Accumulation of Quinolinic Acid Modulates the Pulmonary Immune Response During Influenza Infection

Abstract

Influenza viruses cause a highly contagious, acute pulmonary disease that results in significant mortality each year. These infections trigger the production of interferons, known to induce the expression of the rate-limiting enzyme in the kynurenine degradation pathway in the lungs. As some kynurenine pathway metabolites are biologically active, we aimed to gain a better understanding of their role in influenza A virus infection. The expression of kynurenine pathway enzymes and the levels of their metabolites were quantified in the lungs of C57BL/6 mice 7 days after infection with an H3N2 influenza A virus (IAV). Furthermore, the impact of quinolinic acid supplementation was evaluated on IAV-infected mice and in vitro, in human monocyte-derived macrophages. The expression of key enzymes (IDO1, KMO, and KYNU) increased in mice in the airways of IAV infected mice. High levels of quinolinic acid were produced in the lungs, as revealed by immunohistochemistry in both epithelial cells and immune cells. Oral quinolinic acid supplementation resulted in higher levels of viral mRNA in the lungs and modulated cytokine production, leading to an increased number of neutrophils and interstitial macrophages in lung tissue. In IAV-infected macrophages, the addition of quinolinic acid was associated with higher levels of viral RNA and protein and in increased antiviral and proinflammatory responses (IFN-β, CXCL-1, and TNF-α). These increases were further reduced by memantine, an NMDA receptor antagonist, suggesting that quinolinic acid may modulate the macrophage immune response via NMDA receptors. A deeper understanding of these mechanisms could lead to new therapeutic strategies for influenza infections.

Introduction

Influenza A viruses (IAV) are frequent cause of respiratory infections in humans, causing seasonal epidemics as well as pandemic infections. ¹ Influenza virus replicates in epithelial cells (the first targets of IAV) ² of the upper and lower respiratory tract after binding of hemagglutinin to sialic acid present on the cell surface. ³ The virus is then internalized into an endosome and fuses with its membrane to release viral RNA, which is transported to the nucleus for transcription and then exported to the cytoplasm for translation. Newly synthesized RNA and viral proteins bundle together to form new virions. Innate immunity of epithelial cells can prevent viral replication thanks to different receptors and antiviral molecules present at each step of viral replication. ⁴ Viral RNA can be recognized by the innate immune system both within the endosome by toll-like receptors (TLRs), notably TLR-3, 7, and 8. ⁵ Binding to these receptors induce cytokine and chemokine production by epithelial cells (IFNs, IL-1β, IL-6, IL-8. . .) which are essential for immune cells recruitment such as monocytes/macrophages, or neutrophils. Macrophages have been shown to play a pivotal role in the elimination of IAV. This is exemplified by alveolar macrophages, which are present in airways lumen and alveoli, through the phagocytosis of virus and infected cells. ⁶ In addition, these cells are involved in the antiviral and immune responses by releasing cytokines, notably type I and III IFN. ⁷ The response leading to numerous expression of IFN-stimulated genes constitutes the most powerful antiviral mechanism and a potent inducer of immune cell recruitment. ⁸ Among them, indoleamine 2,3-dioxygenase-1 (IDO1), the first enzyme of the kynurenine pathway (KP), is induced during IAV infection and plays a role in anti-viral defense through its involvement in the KP. ⁹

The KP is the primary degradation pathway for the essential amino acid, tryptophan (Trp), resulting in the production of nicotinamide adenine dinucleotide (NAD+). ¹⁰ This production occurs through the synthesis, mainly by enzymatic reactions, of several metabolites collectively known as kynurenines. ¹¹ The first step of the KP, which converts Try to kynurenine (Kyn) is completed by dioxygenases IDO1; TDO1 and IDO2 (Figure 1). Tryptophan dioxygenase (TDO2) is constitutively expressed in the liver. ¹² Upon inflammation, the expression of IDO1 and IDO2 is dramatically increased. IDO1 is widely expressed, whereas IDO2 expression is restricted to the liver, kidneys and dendritic cells. ¹³ The kynurenine on tryptophan ratio (K/T) is often used as a marker of inflammation.^14,15 Kyn can then be metabolized via 2 different branches. ¹⁶ In the first one, Kyn produces kynurenic acid (KynA) through the action of kynurenine aminotransferases (KYAT 1-4). ¹⁷ In the second one, under the action of kynurenine-3-monooxygenase (KMO) and kynureninase (KYNU), Kyn is finally metabolized to quinolinic acid (QA). QA is then catalyzed by quinolinate phosphoribosyl transferase (QPRT) to produce NAD+. It was initially hypothesized that IDO activation limits the availability of Try, thereby slowing protein biosynthesis and the subsequent growth of pathogens and proliferating cells ¹⁸ but it was subsequently shown that Try metabolites are bioactive molecules with a direct effect on immune response. ¹⁹ These metabolites are known for their effects on the central nervous system, as well as for their involvement in various psychiatric and neurodegenerative disorders. ²⁰ Over the last decade, they have also been implicated in the development of several other diseases, including cardiovascular, autoimmune and chronic kidney diseases, cancer, and diabetes^21,22 and shown to play an important role in controlling inflammation.^11,23 It has been suggested that an imbalance of downstream metabolites leads to impairment in many physiological processes, particularly those related to inflammation. ²⁴ QA was demonstrated to upregulates chemokine production and receptor expression in astrocytes, in addition to its neurotoxic properties.^25-27 QA has also been implicated in the pathophysiology of multiple conditions including infections and proposed as marker of severity in COVID-19. ²⁸

Figure 1.

