Discussion
In this study, we have demonstrated that the stimulation of AGE,
hyperglycemia, and LPS from oral bacterial P. gingivalis can
synergistically dysregulate the DC phenotype, immunometabolism, and
effector functions. Hyperglycemia and infection-activated immune cells
together induce inflammation, which is the key to connect periodontitis
and DM, two chronic inflammatory diseases with a bidirectional
relationship to each other. The Warburg effect-like transition from
mitochondrial oxidative phosphorylation to aerobic glycolysis as the
main pathway for ATP generation is one critical step for DC activation
upon interaction between antigen and TLR (26). Interaction between LPS
and surface TLR on DCs could activate the glycolysis transition via
Akt-induced signaling, initiated by TLR-associated TBK1 and
IKK\(\varepsilon\) proteins, thus the Akt downstream protein mTORC1 is
also determined to be involve in the glucose-sensitive transduction
circuit for glycolysis metabolism transition (27,28). PFKFB3, an major
driver of glycolysis pathway, is identified as a direct target mTOR
under LPS stimulation (29). Our data showed a significant increase in
gene expression of enzymes such as HK2, LDHA, and GLUT that catalyzes
different steps of glycolysis. In the presence of LPS, levels of
L-lactate, which is the final product of glycolysis, increased in DC.
The induction of glycolytic activity by LPS has been well reported in
other immune cells such as macrophages, monocytes, and neutrophils
(30–32). LPS can attenuate mitochondrial OXPHOS activity. In
macrophages LPS suppresses isocitrate dehydrogenase and succinate
dehydrogenase in the citric acid cycle, and blocks the mitochondrial
respiration while promoting glycolysis (33,34). Excessive citrate and
succinate, both upregulate inflammatory molecules such as NO and
IL-1\(\beta\) (33). On the other hand, cells under hyperglycemia shunt
glycolytic intermediates into pentose phosphate pathway as a protective
mechanism, which limits the glycolytic activity (35,36). As macrophages
and DC are closely inter-related in pro- and anti-inflammatory
phenotypes, similar pathways could be involved in DC. Moreover, AGE can
also reduce glycolysis and mitochondrial respiration through
HIF-1\(\alpha\)/PDK4 pathway, thereby reducing ECAR and OCR (37,38).
This may explain the observed inhibition of glycolysis pathway in the
LPS and AGE-stimulated group under hyperglycemia in the cell metabolic
efflux assay. An inhibitory effect of hyperglycemia and AGE on
glycolysis could reduce the stimulating effect from LPS. Under diabetic
microenvironment, glycolysis is not only disrupted by hyperglycemia,
AGE, and LPS, rather other DM-associated metabolites such as ROS, free
fatty acid, and cholesterol could also impact cell immunometabolism
transitions by regulating different signaling pathways (31).
The relative expression of DC surface markers can reflect the cell
activation status. Upregulation of protein CD80, CD86, and MHC-II
complex is a key evidence of DC activation. Here we have shown that
depending on glucose concentration these three co-stimulatory molecules
can polarize the cells toward a hyperactivated status in both
inactivated and activated DC. Similar results were reported in DC
obtained from various lineages under in vitro and in vivostimulations (27,39,40). Thomas et al demonstrated that brief
exposure of hyperglycemic culture could significantly elevate
co-stimulatory surface markers in all immature, mature, and tolerogenic
DC (39). BMDC retrieved from streptozotocin induced-diabetic mice showed
comparable phenotypes as well (27,39). Such increase in CD80, CD86,
CD1\(\alpha\), and MHC-II was also observed when comparing the
hyperglycemia treated-human monocyte derived DC to the normoglycemia
cell culture (40). Glucose and LPS activated-Akt pathway that initiate
glycolysis transition may also upregulate the expression of these DC
surface markers. The transcription factor NF-\(\kappa\)B directly
regulates the receptor genes that are involved in the cell immunogenic
functions (41,42). Note that the HIF-1\(\alpha\) activated by Akt
pathway during DC metabolic reprogramming is also a transcription factor
that regulate CD80 and CD86 expression (43,44).
The metabolic reprogramming in innate immune cells is a TLR induced
event paired with phagocytosis, which is a critical for rapid infection
clearance by APCs. Our data demonstrated that hyperglycemia
significantly impairs the phagocytosis by immature DC in a
glucose-dose-dependent manner. Similar results have been reported in DC
differentiated from human PBMC obtained from T2DM patients, where cells
were unable to take up the FITC-dextran fluorescein, or this uptake was
significantly reduced in hyperglycemic media (45). Similarly, in other
phagocytes such as macrophages and monocytes the phagocytotic abilities
were reduced in both in vitro and in vivo hyperglycemic
environment. High glucose locks the cell in their pro-inflammatory
phenotype that is characterized by dysregulated activation markers and
upregulated pro-inflammatory cytokines (46–48). Similar results were
seen in the present study where hyperglycemia induced a pro-inflammatory
DC with compromised phagocytic effect. Reduced phagocytosis in these
innate immune cells collectively under hyperglycemia can completely
impair the infection clearance; thus leading to the chronic
immune-inflammatory responses seen in periodontitis and DM patients (6).
