Mitochondrial Superoxide Signaling Contributes to Norepinephrine-Mediated T-Lymphocyte Cytokine Profiles

Norepinephrine (NE) produces multifaceted regulatory patterns in T-lymphocytes. Recently, we have shown that NE utilizes redox signaling as evidenced by increased superoxide (O2●-) causally linked to the observed changes in these cells; however, the source of this reactive oxygen species (ROS) remains elusive. Herein, we hypothesized that the source of increased O2●- in NE-stimulated T-lymphocytes is due to disruption of mitochondrial bioenergetics. To address this hypothesis, we utilized purified mouse splenic CD4+ and CD8+ T-lymphocytes stimulated with NE and assessed O2●- levels, mitochondrial metabolism, cellular proliferation, and cytokine profiles. We demonstrate that the increase in O2●- levels in response to NE is time-dependent and occurs at later points of T-lymphocyte activation. Moreover, the source of O2●- was indeed the mitochondria as evidenced by enhanced MitoSOX Red oxidation as well as abrogation of this signal by the addition of the mitochondrial-targeted O2●--scavenging antioxidant MitoTempol. NE-stimulated T-lymphocytes also demonstrated decreased mitochondrial respiratory capacity, which suggests disruption of mitochondrial metabolism and the potential source of increased mitochondrial O2●-. The effects of NE in regards to redox signaling appear to be adrenergic receptor-dependent as specific receptor antagonists could reverse the increase in O2●-; however, differential receptors regulating these processes were observed in CD4+ versus CD8+ T-lymphocytes. Finally, mitochondrial O2●- was shown to be mechanistic to the NE-mediated T-lymphocyte phenotype as supplementation of MitoTempol could reverse specific changes in cytokine expression observed with NE treatment. Overall, these studies indicate that mitochondrial metabolism and O2●--mediated redox signaling play a regulatory role in the T-lymphocyte response to NE.


Introduction
Enhanced activation of the sympathetic nervous system is associated with numerous pathological conditions ranging from hypertension, heart failure, diabetes, and even psychological stress [1][2][3][4]. Sympathoexcitation leads to increased norepinephrine (NE) outflow to peripheral organs including the predominantly sympathetic-innervated lymphoid organs including the bone marrow, lymph nodes, and spleen [5]. Resident immune cells in these lymphoid organs have been shown to possess adrenergic receptors [6,7], and over the last four decades it has become well-accepted that autonomic regulation of the immune system is a tangible phenomenon [8,9]. However, catecholamines appear to elicit a complex pattern of regulation on immune cells dependent upon numerous variables including cell type, activation status, polarization and differentiation, organ of residence, and many others [10,11].
One immune cell type that has been extensively studied for its effects with NE is T-lymphocytes. Early in vitro work with T-lymphocytes demonstrated that NE slows the proliferation of these cells and decreases the amount of pro-inflammatory cytokine production through an inhibition of interleukin 2 (IL-2), and this observation has been validated by numerous laboratories utilizing various populations of T-lymphocytes [11][12][13][14][15]. Moreover, NE appears to produce this inhibitory effect primarily through a β2 adrenergic receptor-mediated mechanism [14]. In contrast, several investigations have shown that NE may enhance the pro-inflammatory state of T-lymphocytes particularly in regards to interferon gamma (IFNγ) production or in the ability to fight infection [16][17][18]. Additionally, other studies have identified NE-mediated effects on T-lymphocytes that are initiated via adrenergic receptors other than just the β2 isoform [15,[19][20][21]. Taken together, the complexity and disparity of observations in regards to NE-mediated effects on T-lymphocytes suggests the potential for multidimensional regulatory mechanisms that are not yet fully understood.
In 2013, Fadel and colleagues observed that human peripheral blood mononuclear cells produced increased reactive oxygen species (ROS), particularly superoxide (O 2 •-), in response to NE, and further suggested this to be an α2-adrenergic receptor-mediated effect [19]. We have recently confirmed and extended these findings specifically in T-lymphocytes both in vivo and in vitro in a mouse model of sympathoexcitation [22]. Our results additionally suggested this increase in O 2 •was causal to the NE-driven effects in the T-lymphocytes as O 2 •--scavenging antioxidant supplementation was able to partially inhibit the NE-mediated T-lymphocyte phenotype [22]. To our knowledge, these findings were the first to report the potential for redox signaling in the regulation of NE-mediated effects in T-lymphocytes.
In the work presented herein, we aimed to expand upon our previous observation and identify the source of O 2 •in NE-stimulated T-lymphocytes. Due to our previous observations that the O 2 •produced in response to NE appeared to be time dependent, we hypothesized that NE may be altering mitochondrial metabolism and in turn affecting mitochondrial-derived O 2 •specifically. We and others have recently reported that metabolism and redox signaling play an integral role in T-lymphocyte activation, polarization, and function [23][24][25][26][27][28], and here we demonstrate that NE may affect these central processes adding to the complexity of catecholaminergic regulation of T-lymphocytes.

