Targeting adrenergic receptors to mitigate invariant natural killer T cells-induced acute liver injury

Summary Invariant Natural Killer T (iNKT) cell activation by α-galactosylceramide (αGC) potentiates cytotoxic immune responses against tumors. However, αGC-induced liver injury is a limiting factor for iNKT-based immunotherapy. Although adrenergic receptor stimulation is an important immunosuppressive signal that curbs tissue damage induced by inflammation, its effect on the antitumor activity of invariant Natural Killer T (iNKT) cells remains unclear. We use mouse models and pharmacological tools to show that the stimulation of the sympathetic nervous system (SNS) inhibits αGC-induced liver injury without impairing iNKT cells’ antitumoral functions. Mechanistically, SNS stimulation prevents the collateral effect of TNF-α production by iNKT cells and neutrophil accumulation in hepatic parenchyma. Our results suggest that the modulation of the adrenergic signaling can be a complementary approach to αGC-based immunotherapy to mitigate iNKT-induced liver injury without compromising its antitumoral activity.


INTRODUCTION
Despite medical advancements in the last decades, cancer remains one of the most significant challenges in clinical practices, especially in cases of advanced or recurrent disease, where treatment options are limited.][3][4] iNKT cells, also known as NKT/NKT-I, are a CD1d-restricted population of non-conventional T lymphocytes with potent modulatory capacity and selectivity for glycolipids, properties shared by human, primate, and non-primate species. 5Indeed, murine experimental models were the first to describe the existence of an invariant T lymphocyte sharing distinct characteristics with T lymphocytes and NK cells, with strong antitumor ability. 1,6After this discovery, experimental models have been extensively used to improve the knowledge of iNKT cells' biology, its agonists, and, thereby, iNKT-based immunotherapy. 7,80][11] Because the efficacy of the antimetastatic effect of aGC treatment depends on the sequential activation of DCs, NK cells, and CD8 T lymphocytes triggered by IFN-g-producing iNKT cells, especially against CD1d-negative tumors, any external events compromising this cascade of immune reactions may impair aGC immunotherapy. 12,13 growing body of evidence supports the idea that the sympathetic nervous system (SNS), through the production of norepinephrine (NE) and primarily via b 2 -adrenergic receptor (b 2 AR) signaling, can act as an immunomodulatory neuronal system.NE impairs the cytotoxic activity of both NK cells and CD8 + T lymphocytes, favors the polarization of CD4 + T lymphocytes toward a Th2 profile, and modulates innate immunity to restrain tissue damage due to uncontrolled inflammatory responses.[14][15][16][17][18][19][20] Although naive Th0 CD4 + are bipotential and Th1 and Th2 responses are mutually exclusive, NE-induced Th2 polarization seems to be mechanistically related to the signaling via b 2 AR, whose expression is sustained under Th1 but not Th2 driving conditions.21,22 In contrast to conventional T lymphocytes, thymus-emigrated iNKT cells exhibit a memory/effector profile and express genes related to different adrenergic receptors, including the Adrb2, which encodes b 2 AR.23,24 Therefore, the expression of ARs in iNKT cells does not seem to be regulated in the same fashion as in conventional T lymphocytes, and their impact on iNKT cells' biology remains to be clarified.
Considering the potent immunomodulatory ability of iNKT cells and their importance as a target for immunotherapy, it is essential to determine the influence of adrenergic signaling on iNKT cell activity and antitumoral functions. 8or this purpose, we used mice genetically deficient for a 2A and a 2C AR (Adra2ac À/À ) as a genetic model of hyperactivation of the SNS. 25,26hese receptors promote a negative feedback mechanism controlling catecholamine secretion, and due to their absence, Adra2ac À/À mice ll OPEN ACCESS exhibit an NE hypersecretion phenotype. 25Using this model and pharmacologic hyperactivation of the SNS, we found that adrenergic stimulation prevents liver damage induced by iNKT cell activation without affecting their antitumoral activity.