Such lowered phagocytotic capability could also result from
hyperglycemia induced altered cell surface receptors. Under
hyperglycemia and hypoxia conditions, in the expression of genes
encoding CD36 and SCARB-1 in human monocyte-differentiated macrophages
is reduced. CD36 and SCARB-1 are proteins responsible for phagocytosis
(47). Expression of FC\(\gamma\) receptor CD32 and CD64 that mediate
phagocytosis with antigen recognition receptor CR3 in monocytes were
also reported to be reduced with increasing HbA1c level
and body mass index in DM patients (48). Interestingly, no significant
differences were observed in phagocytosis protein CD36 on DC from DM
patient but MFGE8, another phagocytosis protein was significantly
reduced (49). MFGE8 is frequently associated with impaired phagocytosis
of various phagocytes generally and under DM conditions (50–52).
However, there is limited knowledge to associate the reduced MFGE8
levels with the aerobic glycolysis transition upon DC activation. Either
way, correcting the expression levels of this protein can result in
profound wound healing potential by promoting DC and macrophages into
immune suppressive phenotypes, which favors mitochondrial respiration
rather than glycolysis (53,54). The antigen presentation ability beyond
phagocytosis mechanism could also be impacted by hyperglycemia. High
glucose evidently delays antigen processing and impairs antigen
presentation in human monocyte-derived macrophages via repressed Rab and
Cathepsin, leading to reduced T cell activation (55). Defect in antigen
presentation has been reported in splenic macrophages from NOD mice
(56).
Exaggerated cytokine secretion is a characterized functional change upon
DC activation. Significant levels of pro-inflammatory cytokine,
TNF-\(\alpha\) and IL-1\(\beta\) from the activated DC indicates that
hyperglycemia, AGE, and LPS could together stress DC to exert a more
immunogenic state. Similar cytokine profiles were reported in other
studies from human and mice DC as each of the three stressors are known
to be pro-inflammatory (40,57,58). Besides their role in interconnecting
the chronic inflammatory states in periodontitis and DM, such increase
in these pro-cytokines may help in developing other autoimmunity in
innate immune cells as well. Recent studies have shown that
TNF-\(\alpha\) could stimulate the synovial fibroblast cells to increase
the glycolysis related-proton efflux and promote rheumatoid arthritis,
which is another risk factors associated with both periodontitis and DM
(59). Similarly, in autoimmune disease Sjögren syndrome as well the
inflammatory response between TNF-\(\alpha\) and DC via TLR-7 and TLR-9
has been confirmed (60). IL-1\(\beta\) can induce autocrine responses
during bacterial infection and reprogram the immunometabolism through
glycolysis (61). If this immunometabolism transition to aerobic
glycolysis can play a role in the trained immunity of DC under
hyperglycemic stress remains unclear, yet it is an important question
worth investigating.
DC have a unique capacity to induce activation and proliferation of
antigen specific T cell responses as its APC function in adaptive
immunity.
The
presence of co-stimulatory molecules CD80, CD86, and MHC-II complex play
an important role in T-cell activation; wherein, signals from DC
synergistically interact with both the T cell receptor and CD28 on T
cells for complete activation (62). On the other hand, lack of any one
of the co-stimulatory signals can result in immune tolerance or
ignorance of activation to prevent aberrant response of T cells.
Increase in the expression of DC co-stimulatory markers should lead to
hyperactivation of the T cell lymphocytes (63); however, contradictory
results were observed in our study. Although DC stimulated with
hyperglycemia, AGE, and LPS had high expression of co-stimulatory
markers , which increases with increase in glucose levels. T cells
cocultured with DC in higher glucose showed attenuated proliferation and
Th-1 type cytokine secretions, indicating a lowered T cell priming
capability. This was expected as the DC cultured in high glucose
concentration were more activated and matured while the phagocytosis
ability of immature DC is only optimal (64). Samuel et al had
also indicated that Th1 is only secondary predominant while Th17 is the
predominant pathway in DM mice (65). Such impairment in the DC function
could also be due to compromised phagocytosis and antigen presentation
capability in hyperglycemia. Antigen presentation on DC is required for
the effective physical interaction and T cell activation in vivo ,
whereas unpulsed DC fail to induce lymphocyte responses as shown by
Ingulli et al (66). Also, a minimum threshold antigen dose is
required for a stable DC-T cell priming (67). With attenuated
phagocytosis ability and therefore the lowered antigen presentation on
DC in hyperglycemia, the T cell activation is weak, and this may lead to
a prolonged inflammation in DM and periodontitis. However, the antigen
presentation using LPS model is a non-specific or bystander response
from T cells, as presence of LPS could induce inflammatory responses
from a wide array of cells. Extension of the adaptive immune responses
using OT-II mice may could help in further understand such compromised
antigen presentation from DC by pulsing DC and T cells co-culture with
specific antigens like ovalbumin (67).
Periodontitis is a chronic disease caused by a multispecies bacterial
biofilm, and DM is a systemic autoimmune disease that involves numerous
inflammatory molecules and immune cells. Therefore, the simple use of
bacterial LPS and hyperglycemia may not comprehensively mimic the
complicated disease microenvironment. In vivo studies utilizing
diabetic mice with induced periodontitis could provide better
understanding of the diseased state and immune responses. While we
performed the transcription analysis of DC surface markers in this
study, it is possible that the hyperglycemia could affect the cell
translation mechanism. Furthermore, the intracellular localization of
the phagocytosed fluorescent microspheres would be an enticing next step
to understand the DC phagocytosis mechanism.
In conclusion, this study highlights that hyperglycemia with AGE and LPS
can polarize DC toward a more pro-inflammatory phenotype and cytokine
profile by upregulating glycolysis, but its phagocytosis and T cell
priming is compromised. Together, this DC function dysregulation may
explain the conflicting adaptive immune cell profiles observed in DM and
periodontitis.