or treated with decitabine, a demethylating agent, we did not observe changes in gene expression of TLR2 or TLR4. These results suggest that, in our model, DNA methylation would not play a predominant role in the reduction of TLR2 or TLR4 surface levels. On the other hand, it has been reported in THP-1 cells, that PAM 3 CSK 4 induced both an increase of TLR2 and an increase of soluble TLR2 (sTLR2). The last one would be mediated by metalloproteinase activity and consequent proteolysis of TLR2 on the membrane surface [78], a process that could be stimulated by PMA and LPS [78][79][80]. These data suggest that the reduction in surface levels of both TLR2 and TLR4 by E. coli LPS and PAM 3 CSK 4 could be explained, at least in part, by non-transcriptional mechanisms, such as recycling, lower export to membrane surface, or proteolysis of these receptors.

LPS Purification
LPS was purified by TRIzol extraction protocol [37], which was modified to remove nucleic acids and protein contaminants, as follows: The obtained bacterial pellets were sonicated with an ultrasonic processor, then incubated with DNase (0.1 mg/mL) (Sigma-Aldrich, St. Louis, MO, USA) and RNase (0.2 mg/mL) (Thermo Fisher, Waltham, MA, USA) at 37 • C overnight [38], and, subsequently, with Proteinase K (1 mg/mL) at 60 • C for 1.5 h. Then, they were homogenized in 2 mL TRIzol with 400 µL chloroform and centrifuged at 3900× g for 10 min at 4 • C. The upper aqueous phase was recovered and lyophilized overnight. The product was dissolved in 0.375 M MgCl 2 in 95% ethanol, centrifuged at 4700× g for 5 min and the precipitate recovered. LPS purity was verified with a Bradford assay (protein detection limit 0.1 mg/mL) and GelRed stained agarose gel (detection limit 0.1 ng). Bacterial lysates were obtained from cell cultures grown to the exponential phase and adjusted to OD 600 = 2.0 [81,82]. Purified LPS and bacterial lysates were visualized on silver-stained Tris-Tricine polyacrylamide gels (SDS-PAGE) [46,83] and compared with molecular weight standard #P7706S (BioLabs, Ipswich, MA, USA).

Flow Cytometry
Cells were placed in a V bottom 96-well plate in PBS for complete staining. Macrophages were stained with Fixable Viability Dye eFluor 780 (Invitrogen, Thermo Fisher, Waltham, MA, USA) 1:1000 in PBS for 30 min at 4 • C in the dark to discard the dead cells. Later, the cells were incubated with a mixture of all fluorochrome-conjugated antibodies at 1:25 each in PBS 2% FBS, for 30 min at 4 • C in the dark. The cells were single-stained under the same conditions as controls. After washing with PBS, the cells were fixed with 2% paraformaldehyde (Invitrogen, Thermo Fisher, Waltham, MA, USA) and analyzed by flow cytometry (LSR Fortessa X-20 ® ; Becton Dickinson Immunocytometry Systems, Franklin Lakes, NJ, USA). Flow cytometric data analysis was performed with Flow Jo v10 software (TreeStar, Ashland, OR, USA). The following mouse monoclonal anti-human antibodies conjugated with their respective fluorochromes were used: CD64 conjugated with Brilliant Violet 510 (BV510) and CD80 conjugated with fluorescein isothiocyanate (FITC) were used to identify M1 macrophage subpopulations; CD163 conjugated with PE/Cy7 and CD206 conjugated with Brilliant Violet 421 (BV421) were used to identify M2 macrophages subpopulations; TLR2 conjugated with APC and TLR4 conjugated with PE were also used (Biolegend, San Diego, CA, USA). Fixable Viability Dye (FVD) eFluor 780 (Invitrogen, Thermo Fisher, Waltham, MA, USA).
THP-1 macrophages were first gated according to their forward-and side-scatter profiles. Negative FVD cells were gated to discard dead cells. For the determination of the frequencies of M1 and M2 macrophage subpopulations, cells were gated over CD64 + CD80 + or CD163 + CD206 + . The threshold in each case was determined by the comparison with unstained samples. The M1/M2 ratio was calculated as the frequency of M1 double-positive macrophages divided by the frequency of the M2 double-positive macrophages in each sample. The intensity of expression of each specific cell surface marker was expressed as MFI.

RNA Expression
TLR2 and TLR4 relative mRNA levels were determined by qRT-PCR, and cDNA was amplified using their appropriate 5 -3 forward primer and 5 -3 reverse primer sets, respectively, as follows: CTCAACACGGGAAACCTCAC and CGTCCACCACTAAGAACG for 18S rRNA, CTCTCGGTGTCGGAATGTC and AGGATCAGCAGGAACAGAGC for TLR2 and CCCTCCCCTGTACCCTTC and TCCCTGCCTTGAATACCTTC for TLR4. All primer sets and reagents (KAPA™SYBR ® Fast; KAPA Biosystems, Woburn, MA, USA) were used in qPCR equipment (StepOnePlus ® ; Applied Biosystems, Singapore) as follows: 95 • C for 3 min and 40 cycles of 95 • C for 30 s and 60 • C for 30 s. To detect nonspecific product formation and false-positive amplification, a melt curve of 95 • C for 15 s, 60 • C for 1 min, and 95 • C for 15 s was performed. Relative quantification of mRNA was analyzed using the 2 -∆∆CT method. The normalization of gene expression was done against the 18S rRNA.

Data Analysis
The Shapiro-Wilk test was used to determine the distribution of the continuous data. Inferential analyses were performed using the Kruskal-Wallis test for non-normally distributed data. The statistical analysis was performed using STATA 12 ® (StataCorpLP, College Station, TX, USA). The figures were performed in Flow Jo v10 software (TreeStar, Ashland, OR, USA) and Graph Prism 5 (GraphPad Software, Inc., San Diego, CA, USA).

Conclusions
In summary, P. gingivalis LPS induces a macrophage polarization profile predominantly towards M1, similarly, although less evidently, than that generated by E. coli LPS, based on their cell surface markers and secreted cytokines. On the other hand, P. endodontalis LPS did not cause a clear immune response in THP-1 macrophages. P. gingivalis and P. endodontalis LPS stimulation did not induce TLR2 or TLR4 levels changes in THP-1 macrophages. These data suggest that macrophage polarization towards M1 during AP progression might be associated with the LPS burden of specific pathogens, such as P gingivalis.