Invariant Natural Killer T cells are refractory to adrenergic signaling
Although previous data showed that iNKT cells express mRNAs encoding a and b-adrenergic receptors, the direct effect of NE signaling on their biological activity remains unclear. 24Therefore, to determine the impact of adrenergic signaling on iNKT cell activity, we first incubated naive splenic cells from C57BL/6J WT mice with an anti-CD3 antibody or with PBS57, an aGC analog loaded in CD1d monomers in the presence or absence of NE (Figure 1).While the maximum iNKT cell response was achieved using the cognate antigen, adrenergic stimulation did not influence the production of IFN-g or TNF-a following stimulation with anti-CD3 or PBS-57 (Figure 1A).In contrast, conventional T lymphocytes exhibited a lower frequency of IFN-g and TNF-a-producing T cells under NE stimulation compared to control samples (Figure 1B).These results are in concordance with previous reports describing that NE signaling impairs the acquisition of a Th1 effector profile by naive CD3 + T lymphocytes upon anti-CD3 stimulation. 18Therefore, iNKT cells, but not conventional T lymphocytes, are refractory to the inhibitory effect of adrenergic signaling.
Next, we took advantage of the Adra2ac À/À mouse, a prototypical model of NE hypersecretion, to study in vivo the influence of adrenergic signaling on the iNKT cells' response to aGC. 26 aGC injection increases systemic levels of several immunomodulatory molecules, such as IL-6, IFN-g, TNF-a, IL-4, IL-10, IL-17, IL-12 and MCP-1, without a significant difference between WT and Adra2ac À/À animals (Figure 2A).Flow cytometric analysis indeed revealed that the production of cytokines by splenic and hepatic iNKT cells following aGC injection was similar in both mouse strains, with rapid production of IFN-g and TNF-a upon TCR engagement (Figures 2B and 2C).
Notably, no difference in the frequency of thymic or peripheral iNKT cells was observed between the C57BL/6J WT and Adra2ac À/À mouse strain (Figure S2).Therefore, these results support the hypothesis that adrenergic signaling does not influence the iNKT cell response to aGC and, consequently, the cascade of events that follows its stimulation.

Adrenergic signaling inhibits a-galactosylceramide-induced liver injury
The production of pro-inflammatory cytokines, especially TNF-a, in response to aGC has been associated with liver injury, characterized by increased systemic levels of hepatic enzymes (ALT and AST) and alterations in the hepatic parenchyma, such as the appearance of necrotic areas. 27Indeed, WT animals treated with aGC exhibited a significant increase in systemic levels of ALT and AST associated with signs of hepatic injury, including necrotic areas and robust inflammatory foci (Figures 3A-3C).In contrast, the Adra2ac À/À animals treated with aGC did not present liver dysfunction, exhibiting systemic levels of ALT and AST comparable to those of control, non-treated animals (Figure 3A).Moreover, histological analysis revealed the preservation of the hepatic tissue in the Adra2ac À/À animals despite a significant influx of inflammatory cells into the liver tissue (Figures 3B and 3C).Therefore, these results demonstrate that although adrenergic signaling did not inhibit the iNKT cells' activity and inflammatory responses triggered by aGC stimulation, it provides a hepatoprotective effect against iNKT-induced tissue damage.