Schematic representation of the kynurenine pathway. The enzymes and metabolites studied in this work are shown in bold.

Abbreviations: IDO, indoleamine 2,3-dioxygenase; KMO, kynurenine 3-monooxygenase; KYAT, kynurenine aminotransferase; KYNU, kynureninase; QPRT, quinolinate phosphoribosyl transferase; TDO2, tryptophan dioxygenase.

Interestingly, an increase in Try metabolism has also been observed in sera of patients with influenza infections. ²⁹ In a murine IAV infection model, several metabolites from the KP were shown to be increased in serum 7 days after infection. ³⁰ Moreover, analysis of lung extracts revealed that kynurenine levels were significantly higher in IAV-infected animals than in the control group. ³¹ To prevent severe complications in influenza infection, it may be important to dampen the response in the early stages of a viral infection. ³²

As modulation of inflammation is one of the primary functions of KP, we aimed to improve our understanding of the role of KP metabolites in the inflammatory and antiviral responses following IAV infection. In a murine IAV infection model, we first analyzed KP enzymes expression and metabolites production in the lungs, and evaluated the impact of an oral QA supplementation. Since macrophages populations are affected during IAV infection, we analyzed the direct role of QA on human monocyte-derived macrophages responding to IAV infection.

Results

Influenza Infection Upregulates Kynurenine Pathway in Lungs

In order to study the KP during IAV infection, we used an experimental model of mice infected with H3N2. Seven days after intranasal infection, as expected, viral RNA is detected in lung tissues with significant increase of il-6, il-1β, ifn-β expressions and secretion in lung and bronchoalveolar lavage (BAL) cells (data not shown). In addition, the presence of immune cells in the peribronchial area, as confirmed by analysis on HE stained lung sections, indicates the development of an inflammatory response.

The levels of ido1, kmo, kynu, and qprt are significantly higher in BAL cells whereas only ido1 is significantly higher in total lung extracts with kmo levels decreasing significantly. qprt and kynu mRNA levels are not significantly different in lung tissue (Figure 2A and B). These effects are associated with a significant increase of Kyn and QA concentrations in lung tissue whereas Try concentration is not affected (Figure 2C). K/T ratio is significantly increased (Figure 2C). In serum, Try concentration is decreased, Kyn and QA concentrations are not significantly modulated but K/T ratio remains significantly increased (Figure 2D). A constitutive expression in bronchial epithelium and at a lower level in alveolar wall and alveolar macrophages of IDO1, KMO, and QA is detected by immunohistochemistry in control non-infected mice. Upon infection, there is a strong increase in the staining of IDO1, KMO, and QA in both the alveolar walls and bronchial epithelium as well as within cells of inflammatory infiltrates (Figure 2E). Immunolocalization of QPRT was not performed since validated antibodies to murine QPRT are not available. However, infection with the H3N2 virus appears to stimulate the expression of KP enzymes in lung tissue, both by airway epithelial cells and by immune cells, leading to increased QA in this tissue. Among all the other metabolites of KP detected in our samples, QA was the most abundant.

Figure 2.

H3N2 infection induces kynurenine pathway in mice lung tissues. Mice were infected intranasally with 50pfu H3N2 and sacrificed 7 days post infection (7dpi). ido1, kmo, kynu and qprt mRNA levels were evaluated by RT-qPCR in total lung tissue (A) or in cells recovered from BAL (B). The results were expressed as the ratio to the expression of hprt-1 as a housekeeping gene. Tryptophan, kynurenine and quinolinic acid concentrations were analyzed by mass spectrometry in total lung extract (C), and in serum (D) at 7dpi; K/T is the ratio between kynurenine and tryptophan concentrations. (E) Staining by immunohistochemistry of IDO1, KMO, and QA was performed on paraffin lung sections of control (PBS) and infected mice (H3N2; bar = 50 µm). The results are expressed as mean ± SEM.

*P < .05. **P < .01. ***P < .001. ****P < .0001, (Mann–Whitney; n = 3).