Mitochondrial bioenergetics analysis
To measure T-lymphocyte mitochondrial bioenergetics, a Seahorse Bioscience XFp extracellular flux analyzer was used. This device utilizes specialized microplates to create a closed chamber able measure real-time oxygen consumption by mitochondria in live cells exposed to various stimuli through multiple designed injection ports. Optimal seeding density of T-lymphocytes was established at 200,000 cells per well and was utilized for all experiments. Additionally, mitochondrial agents (Seahorse Bioscience Cell Mito Stress Test Kit #103015-100, Boston, MA) were pre-optimized at 1 μM oligomycin, 1 μM FCCP, and 10 μM rotenone/antimycin A to elicit maximal effects on mitochondrial respiration. Microplates were pretreated with 1 μg/cm 2 Cell-Tak (Corning #354240, Corning, NY) to allow T-lymphocyte adhesion to the chamber. For acute NE treatment, T-lymphocytes were freshly isolated, counted, and immediately seeded on microplates in minimal DMEM (Seahorse Bioscience XF Base Medium #102353, Boston, MA) supplemented with 11 mM D-glucose and 2 mM L-glutamine. NE was administered by way of Seahorse injection port and cells analyzed for 30 minutes prior to mitochondrial stress test. For treatment of NE in activated T-lymphocytes, T-lymphocytes were cultured for 96 hours as previously described in the presence of 1 μM NE. Cells were harvested, counted, and seeded on microplates in minimal DMEM supplemented with 11 mM D-glucose and 2 mM L-glutamine. Mitochondrial stress test was performed immediately without any additional NE administration. Cells from one mouse (vehicle and NE treated) were run on a single plate in technical replicates; biological replicate plates were run and data pooled for analysis.

Cytokine analysis
Extracellular secreted cytokine analysis was performed on cell media from T-lymphocytes cultured for 96 hours. Analysis was performed using the mouse Th1/Th2/Th17 cytometric bead array (BD Biosciences #560485, San Jose, CA) as per manufacturer's instructions. Briefly, media was combined with antibody-coated fluorescent beads for 7 specific cytokines used to profile various T-lymphocyte subtypes. Beads were analyzed on a LSRII flow cytometer for quantification of specific cytokines.

Statistics
Data are presented as mean ± standard error of the mean (SEM). For two group comparison, significance was assessed using the paired Student's t-test. For multiple group comparison, significance was assessed using 2-way ANOVA followed by Bonferroni post-hoc analysis. Differences were considered significant at p<0.05.

NE increases O 2 •in both CD4+ and CD8+ T-lymphocytes
In our previous report, we demonstrated that NE elevates steady-state O 2 •levels in unfractio-

Mitochondria are the source of O 2 •in NE-stimulated T-lymphocytes
Under certain conditions, NE is known to generate free radical species independent of a receptor-mediated or cell-dependent mechanism [38][39][40]. To ensure the increase in DHE oxidation observed with NE-stimulated T-lymphocytes was not due to direct interaction of the NE with DHE, we tested the ability of NE to oxidize DHE in a cell free environment. We found no evidence that NE was directly increasing DHE oxidation (S3 Fig

Mitochondrial bioenergetics are altered by NE in T-lymphocytes
Within a mitochondrion, there are many potential sources of O 2 •generation, but the primary source is due to electron leak onto oxygen from the electron transport chain [43,44]. Due to this, we measured mitochondrial respiration in NE-stimulated T-lymphocytes utilizing a Seahorse Bioscience extracellular flux bioanalyzer. This technology allows for real-time measurements of mitochondrial metabolism via the measurement of oxygen consumption in response to specific mitochondrial respiratory chain inhibitors [45]. When examining the effects of an acute exposure (30 minutes) of NE to T-lymphocytes, no changes were observed in baseline oxygen consumption or with any mitochondrial inhibitor (Fig 3A), which correlated with the lack of any detectable ROS production at this time point as well (Figs 1B and 2B). In contrast, T-lymphocytes activated in the presence of NE for 96 hours demonstrated a decreased respiratory capacity response to the uncoupling agent carbonilcyanide p-triflouromethoxyphenylhydrazone (FCCP; Fig 3B). Increased ROS production from the mitochondria and decreases in respiratory capacity are often associated, and are additionally correlated with a numerous pathologies suggesting a possible mechanism behind altered T-lymphocyte function in response to NE [46][47][48][49]. Moreover, NE does not appear to affect T-lymphocyte mitochondrial ATP levels as the addition of oligomycin demonstrated similar inhibition in both NE and control treated cells (Fig 3B). Overall, these data further support the mitochondria as the source of O 2 •in NE-treated T-lymphocytes and suggest bioenergetic dysfunction may play a role in the observed phenotype.