Adrenergic signaling does not inhibit the control of melanoma growth by a-galactosylceramide immunotherapy
Previous studies reported that b 2-adrenergic receptor signaling undermines the immunosurveillance against tumor cells by impairing the activity of NK cells and CD8 + T lymphocytes. 16,28In concordance with these findings, Adra2ac À/À mice are more susceptible to B16F10 lung metastasis than their WT counterparts (*p < 0.05; Figure 4D).Therefore, considering that the effectiveness of aGC immunotherapy relies on a cascade of events triggered by the production of IFN-g by iNKT cells, it was essential to determine the impact of NE hypersecretion on the antitumor efficacy of aGC treatment. 12he aGC-treated WT group exhibited a substantial increase in systemic IFN-g levels, hepatic enzymes (ALT and AST), and a significant reduction in the number of lung nodules in comparison to the non-treated group (Figures 4A-4D), reinforcing the notion that in addition to controlling tumor growth, aGC immunotherapy also results in liver injury.The tumor-bearing Adra2ac À/À mice responded positively to aGC injection, exhibiting a significant increase in systemic IFN-g and control of tumor growth without any significant alteration in the systemic levels of hepatic enzymes, thus corroborating the hepatoprotective effect of adrenergic signaling (Figures 4A-4D).Therefore, while the sympathetic nervous system activation might impair natural immunosurveillance against tumors, the iNKT-mediated immune response to melanoma cells is refractory to its immunosuppressive effects.
Pharmacological stimulation of the sympathetic nervous system inhibits a-galactosylceramide-induced liver injury Considering the hepatoprotective effect resulting from the hyperactivity of the adrenergic fibers observed in Adra2ac À/À mice, we hypothesized that a punctual stimulation of the SNS should also be able to mitigate liver damage without affecting the aGC antimetastatic ability.To test this idea, WT animals received a single injection of yohimbine (Yoh), an a 2 -adrenergic antagonist that mimics the lack of control of NE secretion due to the absence of the a 2A/C AR, before aGC treatment. 29,30Figure 5 shows that administering Yoh before aGC treatment did not affect IFN-g production or the antitumor effect associated with iNKT cell activation, compared to control, non-treated mice (Figures 5A and 5B).Moreover, systemic levels of hepatic enzymes in the Yoh group were comparable to those detected in control mice, in contrast to the aGC group, which exhibited a significant increase in systemic levels of ALT and AST (Figures 5C and 5D).Besides aGC, other specific agonists have been developed to improve iNKT-based immunotherapy.Thus, we used the same approach to study the impact of adrenergic signaling on the antimetastatic effect of PBS44, an aGC analog with antitumor activity. 9Despite controlling lung metastasis, PBS44 treatment also increased the levels of ALT and AST, which were controlled by pre-treatment with Yoh (Figures 5E-5G).Therefore, these results support the notion that the pharmacological modulation of sympathetic fiber activity during iNKT cells stimulation improves immunotherapy outcomes.

Adrenergic signaling mitigates invariant natural killer T-induced liver injury through the a 1 receptor
As adrenergic signaling did not exert a direct suppressive effect on the immune responses triggered by iNKT cell activation, we hypothesized that it could act directly on hepatocyte survival.Analyzing datasets from two different RNAseq liver libraries, we found that, under steady-state conditions, mouse hepatocytes primarily express the Adra1b genes. 31,32However, a recent study demonstrated that hepatic injury increases the expression of b 2 AR, which promotes liver regeneration. 33Therefore, we employed two strategies to study the stimulatory pathways driving the hepatoprotective effect provided by adrenergic signaling.
First, we used a triple KO model, Adra2ac À/À Adrb2 À/À , a mouse strain that exhibits hyperactivity of the sympathetic nervous system along with the absence of the b 2 AR.Similar to what was observed in Adra2ac À/À mice, no significant alterations in ALT and AST levels were observed after aGC administration, indicating that in our model, b 2 AR signaling is dispensable for the hepatoprotective effect promoted by increased NE release (Figure S3).(A-C) C57BL/6J WT and Adra2ac À/À mice were treated with aGC (2mg/animal), and serum cytokine production was evaluated after 4 h (A).Mean G SEM of 3 different experiments (n = 4-6), all data had a minimum p < 00.5 versus the respective control group.For the ex vivo, intracellular cytokine assay, splenic (B) and hepatic (C) cells were harvested 90 min after aGC administration and incubated with brefeldin and monensin for 2 h.Subsequently, staining was performed for the detection of IFN-g and TNF-a production by iNKT cells.Data represent the mean G SEM of 2 independent experiments (n = 5), p < 0.0001 versus the respective control group.
Next, we pre-treated C57BL/6J WT mice with phenylephrine (Phe), an a 1 AR agonist, before aGC administration. 34This approach did not affect the overall development of a systemic pro-inflammatory response or the specific production of IFN-g or TNF-a by iNKT cells induced by the in vivo administration of aGC (Figures 6A-6C).Despite high amounts of pro-inflammatory cytokines, no significant alterations in the ALT and AST levels were observed compared to control mice, in contrast to the aGC-treated group (Figures 6D and 6E).