QA Supplementation Impairs Immune Response to H3N2 Infection

To understand the effect of QA accumulation after H3N2 infection on the lung immune response, we chose to increase the concentration of QA by adding it to the drinking water, which is the least traumatic method of administration for infected mice. No significant change in drinking volume was reported in the presence of QA. However, infected mice drank less on day 5 after infection (Supplemental Table 1). After QA treatment, the serum concentration of this metabolite increases (Supplemental Table 2). This supplementation is associated with a significant increase in IAV mRNA in the lungs of mice (Figure 3A). IAV-induced weight loss is not significantly enhanced by QA supplementation (P = .062; Figure 3D). Antiviral response to IAV is characterized by the induction of tlr-3, tlr-7, and ifn-γ mRNA expression in lung tissue and this expression is significantly increased by QA supplementation (Figure 3B and E). However, ifn-β expression and IFN-β and IFN-γ concentrations were not significantly increased in total lung extracts (Figure 3E and F). Lung damage is evaluated by histopathological scores including bronchial and alveolar lesions and demonstrate a worsening in alveolar lesions with QA supplementation (Figure 3C). This similar effect is also observed for il1-β, il-6, and cxcl1 IAV-induced expression in lung tissue (Figure 4A). Although il-10 expression in lung is increased by H3N2 infection its expression is not significantly affected by QA supplementation (Figure 4A). In contrast, IL1-β, IL-6, IL-10, and CXCL-1 IAV-induced concentrations are significantly lower in lung tissue after QA supplementation (Figure 4B). Interestingly, QA has no significant effect on TLRs or cytokines expressions nor on cytokines lung concentrations in not infected mice (Figure 3).

Figure 3.

QA oral administration during H3N2 infection impairs mice antiviral response. Mice were infected intranasally with 50pfu H3N2 and QA was added or not in drinking water (100 mg/L) since the first day of infection. Mice were sacrificed 7 days post infection (7dpi). Non-infected control groups (Control, white) or infected groups (H3N2, grey) with (square) or without (circle) QA in drinking water. (A) Viral RNA was evaluated in total lung by RT-qPCR. (B) tlr-3 and tlr-7 mRNA levels were evaluated by RT-qPCR in total lung tissue. (C) Bronchial and alveolar score analysis. (D) Body weight at day 7 is expressed as a percentage of initial body weight à D0. (E) Ifn-β and ifn-γ mRNA levels were evaluated by RT-qPCR in total lung tissue. (F) IFN-β and IFN-γ protein levels were evaluated by ELISA in total lung extracts (pg/mL). The results are expressed as mean ± SEM.

*P < .05. **P < .01. ***P < .001. ****P < .0001 (2-way ANOVA).

Figure 4.

QA oral administration during H3N2 infection impairs mice inflammatory response. Mice were infected intranasally with 50pfu H3N2 and QA was added or not in drinking water (100 mg/L) since the first day of infection. Mice were sacrificed 7 days post infection (7dpi). Non-infected control groups (Control, white) or infected groups (H3N2, grey) with (square) or without (circle) QA in drinking water. (A) il1-β, il-6, il-10, and cxcl-1 mRNA levels were evaluated by RT-qPCR in total lung tissue. The results were expressed as fold increase compared to control non-infected-mice with no QA in drinking water, using expression of hprt-1 as a housekeeping gene. (B) IL-1β, IL-6, IL-10, and CXCL-1 protein levels were evaluated by ELISA in total lung extracts (pg/mL). (C) CD45+/CD19+ B lymphocytes; CD45+/NK1-1-/TCRβ+/CD8+ T lymphocytes, CD45+/Ly6G+/CD11b+ neutrophils, CD45+/Ly6G-/F4-80+/CD11c-/Ly6C+/CD64-/CCR2+ inflammatory monocytes, CD45+/Ly6G-/F4-80+/CD11c+ alveolar macrophages and CD45+/ Ly6G-/F4-80+/CD11c-/CD64+ interstitial macrophages were evaluated in lung tissue. The results are expressed as mean ± SEM.

*P < .05. **P < .01. ***P < .001. ****P < .0001 (2-way ANOVA).

IAV infection is associated with an increased immune cell recruitment in lung tissues after 7 days of infection. Following infection, there is no significant recruitment of B lymphocytes (CD45+/CD19+) or CD8+ T lymphocytes (CD45+/NK1-1-/TCRβ+/CD8+), but the recruitment of CD8+ T lymphocytes is significantly enhanced following the oral administration of QA (Figure 4C). Neutrophils (CD45+/Ly6G+/CD11b+), inflammatory monocytes (CD45+/Ly6G-/F4-80+/CD11c-/Ly6C+/CD64-/CCR2+) and alveolar macrophages (CD45+/Ly6G-/F4-80+/CD11c+) are recruited to the lung following infection. Oral administration of QA further enhances neutrophil recruitment and significantly increases the number of interstitial macrophages (CD45+/ Ly6G-/F4-80+/CD11c-/CD64+), which are not significantly affected by infection alone (Figure 4C). Other cell populations were not significantly affected by QA oral administration (data not shown).

These results suggest that the presence of QA increases the virus replication, the antiviral response and inflammatory cell recruitment in infected mice.