NE-mediated redox signaling appears dependent upon adrenergic receptors
We next attempted to understand the mechanism by which NE was causing increased mitochondrial O 2 •in T-lymphocytes. Aforementioned, NE is able to auto-oxidize to generate free radicals independent of cell or enzyme dependent mechanisms [38][39][40]. While we demonstrated NE was not directly oxidizing DHE, we postulated that if NE was able to enter the cells it could possibly directly generate O 2 •intracellularly. However, NE is not cell permeable and requires specific transporters to import the catecholamine across the cell membrane. T-lymphocytes have been shown to express various catecholamine transporters as well as generate their own NE which is transported across the cell membrane and utilized in an autocrine fashion[21, 32, [50][51][52][53]. Due to this, we utilized the NE transporter (NET) inhibitor atomoxetine to see if the intracellular import of NE was leading to the increase in O 2 •in T-lymphocytes. In contrast to our hypothesis, the addition of atomoxetine alone increased O 2 •within T-lymphocytes, and moreover, produced an additive effect with the supplementation of NE (Fig 4). Additionally, atomoxetine alone demonstrated a dose-dependent decrease in cell number and in combination with NE enhanced the proliferative defect (S4 Fig). Taken together, these data suggest NE transport is not the primary mechanism by which NE mediates intracellular O 2

•-
production. Furthermore, the inhibition of NE uptake into the cell appears to enhance the NE- mediated alterations suggesting T-lymphocytes may normally utilize this mechanism to metabolize the catecholamine and limit its effects. + T-lymphocytes cultured for 96 hours, the specific antagonists demonstrated variable responses in the respective cell types (Fig 5). CD4+ cells showed a significant reduction in NEdriven O 2 •only when treated with the α2 antagonist (alone or in combination with α1 antagonism). Interestingly, individual β antagonism had no effect, but both β1 and β2 blockade suggest that NE mediates intracellular O 2 •production in T-lymphocytes most likely through the binding and activation of specific adrenergic receptors dependent upon cell type, and moreover, these redox signaling events appear to be discordant to the mechanism driving decreases in T-lymphocyte proliferation.
NE alters T-lymphocyte cytokine profiles in part due to mitochondrial O 2

•-
We and others have reported that NE is able to alter cytokine production in T-lymphocytes, however, the specific responses appear to be dependent upon several factors including experimental setup and activation status of the cells [15,17,22,33]. We hypothesized that the changes in cytokine production may be partially mediated by the increase in mitochondrial O 2 •produced by NE in T-lymphocytes. To address this, we performed cytokine arrays on media from CD4+ or CD8+ T-lymphocytes cultured with NE in the presence or absence of O 2 •scavenging antioxidants. In CD4+ cells, NE significantly reduced IL-2, IFNγ, tumor necrosis factor α (TNFα), and IL-10 levels, while increasing IL-17A (Fig 6). Interestingly, the addition of Mito-Tempol was able to significantly restore IL-2, IFNγ, and IL-17A levels in these cells (Fig 6). In CD8+ cells, NE significantly reduced the same four cytokines as CD4+ cells (i.e. IL-2, IFNγ, TNFα, and IL-10), but also increased both IL-17A and IL-6 levels (Fig 7). In these cells, Mito-Tempol was able to significantly reestablish IL-6, IL-17A, and IL-10 levels (Fig 7). Interestingly, growth was not restored in either CD4+ or CD8+ T-lymphocytes treated with antioxidants (Figs 6 and 7), which further supports differential regulation between growth and cytokine production in regards to NE-mediated redox signaling. Taken together, these data further support the hypothesis of differential redox regulation in CD4+ and CD8+ cells, but also suggest that increased mitochondrial O 2 •signaling is only partially contributing to the NE-mediated regulation of T-lymphocytes as not all cytokines or growth could be rescued with the attenuation of the NE-mediated mitochondrial redox signaling.