Adrenergic signaling fails to increase hepatocyte resistance to apoptosis
To determine the effect of adrenergic signaling on hepatocyte survival, we took advantage of the murine-derived hepatocyte cell line, AML12.Although AML12 hepatocytes proved to be insensitive to TNF-a-induced apoptosis, treatment with NE or Phe failed to inhibit the increase in TNF-a-induced Fas expression (Figures 7A and 7B).Next, AML12 were cultured under the same conditions in the presence of FasL and TNF ligand enhancer.Under this condition, activation of the FasL/Fas pathway induced hepatocyte death, which was not inhibited by adrenergic signaling (Figure 7C).Hence, adrenergic signaling does not enhance the hepatocyte resistance to TNF-a/FasL-Fas-induced cell death.
We then co-cultured aGC-pulsed AML12 cells with hepatic leukocytes (HL) for 4 h under two conditions: 1) AML12 cells were pre-treated with NE or Phe for 2 h and subsequently co-cultured with HL; 2) AML12 cells were incubated with HL previously treated with NE or Phe for 2 h.
Pre-treating AML12 cells with NE or Phe did not protect the hepatocytes from aGC-induced apoptosis (AnnexinV + ) (Figure 7D).Similarly, pre-incubating HL with NE or Phe did not inhibit AML12 apoptosis, thus supporting the idea that adrenergic signaling is ineffective in inhibiting iNKT-induced hepatocyte apoptosis (Figure 7E).In addition to the direct cytotoxic effect on CD1d-expressing target cells, aGC-induced iNKT cell activation has also been associated with the early appearance of pro-inflammatory neutrophils, which contribute to the extension of liver damage. 35In this regard, flow cytometric analysis of HL revealed that pre-treatment with NE or Phe impaired the shift of neutrophils toward a more inflammatory profile (Figures 7F and 7G).

Adrenergic signaling inhibits a-galactosylceramide-induced migration of inflammatory cells to hepatic parenchyma
To determine the influence of adrenergic signaling on the influx of inflammatory cells into the liver after iNKT cell activation, we first analyzed the profile of neutrophils and inflammatory macrophage-like cells (CD11b + F4/80 + ) in the liver of Adra2ac À/À mice.Although aGC administration induced the influx of inflammatory leukocytes into the hepatic parenchyma in both the WT and Adra2ac À/À mice, the latter animals exhibited fewer infiltrating cells (Figure 8A).As a result, the total number of neutrophils and CD11b + F4/80 + was lower in the Adra2ac À/À mice (Figures 8B and 8C).Flow cytometric analysis also revealed a reduction in the expression of CD11b in both cellular populations in these animals, indicating a lower pro-inflammatory profile compared to the WT group (Figures 8D and 8E).
Therefore, these data support the notion that although adrenergic signaling does not influence the iNKT cells response to aGC, Adra2ac À/À mice exhibit a reduced inflammatory reaction due to adrenergic signaling on myeloid-derived cells.Indeed, pharmacological modulation of the adrenergic response by Yoh or Phe treatment replicated the same inflammatory profile observed in Adra2ac À/À mice, as both treatments reduced the influx of inflammatory cells (Figures 8F-8H).Notably, Phe's influence on myeloid-derived cells appears more pronounced than that of Yoh, suggesting a more suppressive effect for a 1 AR adrenergic stimulation (Figures 8I and 8J).