QA Enhances Inflammatory Response to H3N2 Infection in Human Primary Macrophages

Macrophages are essential for protecting against IAV infection. To study modulation by QA, human monocyte-derived macrophages were infected with the H3N2 virus for 24 hours. mRNA levels of ido-1, kmo, and kynu are strongly enhances by viral infection while the mRNA level of qprt is decreased (Figure 5A). The mRNA levels of tlr-3 and tlr-7 are also increased by viral infection but this increase is reduced in the presence of QA (Figure 5C). The mRNA level of ifn-β and IFN-β secretion are induced by H3N2 infection and are further enhanced by the presence of QA (Figure 5D). Upon addition of QA, IAV mRNA is increased (Figure 5B). This result has been confirmed by labelling IAV nucleoprotein NS1 with a specific antibody: a higher percentage of NS1-positive cells were observed in the presence of QA, suggesting that viral multiplication is more active in the presence of QA (Figure 5E). The mRNA levels of il-6, il-8, cxcl-1, and tnf-α are enhanced by H3N2 infection and further amplified by QA addition, whereas IAV-induced il-10 mRNA is significantly decreased by QA (Figure 6A). Analysis of the supernatant confirms a significant increase in H3N2-induced IL-6, IL-8, CXCL-1, and TNF-α secretion in the presence of QA. However, IL-10 secretion decreases in the presence of QA (Figure 6B). Taken together, these results suggest that the presence of QA facilitates viral replication and enhances the inflammatory response in human macrophages during H3N2 infection.

Figure 5.

QA modulates antiviral response of human primary macrophages after H3N2 infection. Monocyte derived primary macrophages were exposed or not to QA (100 µM) and infected with H3N2 (MOI 0.5); 24 hours after infection supernatants were collected and mRNA was extracted. Non-infected cells (Control, white) or infected cells (H3N2, grey) with (square) or without (circle) QA exposure. (A) ido1, kmo, kynu, and qprt mRNA levels were evaluated by RT-qPCR. (B) Viral RNA was evaluated in total cell extract by RT-qPCR. (C) tlr-3 and tlr-7 mRNA levels were evaluated by RT-qPCR in total cell extracts. (D) ifn-β mRNA level was evaluated by RT-qPCR in total cell extracts and IFN-β secretion was evaluated by ELISA (pg/mL) in supernatants. (E) Expression of viral protein was visualized on macrophages in control (PBS), infected (H3N2) or infected with QA exposition (H3N2+QA); bar = 100 µM. The percentage of positive cells was evaluated on different samples (n = 7). The results are expressed as mean ± SEM, n = 3.

*P < .05. **P < .01. ***P < .001. ****P < .0001 (Mann–Whitney analysis for A, and 2-way ANOVA for B, C, D, E).

Figure 6.

QA modulates inflammatory response of human primary macrophages after H3N2 infection. Monocyte derived primary macrophages were exposed or not to QA (100 µM) and infected with H3N2 (MOI 0.5); 24 hours after infection supernatants were collected and mRNA was extracted. Non-infected cells (Control, white) or infected cells (H3N2, grey) with (square) or without (circle) QA exposure. (A) Il-6, il-8, il-10, cxcl-1, and tnf-α mRNA levels were evaluated by RT-qPCR in total cell extracts. The results are expressed as fold increase compared to control group (no H3N2 and no QA) using hprt-1 expression as a housekeeping gene. (B) IL-6, Il-8, IL-10, CXCL-1, and TNF-α secretion was evaluated by ELISA (pg/mL) in supernatants. The results are expressed as mean ± SEM.

*P < .05. **P < .01. ***P < .001. ****P < .0001 (2-way ANOVA).

It has been demonstrated that QA activates the N-methyl-D-aspartate (NMDA) receptor, which is expressed on macrophages.^33,34 To test whether NMDA receptors are involved in QA effect in our model, we preincubated cells with an NMDA antagonist, the memantine^34,35 before viral infection. Pre-exposure to memantine decreases QA-induced IAV RNA, but this effect remains not significant (Figure 7A). The expression of the NMDA receptor by primary macrophages was assessed by analyzing the expression of the mRNA of 2 NMDA receptor subunits (GRIN1 and GRIN2D). grin1 mRNA expression was undetectable (data not shown), whereas grin2d mRNA was expressed and appeared to be enhanced by infection (Figure 7B). The addition of memantine limits the QA-induced enhancement observed for expression and secretion of IFN-β, CXCL1 and TNFα (Figure 7C and D). Memantine partially counteracts the inhibitory effect of QA on il10 expression (Figure 7C) but not on IL-10 supernatant levels (Figure 7D). These results suggest that activation of NMDA receptors on macrophages may be part of the mechanisms involved in modulating the immune response by macrophages.

Figure 7.