Discussion and Conclusion
Investigations into the crosstalk between the nervous and immune systems have been ongoing for several decades. Previously, we observed a new paradigm of redox control in the regulation of T-lymphocytes exposed to NE [22]. In the work presented herein, we have expanded this original observation to identify that mitochondria (and possibly mitochondrial metabolic dysfunction) are the source of the increased O 2 •in response to NE.
In recent years, it has become appreciated that ROS are necessary for proper T-lymphocyte activation and function. Initial studies identified that both O 2 •and hydrogen peroxide (H 2 O 2 ) were produced upon crosslinking of the T-lymphocyte receptor, which stimulated the ERK signaling pathway modulating T-lymphocyte activation [54]. Soon after this work, T-lymphocytes were identified to possess a distinctly unique NAPDH oxidase that was the source of this ROS upon T-lymphocyte receptor stimulation [55]. Moreover, we and others have shown that mitochondrial derived ROS are also critical in the development and function of T-lymphocytes[23, 25]. However, while ROS have demonstrated a critical functional role in T-lymphocytes, the actual redox signaling mechanisms involved in these processes remain elusive. In the work presented here, we confirm the finding that mitochondrial O 2 •is increased with T-lymphocyte activation over time and report for the first time that NE-stimulation is able to potentiate this specific ROS production which can alter T-lymphocyte cytokine production. How NE facilitates its effects on the mitochondria remains unclear, but we hypothesize the mediator may be cyclic AMP (cAMP) or its derivatives. Early work examining NE effects on T-lymphocytes demonstrated the observed phenotype is highly attributed to a significant induction of cAMP via the classic G protein-coupled receptor pathway [5,9]. Recently, it has been observed that cAMP derivatives may affect mitochondrial ROS and metabolism via altering the mitochondrial permeability transition pore (MPTP) [56,57]. This pathway may explain how NE is able to exert its effects on the mitochondria in T-lymphocytes. Moreover, this mechanism could explain why we observed increases in O 2 •over time and not acutely, as a buildup of cAMP may be required to significantly alter MPTP function, mitochondrial polarization, and ROS production. Overall, this hypothesis warrants further investigation, and is a focus of current work in our laboratory. In addition to ROS, cellular metabolism has become accepted as a primary regulator of Tlymphocyte activation and differentiation. Work from Pearce and colleagues has identified that T-lymphocytes shift their metabolic profiles dependent upon activation and differentiation status [26][27][28]58]. For example, naïve T-lymphocytes reside in a relatively quiescent state with minimal metabolic needs, however, upon activation to effector T-lymphocytes these cells utilize the Warburg Effect and heavily rely upon glycolysis over mitochondrial oxidative phosphorylation to fulfill their metabolic demand to proliferate and function [28]. In contrast, memory T-lymphocytes significantly enhance their mitochondrial biomass and rely exclusively on oxidative phosphorylation for their function, which is believed to provide a significant advantage to rapidly respond to a secondary immune insult [59][60][61][62][63]. Mitochondrial respiratory capacity is linked to T-lymphocyte memory cell development [61], and furthermore, others have shown that NE does in fact affect memory T-lymphocyte function [34]. Our data presented herein indicates that NE significantly decreases mitochondrial respiratory capacity, and this mechanism may explain the NE-mediated changes in memory T-lymphocyte function previously reported. However, it remains unclear at this time if the NE-mediated mitochondrial dysfunction leads to the increase in mitochondrial O 2 •or vice versa.
Our data utilizing mitochondrial-targeted antioxidant supplementation demonstrate a partial rescue of the NE-mediated changes in T-lymphocytes, however, the attenuation of O 2