DISCUSSION
Alterations in neuro-endocrine-immune homeostasis promoted by the activation of the sympathetic nervous system are linked to susceptibility to tumor development and resistance to chemo/radiotherapies. 36,37 As aGC-based immunotherapy is highlighted as an alternative approach to overcome tumor resistance to conventional therapies, we aimed to determine the impact of adrenergic signaling on the antimetastatic activity of iNKT cells.We observed that a combination of SNS stimulation and iNKT cell activation could promote tumor elimination without the deleterious liver damage associated with aGC therapy.
Norepinephrine signaling through the b 2 AR during the ex vivo stimulation of conventional naive CD4 + T lymphocytes impairs IFN-g production, leading to a skewing toward the Th2 profile. 19,21Therefore, considering that aGC's ability to control tumor growth depends on the production of IFN-g by iNKT cells, which express a and bARs, our first question was to determine whether adrenergic signaling impairs their activation toward a type 1 profile. 24,38In this context, in vitro treatment with NE did not inhibit the production of IFN-g or TNF-a by splenic iNKT cells upon cognate antigen stimulation.9][40][41] In line with the concept that iNKT cells coordinate various immune responses, both WT and the Adra2ac À/À mouse strains exhibited a substantial and similar increase in systemic levels of several immunomodulatory molecules following aGC administration. 42These results thus supported the idea that iNKT cells, unlike other immune cells, are refractory to the suppressive effect of adrenergic signaling.
An important aspect of the iNKT cell biology is that their rapid activation by aGC stimulation results in a dual-edged immune reaction.While it potentiates host resistance against pathogens or cancer through type-1 immunity, it also triggers a cascade of inflammatory events leading to liver injury.Several experimental studies have implicated TNF-a-producing iNKT cells in hepatocyte damage, characterized by elevated systemic ALT and AST levels and the emergence of necrotic areas in liver parenchyma. 27,43,44n contrast to what was observed in WT mice, aGC administration in Adra2ac À/À animals did not result in liver dysfunction despite the cytokine storm, the presence of hepatic TNF-a-producing iNKT cells, in situ, and a significant influx of inflammatory cells into liver parenchyma.Therefore, these data suggested that instead of suppressing the aGC-induced immune response, adrenergic signaling promotes hepatocyte protection against iNKT-mediated cytotoxicity, raising the question of whether this phenomenon could be extended to malignant cells.To answer this question, we decided to determine the impact of adrenergic signaling on the antimetastatic capacity of aGC-activated iNKT cells.While the sympathetic nervous system's hyperactivity rendered Adra2ac À/À mice more susceptible to lung metastasis, aGC treatment overcame the pro-metastatic effect of the adrenergic signaling, promoting melanoma control with the same efficiency observed in WT mice.Despite the antitumor activity, we did not detect any signal of liver injury in the aGC-treated Adra2ac À/À mice, in contrast to WT mice injected with aGC.Therefore, these results underscored adrenergic signaling as a target for modulating liver lesions during iNKT-based immunotherapy.
Considering adrenergic signaling as a target to improve aGC immunotherapy, we hypothesized that the punctual stimulation of NE release could maintain liver physiology despite iNKT cell activation.To mimic the Adra2ac À/À mice phenotype, we pre-treated C57BL/6J WT mice with Yoh, an a 2 AR antagonist. 29Alongside the positive response to the antimetastatic effect of aGC treatment, the liver physiology was preserved in tumor-bearing Yoh-pre-treated mice, further supporting the idea that adrenergic signaling protects hepatocytes from iNKT-mediated injury.Notably, the hepatoprotective effect provided by Yoh pre-treatment extended to PBS44, an aGC analog (C) AML12 cells were pre-treated with NE or Phe for 2 h, stimulated with TNF-a (20 ng/ml) for 4 h, and treated with FasL (100 ng/ml) in the presence of TNF ligand enhancer (1 mg/ml) for additional 12 h; cytotoxicity was evaluated by LDH release in the supernatant, and graphs represent the percentage of cell death.(D) AML12 cells were pulsed with aGC (250 ng/ml), pre-treated with NE or Phe for 2 h and co-cultured with liver leukocytes for 4 h; graphs represent the percentage of apoptotic hepatocytes (SSC/FSC high CD45 À Live/dead À AnnexinV + ).(E-G) Liver leukocytes were pre-treated with NE or Phe for 2 h and co-cultured with AML12 pulsed with aGC for 4 h.Graphs represent the percentage of apoptotic hepatocytes (E), CD11b high neutrophils (SSC/FSC int CD45 + CD11b + Ly6G + ) (F), and CD11b MFI (G).Data represent the mean G SD from a single set of two independent experiments (n = 5) **p < 0.01, ***p < 0.001.that shares its antimetastatic ability, serving as another potential immunotherapeutic agent. 45These results thus corroborated the idea that stimulating sympathetic fibers is an alternative to improving the outcomes of iNKT-based immunotherapy.