Effect of memantine on QA-induced response of human primary macrophages after H3N2 infection. Monocyte derived primary macrophages were exposed or not to QA (100 µM) and infected with H3N2 (MOI 0.5) in presence or not of 100 µM memantine; 24 hours after infection supernatants were collected and mRNA was extracted. Non-infected cells (Control, white) or infected cells (H3N2, grey) with (square) or without (circle) QA exposure. (A) Viral RNA was evaluated in total cell extract by RT-qPCR. (B) grin2d mRNA levels were evaluated by RT-qPCR in total cell extracts. The results were expressed as the ratio to the expression of hprt-1 as a housekeeping gene. (C) inf-β, il-10, cxcl-1, and tnf-α mRNA levels were evaluated by RT-qPCR in total cell extracts. The results are expressed as fold increase compared to control group (no H3N2 and no QA) using hprt-1 expression as a housekeeping gene. (D) IFN-β, IL-10, CXCL-1, and TNF-α secretion was evaluated by ELISA (pg/mL) in supernatants. The results are expressed as mean ± SEM.

*P < .05. **P < .01. ***P < .001. ****P < .0001 (2-way ANOVA).

Discussion

Previous studies have shown that IAV induces the production of Kyn by inducing the expression of interferons (IFNs), which in turn upregulate IDO1 expression and induce the biosynthesis of Kyn. ³¹ We demonstrate here that the main enzymes of the KP (IDO1, KMO, KYNU, and QPRT) are induced by influenza infection in BAL cells. In the total lung extract, this induction is significant only for ido1, while kmo expression appears decreased. We hypothesize that these unexpected results are related to the presence of structural cells within the lung tissue, but not within the BAL, or to compartmentalized responses between the airway lumen and lung tissue, as has been reported in other contexts. These changes in enzyme expression are accompanied by an increase in metabolite concentration in the pathway, particularly of Kyn and QA in lung tissue. These variations were more striking in lung extracts that in serum that recapitulates the metabolic condition of a whole organism. For example, the significant decrease observed in serum Try concentration in infected mice might be due to IDO1 activation but also to a change in food-intake or a virus-induced dysbiosis that have been reported recently. ³⁶ In addition, a higher increase in kynurenine and in the K/T ratio in serum was detected in patients suffering from community-acquired pneumonia compared to controls. ³⁷ This normalizes with the resolution of inflammation, showing that activation of the kynurenine pathway occurs also upon lung infection in humans.

Our immunohistological analyses demonstrate that QA is abundantly produced locally in the lung both by the epithelial cells and by the immune infiltrating cells. We have therefore wondered what effect QA accumulation might have on the pulmonary immune response. To this end, we administered QA in the drinking water of infected mice. The oral administration of QA was chosen because this method is non-traumatic for the mice and ensures a constant presence of QA in the organism, since QA is known to be filtered by the kidney, and eliminated from the organism within 24 hours. ³⁸ Drinking water administration has previously been used to study the role of KynA supplementation. ³⁹ Adding QA to the drinking water did not change the daily drinking intake, or induce any lethality. Feeding behavior was not monitored in this study. We observed a reduction in drinking especially 4 days after infection, which is a behavior that has been previously described as resulting from influenza infection. ⁴⁰ In mice lung extracts, we observed an increase in cytokine expression (ifn-γ, il1-β, il-6, il-10, and cxcl-1) at the mRNA level between infected animals and QA supplemental animals. However, the corresponding cytokine protein levels decreased in the QA-supplemented group. In the absence of any kinetic analysis of these parameters, we can hypothesize that QA has induced higher cytokine expression coupled with quicker release of cytokines from lung tissue or cytokine binding to their respective receptor on recruited immune cells. As an alternative explanation, it cannot be ruled out that the post-transcriptional modulation of cytokine production in the lung cells of QA-treated mice has been altered.

Our study was complemented by the use of human macrophages. In this model, QA addition has enhanced the IAV RNA and decreased tlr-3 and tlr-7 expression in macrophages. We hypothesize that in macrophages, QA increases viral replication, consequently increasing the secretion of pro-inflammatory cytokines such as IL-6, IL-8, TNFα, and CXCL-1, while decreasing the secretion of the anti-inflammatory cytokine IL-10. The presence of QA can interfere with the recognition of viruses and the efficient initiation of the immune response and recruitment of immune cells. This can lead to the development of viruses at the onset of infection. Interestingly, the amplification of the inflammatory reaction results in the amplification of the lung lesions consecutive to IAV infection and in an enhanced recruitment of neutrophils. This is probably consecutive to the higher observed secretion of CXCL-1, a chemokine implied in their recruitment. ⁴¹ The clearance of respiratory viruses relies on the activation of CD8+ T cells that recognize viral epitopes and limit viral reservoirs by lysing infected lung epithelial cells. ⁴² Interstitial macrophages were shown to mediate the clearance of apoptotic alveolar epithelium during influenza infection, ⁴³ therefore their increase after QA administration during H3N2 infection may reveal a higher level of apoptosis in the respiratory epithelium In addition, lung interstitial macrophages were shown to trigger regulatory T cells and to secrete IL-10, which might further attenuate lung immune responses.^44,45 We have demonstrated that QA can stimulate the production of the viral nucleoprotein protein NS1. This protein can counteract antiviral responses to facilitate viral replication, as it greatly limits the production of several Interferon Stimulated Genes (ISGs). ⁴⁶ Kinetic studies should be conducted to elucidate the sequence of events that leads to an increase in viral replication. In addition, studies on respiratory epithelial models will also provide insight into the global mechanisms that occur within lung tissue.