•-
was not able to fully restore growth or all cytokine levels back to normal. These data suggest that these processes are controlled, at least in part, by NE signaling through non-redox regulated mechanisms (i.e. cAMP and PKA [9,64], changes in cyclin expression[22], etc.) or that alterations in mitochondrial metabolism may be upstream of the increased O 2 •production.
Thus, attenuating the ROS via O 2 •scavenging restores the redox-regulated signaling processes but does not eliminate the continued metabolic defect, which may be causal to other aspects of NE-driven changes in T-lymphocytes. For example, metabolite alterations can directly affect the production of specific T-lymphocyte cytokines through regulation of their mRNA via post-transcriptional modifications [65]. Understanding that NE appears to have multifaceted regulation on various cellular processes including metabolism, it is intuitive that the addition of an antioxidant would not be sufficient to reverse all NE-mediated processes in the cells, but only those that are specifically controlled via redox mechanisms (Fig 8).
As part of the phenotype observed in this work, we identified significant changes to cytokine profiles in T-lymphocytes treated with NE. First, we identified that IL-2, IFNγ, and TNFα were all significantly down-regulated in both CD4+ and CD8+ T-lymphocytes treated with NE; a finding that we and others have reported previously[9, 15,22,64]. Additionally, we also observed changes in IL-17A, IL-10, and IL-6. These cytokines are not always reportedly altered in T-lymphocytes in response to NE, but we attribute these conflicting results to variances in experimental setup (i.e. mouse strain, activation stimulus, time course, NE dosage and regimen, etc.) as previously discussed [10,11]. Interestingly, these three cytokines were all demonstrated to be redox regulated, as antioxidant supplementation could reverse their aberrant expression. Moreover, their pattern of expression (i.e. increased IL-17A and IL-6, decreased IL-10) suggests a pro-inflammatory profile with increased abundance of T H 17 polarization. T H 17 cells primarily produce the pro-inflammatory cytokine IL-17A, and while specific cytokines are known to augment the polarization of T-lymphocytes to this subtype, the intracellular mechanisms controlling this differentiation have yet to be fully elucidated. To date, the major transcription factors known to contribute to T H 17 differentiation are the retinoic acid receptorrelated orphan receptor gamma (RORγ) and Janus kinase/signal transducer and activator of transcription (JAK/STAT) pathways [66]. It may be possible that NE signaling via non-canonical (i.e. increased mitochondrial O 2 •or altered cellular metabolites) pathways activates these pro-T H 17 signaling pathways, thus leading to the redox-regulated increase in IL-17 in T-lymphocytes treated with NE, as reported herein. Interestingly, T H 17 cells are often associated with autoimmune diseases, and recent reports have demonstrated diseases of increased sympathetic drive appear to be associated with higher abundance of autoimmune and T H 17 T-lymphocytes [67][68][69]. For example, patients suffering from post-traumatic stress disorder (PTSD) possess increased incidence rates of the autoimmune disease rheumatoid arthritis [70,71]. Because PTSD is known to elevate sympathetic drive [72], and that rheumatoid arthritis is also known to be affected by elevated levels of NE [73,74], it suggests autonomic dysfunction as a possible link between psychological and autoimmune diseases. To date, mitochondrial-targeted antioxidant supplementation has not been examined as a potential treatment in these diseases, and as such our data may support further investigation into the potential of this therapeutic modality. This study does possess some limitations that will need to be addressed in future work. First, we exclusively utilized CD3/CD28 stimulation as the mechanism to activate T-lymphocytes. While this is an established immunologic procedure, this methodology precludes examination as to whether the redox effects we observed are also produced during antigen-specific activation of T-lymphocytes in vitro as well as in vivo. These studies along with culturing the T-lymphocytes in polarizing conditions (i.e. T H 1, T H 2, T reg ) may provide additional information as to how specific subtypes of T-lymphocytes utilize redox regulation in response to NE. Furthermore, our studies were carried out ex vivo, in ambient (approximately 21%) oxygen conditions, as well as in the presence of standard high glucose media. Follow-up studies will use in vivo models of sympathoexcitation as well as utilize gas controlled work stations to mimic physiological concentrations of oxygen as well as varying metabolic substrates to gain a deeper understanding of the metabolic and redox effects of NE on T-lymphocytes under different physiological conditions. Lastly, examination of the adrenergic receptors responsible for initiating NE-mediated O 2 •production depicted a complex pattern of regulation. In CD4 + cells, the α2 receptor appeared to be primarily responsible for the NE-driven increase in O 2

•-
, and this result agrees with a previous report of O 2 •production in human immune cells [19]. However, in CD8+ cells the pattern of adrenergic receptor regulation was complex and appeared to involve all receptors analyzed. It is unclear at this time why and how these patterns are divergent, but may be due to differential receptor density, sensitivity, or even mitochondrial content and signaling [75] between CD4+ and CD8+ T-lymphocytes. Investigations into these differences and their responses to adrenergic stimulation and blockade are highly warranted as patients currently taking adrenergic inhibitors may be subject to altered immune regulation [76] due to disproportionate binding of physiological NE concentrations to the remaining available adrenergic receptors.
In conclusion, our results present the unique observation that NE modulates T-lymphocyte function via alterations in mitochondrial metabolism and redox status, primarily through the generation of mitochondrial O 2

•-
. The interplay of the nervous and immune systems is incredibly complex, and this study illuminates a single possible mechanism of regulation. It remains unclear how other neurotransmitters such as epinephrine, acetylcholine, neuropeptide Y, or substance P affect T-lymphocyte redox signaling. Moreover, the understanding of how these compounds affect the oxidative status of other components of the immune system (i.e. B-lymphocytes, macrophages, dendritic cells, etc.) remains unknown. The understanding of how these neural components dictate immune function may prove to be essential in the understanding of diseases with increased autonomic activity, and thus provide novel avenues for therapeutic intervention.