Finally, we directed our focus toward the signaling pathways driving adrenergic-mediated hepatocyte protection.Under steady-state conditions, the most expressed adrenergic receptor in hepatocytes is the a 1B AR (Adra1b>Adra2b>Adrb3>Adra1a), which could account for the maintenance of liver physiology/structure in the Adra2ac À/À mice. 31Despite a study showing the induction of b 2 AR following partial hepatectomy (PHx), our findings did not associate b 2 AR signaling with the protection against iNKT-induced hepatotoxicity. 33Using the Adra2ac À/À Adrb2 À/À mice, we demonstrated that even with b 2 AR deficiency, the liver injury induced by aGC administration remained under the control of the adrenergic signaling promoted by the absence of Adra2ac, indicating that b 2 AR is dispensable for hepatocyte protection against immune-mediated injury.Although our approach cannot entirely exclude the participation of the other members of bAR family, no evidence supports a role for b 1 or b 3 AR in injury-induced hepatocyte survival/proliferation. 33 Regarding the potential implication of a 2B AR, a recent study demonstrated that in the case of PHx, the regenerative effect of an a 2 AR agonist is due to macrophage differentiation toward an anti-inflammatory profile and not from direct action on hepatocytes. 46Since we did not find any indication of alteration in the pro/anti-inflammatory cytokines ratio due to the adrenergic signaling, we supposed that this was not the principal mechanism maintaining liver physiology after iNKT cell activation.Consequently, we opted to pre-treat mice with Phe, a selective a 1 AR agonist, before aGC administration. 47This approach inhibited the development of liver injury induced by iNKT cell activation without altering the production of pro-inflammatory cytokines induced by aGC.
Previous studies have demonstrated that liver injury due to aGC-induced iNKT cell activation depends on the FasL/Fas pathway and TNF-a production. 13,27Although TNF-a alone cannot induce hepatocyte apoptosis, it does stimulate Fas expression on their surface, rendering them susceptible to FasL-induced cell death. 48Considering that the adrenergic signaling does not impair TNF-a production, our initial hypothesis was that it might protect hepatocytes from apoptosis by inhibiting the FasL/Fas pathway.To test this assumption, we stimulated AML12 cells with NE or Phe and then incubated them with TNF-a or FasL.(A-E) C57BL/6J WT and Adra2ac À/À mice were treated with aGC (2mg/animal), and liver leukocytes were analyzed by flow cytometry after 3 h.Graphs represent the total number of liver leukocytes (CD45 + ) (A), neutrophils (CD11b + Ly6G + ) (B) and CD11b + F4/80 + cells (C), as well as the CD11b MFI in neutrophils (D) and CD11b + F4/80 + cells (E).(F-J) C57BL6/J WT animals were pre-treated with Yoh (5 mg/kg) or Phe (10 mg/kg) and then treated with aGC (2mg/animal).Liver leukocytes were analyzed by flow cytometry after 3 h, and graphs represent the total number of liver leukocytes (CD45 + ) (F), neutrophils (CD11b + Ly6G + ) (G) and CD11b + F4/80 + cells (H), as well as the CD11b MFI in neutrophils (I) and CD11b + F4/80 + cells (J).Data represent the mean G SD from a single set of two independent experiments (n = 5).*p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.
However, both stimuli failed to inhibit TNF-a-induced Fas expression on the AML12 cell surface or prevent their death via the FasL/Fas pathway.This suggested that adrenergic signaling does not have a direct impact on hepatocytes.To determine the influence of adrenergic signaling on iNKT-mediated hepatocyte apoptosis, we then co-cultured aGC-pulsed AML12 cells with hepatic leukocytes pre-incubated with NE or Phe.Corroborating the notion that iNKT cells are resistant to the suppressive effect of adrenergic signaling, aGC-pulsed AML12 cells underwent apoptosis upon contact with hepatic leukocytes.Therefore, these results demonstrated that adrenergic signaling does not directly affect hepatocytes or iNKT-induced apoptosis.
In addition to the implication of the TNF-a/FasL axis in hepatic apoptosis, iNKT cell activation also induces neutrophil accumulation in the liver parenchyma as early as 3 h, accompanied by an increase in caspase3. 35Neutrophil depletion through anti-Ly6G administration inhibited liver injury, in contrast to STAT1 À/À deficiency, which increased infiltrating neutrophils and hepatic damage despite lower levels of Fas expression compared to the WT mice. 35Therefore, apart from TNF-a/FasL-mediated hepatocyte apoptosis, early neutrophil infiltration contributes to liver injury.
In this context, we observed that in vitro adrenergic stimulation impaired the acquisition of a more inflammatory/invasive profile by neutrophils (Ly6G + CD11b high ) treated with NE or Phe before incubation with aGC-pulsed AML12 cells. 49,50In line with this, WT animals exhibited higher hepatic leukocyte counts after aGC administration than the Adra2ac À/À mice.Furthermore, in the latter, the infiltrating neutrophils and myeloid-derived CD11b + F4/80 + mice expressed lower levels of CD11b on their surface, indicating a lesser inflammatory/invasive profile due to adrenergic signaling.Pharmacological stimulation of adrenergic signaling in the WT mice, by Yoh or Phe, replicated the lower inflammatory profile after aGC administration.The hypothesis that NE and Phe modulate the inflammatory profile aligns with previous studies showing the influence of adrenergic receptors on innate immunity. 20n short, aGC administration triggers two hepatotoxic self-limiting pathways.One pathway depends on iNKT-mediated activation of the TNF-a/FasL axis, while another is promoted by the influx of myeloid-derived inflammatory cells, leading to more extensive hepatic damage as reflected by the increase in serum ALT/AST levels.In this context, although adrenergic signaling fails to modulate the cognate activation of iNKT cells and the TNF-a/FasL-induced hepatocyte death, it does restrain the inflammatory response of the myeloid cells and, consequently, mitigates liver injury.
Although further experiments are required to elucidate the mechanisms underlying the protective effect observed upon sympathetic stimulation, it is plausible to assume that signaling through the a 1 AR is the primary pathway that prevents iNKT-induced liver damage.
In conclusion, our results demonstrate that modulating adrenergic signaling during aGC immunotherapy preserves liver physiology without compromising the control of tumor growth.Beyond enhancing the aGC therapeutic outcomes, this approach can also be extended to other iNKT agonists and offers new perspectives in managing immune-mediated liver diseases.