The induction of the kynurenine pathway following infection with a neurotropic influenza A virus has already been demonstrated in primary cultures of mouse brain cells and in the brains of mice on day 3 after birth. Following this infection, levels of kynurenic acid increased, which was suggested to counteract the potential excitotoxic effects of QA on NMDA receptors. ⁴⁷ Indeed, several deleterious effects of QA on the central nervous system are associated with NMDA receptor activation. However, its antimicrobial response remains unclear. Our study suggest that QA may act on macrophages to modulate inflammatory response via NMDA receptor in our in vitro model, but this remains to be confirmed in vivo. The nature of NMDA receptor subunits in primary macrophages is poorly documented. Nevertheless, a recent study described GRIN1 expression only in M2 macrophages and not in unpolarized macrophages. ⁴⁸ This study reported that QA can activate NMDA receptors on macrophages, thereby inducing the priming of macrophages toward an M2-like phenotype. ⁴⁸ In our model, monocyte-derived macrophages are polarized to an M1 phenotype using GM-CSF, ⁴⁹ and displayed an enhancement of the inflammatory response to H3N2 which is abrogated by an NMDA antagonist. Our data showed that GRIN1 mRNA (grin1) is not expressed whereas GRIN2D (grin2d) is expressed in our macrophages. To better interpret our results, a comprehensive study of NMDARs in human macrophages, including a description of the subunits and their functionality, would be necessary. In addition, we have to demonstrate the same effect on alveolar or lung interstitial macrophages in order to conclude that this mechanism is relevant in the lung physiology since experiments using cultured macrophages must be interpreted with caution, as the macrophage phenotype is clearly dependent on its environment.^50,51 Nevertheless, it has also been demonstrated that QA can also activate the aryl hydrocarbon receptor in skin cells during psoriasis. ⁵² Therefore, the interaction with NMDA receptor may not be a unique mechanism that induce an effect on macrophages.

A key finding is that QA alone does not affect the antiviral or inflammatory responses, either in vivo or in macrophages. Therefore, the effects of QA on immune responses are contingent on external stimulation. This dependence raises the question of QA’s mechanism of action. The present study hypothesizes that recognition of viral pathogen-associated molecular patterns could be a prerequisite for allowing the effects of QA, potentially rendering its action TLR-dependent. Then the significance of QA in other infections, such as bacterial infections, may be entirely different. Additionally, prior influenza vaccination may affect the accumulation of QA in the lungs and the overall activation of the KP. We hypothesize that precisely modulating QA in the lungs during influenza infection could reduce the pathogenesis of IAV infection and prevent excessive host inflammatory responses. However, QA may also play an important role in replenishing levels of NAD+, a molecule implicated in modulating the outcomes of viral infections and in meeting the body’s energy demands in response to heightened cellular stress. ⁵³ QPRT is a key enzyme that is known to ensure the production of NAD+ and the catabolism of QA. ¹¹ We observed a decrease in qprt expression in primary macrophages infected with H3N2. Consequently, the reduction in QPRT may prevent the chemical degradation of QA, thereby contributing to its accumulation in the lungs. This phenomenon has been observed in the nervous system, where the accumulation of QA is due to the recruitment of macrophages and the saturation of QPRT enzymatic activity. This leads to the accumulation of QA in the central nervous system, resulting in neuronal degeneration, inflammation, and the development of neurological diseases. ⁵⁴

This study describes the accumulation of QA in the lungs of influenza-infected mice and examines its impact on macrophages 1 main player in the antiviral response. Our study suggests that QA could have a detrimental effect on the pulmonary immune response to viral infection. Future studies are required to investigate the contribution of other cell types, such as epithelial cells, as well as the efficacy of counterbalancing QA production by activating the enzymatic activity of QPRT using drugs, to identify new therapeutic targets for combatting the deleterious outcomes of influenza viral infections.

Methods

Animals

Six-week-old male wild-type C57BL/6 (H-2Db) mice were purchased from Janvier (SOPF animal facility, Le Genest-Saint-Isle, France). All animal work conformed to the guidelines of the Animal Care and Use Committee of Nord-Pas-de-Calais (agreement no. AF 16/20,090). The mice were housed in a temperature-controlled facility (23°C) with a strict 12-hour light/dark cycle, and were given food and water ad libitum. In some experiments, 100 mg/L of QA (MedChemExpress, Monmouth Junction, NJ, USA) was added to the drinking water. The drinking water containing QA was changed every 2 days.

Viral Challenge

To induce infection, the frozen working stocks of the pathogenic, murine-adapted H3N2 influenza A virus strain (Scotland/20/74) ⁵⁵ were diluted in phosphate-buffered saline (PBS) were diluted in phosphate-buffered saline (PBS). Intranasal infection was performed under anesthesia with xylazine (16 mg/kg) and ketamine (80 mg/kg), with 50 pfu of IAV administered per mouse. Control mice received 50 μL of PBS.

Sample Collection and Processing

The mice were euthanized 7 days after infection. Bronchoalveolar lavage (BAL) fluids, lungs and blood samples were collected and kept on ice until processing, or frozen immediately in liquid nitrogen. BAL was performed by instilling 0.5 mL of sterile PBS 3 times. After centrifugation at 400× g for 6 minutes at 4°C, the cell-free BAL fluid supernatant was stored at −20°C for cytokine analysis (ELISA).

Lung Histology

The posterior lobes of the lungs were fixed in 10% formalin (Bio-Optica, Milan, Italy) for histopathological studies. After embedding in paraffin wax, lung sections were stained with haematoxylin and eosin. Lung injury was scored, as previously defined. ⁵⁶ For immunohistochemistry, paraffin-embedded sections were deparaffinized with Q Path^® Safesolv (VWR, Radnor, PA, USA), rehydrated and subjected to epitope unmasking as previously described. ⁵⁷ The primary antibodies used were anti-IDO1 (bs-15493R, Bioss, Woburn, MA, USA), anti-KMO (10698-1-AP, Proteintech, Manchester, UK) and anti-QA (IS1010, Immusmol, Bordeaux, France). These were all rabbit polyclonal antibodies. The rabbit-specific HRP/DAB Detection IHC Kit (ab236469, Abcam, Cambridge, UK) was used according to the manufacturer’s recommendations. Counterstaining was performed with hematoxylin. A negative control was performed using no primary antibody, revealing no immunostaining.

Flow Cytometry

The total cells of the lungs were incubated with the appropriate panel of antibodies for 30 minutes in PBS 2% FCS. Following conjugated antibodies against mouse were used: CD5 (FITC), NK1.1 (PerCp-Cy5.5), CD4 (PE-Cy7), CD25 (APC), TCRγδ (APC-Vio770), TCR-β (V450), CD8 (V500), I-Ab (FITC), F4/80 (PE), CD45 (V500), and CCR2 (APC) from Miltenyi Biotech, Paris, France. Antibodies against mouse CD69 (Alexa700), CD103 (PerCP-Cy5.5), CD11c (PE Cy7), CD86 (Alexa-700), Ly6G (APC-H7), CD11b (V45O), and CD64 (BV780) were from BD Biosciences, Franklin Lakes, USA. Antibodies against mouse CD45 (BV605) and Ly6C (BV605) were from Biolegend, San Diego, USA. Data were acquired on an LSR Fortessa (BD Biosciences, Franklin Lakes, USA) and analyzed using FlowJo™ software v7.6.5 (Stanford, CA, USA). The gating strategy has been previously described.^{56 -58} Absolute cell numbers were calculated according to the total cell numbers and the frequency of each population among the CD45+ immune cells. Absolute cell numbers were calculated. This was done according to the total cell numbers and the frequency of each population among the CD45+ immune cells.

Cell Culture

Whole blood was obtained from anonymous healthy adult donors at the Etablissement Français du Sang (French National Blood Service; agreement 2024/174). Peripheral blood mononuclear cells (PBMCs) were isolated using Ficoll-Paque™ PLUS density gradient medium (Cytiva, Wilmington, DE, USA). Human monocytes were then purified from the PBMCs using a MACS column and anti-CD14 monoclonal antibody-conjugated microbeads (Miltenyi Biotech). Immature human monocyte-derived macrophages were generated by culturing the monocytes for 7 days in RPMI 1640 + Glutamax (Gibco, Invitrogen, Carlsbad, CA, USA), supplemented with 10% heat-inactivated fetal calf serum (FCS; Invitrogen), GM-CSF (20 ng/mL; PromoCell, Heidelberg, Germany), and 100 U/mL penicillin/streptomycin. Influenza infection was performed with H3N2 with a multiplicity of infection (MOI) of 0.5. According to the experimental design, the cells were either exposed to 100 µM of QA or left unexposed. QA toxicity had been previously evaluated by an MTT assay ⁵⁹ on primary macrophages. Briefly, QA was added at a concentration ranging from 10 to 150 µM during 24 hours. Following QA treatment, MTT reagent (3-(4,5-Dimethyl-2-thiazolyl)-2,5-diphenyl-2H-tetrazolium bromide, Merck, Darmstadt, Germany, 5 mg/mL) was added (1 mg/mL final) to each well. After 24 hours incubation, medium was removed, and cells treated with 100 µL of 10% SDS solution during 24 hours. Absorbance was then evaluated at a wavelength of 571 and 620 nm as reference. A concentration of 100 µM was the highest concentration that induced no toxicity in primary macrophages, which may reflect the high concentrations that can be found locally in their vicinity, as suggested by immunohistochemical experiments (Figure 2E). NMDA receptors were blocked using the inhibitor memantine hydrochloride (M9292, Sigma-Aldrich, St. Louis, MI, USA). Cells were exposed, 24 hours before viral infection, to a concentration of 50 µM, a concentration shown to be non-cytotoxic on macrophages. ³⁴