METHOD DETAILS
In vivo treatment with a-galactosylceramide Animals were treated with 2mg of aGC via intraperitoneal injection in 200mL of 0.9% saline.The control group received saline as a vehicle solution.To analyze the iNKT response to aGC ex vivo through intracellular staining, spleen and liver samples were collected 90 min after aGC administration, and cells were cultured for 2 hours in the presence of brefeldin A and monensin. 52The experiments with Yoh or Phe followed the same protocol, but animals were respectively treated with the pharmacological a 2 AR antagonist (5 mg/kg) or the a 1 AR agonist (10 mg/kg) via IP route in 0.9% saline, 30 minutes before aGC.

Spleen cells activation in vitro
For in vitro activation assay, splenic cells were incubated with or without NE (1mM) for 30 minutes, stimulated with anti-CD3/CD28 (2 mg/ml) or mCD1d/PBS57 monomer (1 mg/ml) and anti-CD28(2 mg/ml) for 1 hour, followed by additional 2 hours in the presence of brefeldin A and monensin.

AML12 in vitro assay
For the evaluation of CD95 expression, cells were incubated for 2 hours with or without NE or Phe (10mM) and then treated with recombinant TNF-a (20 ng/ml) for 16 hours.To analyze cell death, AML12 cells were incubated for 2 hours with or without NE or Phe (10mM), followed by treatment with recombinant TNF-a (20 ng/ml) for 4 hours, and then treated with recombinant FasL (100 ng/ml) in the presence of TNF ligand enhancer (1 mg/ml) for an additional 12 hours.Cytotoxicity was assessed by measuring the release of LDH in the supernatant, using triton X-100 (1%) as a positive control for cell death.To evaluate aGC-induced cell death, AML12 cells were pulsed with aGC (250 ng/ml) overnight, followed by incubation with or without NE or Phe (10mM) for 2 hours.Subsequently, these cells were co-cultured with liver leucocytes, pretreated with NE or Phe, for 4 h.
For analysis of the early infiltrating content, the liver was collected 3 hours after aGC treatment and stained with markers against CD11b, F4/80 and Ly6G.The cellular content and CD11b expression in neutrophils (CD11b + Ly6G + ) and CD11b + F4/80 + cells were assessed.

Quantification of in vivo cytokine production
Blood was collected through submandibular puncture of the facial vein 4 hours after aGC administration to assess systemic levels of pro-inflammatory molecules using CBA or ELISA assays.In certain experiments, the systemic IFN-g levels were measured 16 hours after aGC injection.

Analysis of liver injury
Liver injury was assessed by measuring the systemic levels of the hepatic enzymes alanine aminotransferase (ALT) and aspartate aminotransferase (AST), in serum samples collected 16 hours after aGC administration, using commercial kits (Bioclin, BR).In some experiments, organs were harvest, fixed, and paraffin-embedded, and tissue slices with a thickness of 5mm were stained with hematoxylin/eosin solutions for histological analysis.Briefly, digital images were captured using a 4x and 10x objective on a light microscope (Olympus, PA) equipped with a digital camera (Moticam 500, Motic, CA) for histopathologic evaluation.Samples were coded and subsequently evaluated by a trained pathologist unaware of the treatment modalities, scoring them according to observed alterations, such as inflammatory infiltrates and necrotic areas,.The scoring system ranged from zero (no alterations) to 3 (severe alterations).The results represent the mean score obtained for each parameter (0-3).

Figure 2 .
Figure2.Hyperactivation of the adrenergic system does not inhibit cytokine production following iNKT cell activation in vivo (A-C) C57BL/6J WT and Adra2ac À/À mice were treated with aGC (2mg/animal), and serum cytokine production was evaluated after 4 h (A).Mean G SEM of 3 different experiments (n = 4-6), all data had a minimum p < 00.5 versus the respective control group.For the ex vivo, intracellular cytokine assay, splenic (B) and hepatic (C) cells were harvested 90 min after aGC administration and incubated with brefeldin and monensin for 2 h.Subsequently, staining was performed for the detection of IFN-g and TNF-a production by iNKT cells.Data represent the mean G SEM of 2 independent experiments (n = 5), p < 0.0001 versus the respective control group.

Figure 5 .
Figure 5. Pharmacological stimulation of the sympathetic nervous system by yohimbine controls hepatic damage induced by iNKT activation without impairing its antitumoral activity (A-D) C57BL/6J WT mice were treated with Yoh prior to grafting with B16F10 cells and subsequent aGC treatment (2mg/animal).Serum levels of IFN-g (A), ALT (C) and AST (D) were evaluated after 16 h.Pulmonary metastatic nodules were quantified after 15 days (B).(E) C57BL/6J WT mice were treated with Yoh prior to grafting and subsequent PBS44 treatment (2mg/animal).Pulmonary metastatic nodules were quantified after 15 days.(F and G) Serum levels of ALT (F) and AST (G) were evaluated after 16 h.Data represent the mean G SEM of 2 independent experiments (n = 5).****p < 0.0001; ***p < 0.001; **p < 0.01; *p < 0.05.

Figure 6 .
Figure 6.Pharmacological stimulation of a 1 adrenergic receptor controls hepatic damage induced by activated iNKT cells (A) C57BL/6J WT mice were treated with Phe 30 min prior to aGC treatment (2mg/animal).Serum cytokine production was evaluated after 4 h.(B and C) For the ex vivo intracellular cytokine assay, splenic cells were harvested 90 min after aGC administration and incubated with brefeldin and monensin for 2 h, followed by staining for the detection of IFN-g (B) and TNF-a (C) production by iNKT cells.(D and E) ALT (D) and AST (E) levels in serum were evaluated after 16 h.Data represent the mean G SD from a single set of two independent experiments (n = 6, B and C); Mean G SEM of 2 independent experiments, (n = 4-5, A, D, and E).***p < 0.001; **p < 0.01; *p < 0.05.

Figure 7 .
Figure 7. Adrenergic signaling fails to increase hepatocyte resistance to apoptosis (A and B) AML12 cells were pre-treated with NE or Phe for 2 h, treated with TNF-a (20 ng/ml) for 16 h, and the expression of CD95 was evaluated by flow cytometry; graphs represent the median fluorescence intensity (MFI).(C)AML12 cells were pre-treated with NE or Phe for 2 h, stimulated with TNF-a (20 ng/ml) for 4 h, and treated with FasL (100 ng/ml) in the presence of TNF ligand enhancer (1 mg/ml) for additional 12 h; cytotoxicity was evaluated by LDH release in the supernatant, and graphs represent the percentage of cell death.(D) AML12 cells were pulsed with aGC (250 ng/ml), pre-treated with NE or Phe for 2 h and co-cultured with liver leukocytes for 4 h; graphs represent the percentage of apoptotic hepatocytes (SSC/FSC high CD45 À Live/dead À AnnexinV + ).(E-G) Liver leukocytes were pre-treated with NE or Phe for 2 h and co-cultured with AML12 pulsed with aGC for 4 h.Graphs represent the percentage of apoptotic hepatocytes (E), CD11b high neutrophils (SSC/FSC int CD45 + CD11b + Ly6G + ) (F), and CD11b MFI (G).Data represent the mean G SD from a single set of two independent experiments (n = 5) **p < 0.01, ***p < 0.001.