Viral Protein Labelling

Viral protein labelling was performed on primary human macrophage cultures 24 hours after viral infection. After recovery of the supernatant, the cells were exposed to an anti-Influenza A nucleoprotein mouse primary antibody (clone AA5H, Monoclonal 011MON6100, ARP, Waltham, MA, USA) for 1 hour, at a dilution ×1000 in PBS. Alexa Fluor 488 donkey anti-mouse secondary antibody (A21202, Thermo Fisher, Waltham, MA, USA) was then added for 1 h diluted ×1000 in PBS. Analysis was performed on an EVOS M50000 (EVOSTM LED Cube. GFP 2.0, Thermo Fisher).

Cytokine Quantification by ELISA

Lung protein extracts were prepared by mechanically dissociating (Precellys, Bertin Technologies SAS, Montigny-le-Bretonneux, France) 1 lung lobe in 1 mL of lysis buffer (Tissue-Protein Extraction Reagents, Life Technologies, Villebon-sur-Yvette, France) and a protease inhibitor cocktail (Roche, Newburyport, MA, USA) at 4°C. The resulting supernatant was collected for total protein analysis using a Pierce BCA Protein Assay Kit (Life Technologies). Levels of mouse IL-1β, IL-6, IL-10, CXCL-1, IFN-β, and IFN-γ were determined in the lungs using an enzyme-linked immunosorbent assay (ELISA) according to the manufacturer’s recommendations (R&D Systems).

In human macrophage cell cultures, the levels of IL-6, IL-8, CXCL-1, TNF-α, IFN-β, and IL-10 were also determined using an ELISA kit (R&D Systems) according to the manufacturer’s instructions.

mRNA Expression Quantification by Reverse Transcription-Polymerase Chain Reaction (RT-qPCR)

Total RNA was extracted from the mouse lungs using the NucleoZol reagent (Macherey-Nagel, Hoerdt, France), following the manufacturer’s instructions. Reverse transcription (RT) was performed using a High-Capacity cDNA Reverse Transcription Kit (Applied Biosystems, Waltham, MA, USA) according to the manufacturer’s recommendations. Real-time PCR was performed using AceQ Universal SYBR Green Master (Q511-03, Vazyme, Nanjing, China) for some genes. The primers used in this study, which were obtained from Eurofins Genomics, are described in Tables 1 and 2. The expression of other genes was studied using a PrimeTime probe assay (IDT, Leuven, Belgium; see Supplemental Table). Quantitative PCR was performed using the QuantStudio 5 Real-Time PCR System (Applied Biosystems, Waltham, MA, USA) to quantify the mRNA of interest. hprt-1 was used as the internal reference gene to normalize the transcript levels. The relative mRNA level (2^{(−ΔΔ Ct)}) was determined by comparing the PCR cycle thresholds (Ct). The results were expressed as a fold increase compared to the control sample. However, when the control sample showed no expression, the results were expressed as a ratio of hprt-1 expression. Detection of IAV was performed using a method based on dual-labelled oligonucleotides and the exonuclease activity of the Taq polymerase enzyme (see primer sequences in Table 1). This method was performed using the qPCRBIO Probe Mix Lo-ROX kit (Eurobio).

Table 1.

Sequence of Primers Used for Murine and IAV qPCR.

Gene	Sequences
Mouse (SYBR)
hprt1	Sens: 5′- TCCTCCTCAGACCGCTTTT -3′
	Antisens: 5′- CCTGGTTCATCATCGCTAATC -3′
cxcl1	Sens: 5′- ACTGCACCCAAACCGAAGTC -3′
	Antisens: 5′-TGGGCACACCTTTTAGCATCTT -3′
ifnb	Sens: 5′- TCCACCAGCAGACAGTGTTTCT -3′
	Antisens: 5′- GGTACCTTTGCACCCTCCAGTA -3′
ifng	Sens : 5′- CAA CAG CAA GGC GAA AAA GG -3′
	Anti sens: 5′ - GTG GAC CAC TCG-GAT GAG CT 3′
il1b	Sens: 5′- TGAAGGGCTGCTTCCAAACCTTTGACC -3
	Antisens: 5′- TGTCCATTGAGGTGGAGAGCTTTCAGC -3′
il10	Sens: 5′- CATTTGAATTCCCTGGGTGAGA -3′
	Antisens: 5′- TGCTCCACTGCCTTGCTCTT -3′
tlr3	Sens: 5- TCATTCTCCCTTGCTCACTCTCA -3′
	Antisens: 5′- AAGGCCGAGATCAAGTATCCTGA -3′
tlr7	Sens: 5′- TGTTCTATGGAGAGCCGGTGA -3′
	Antisens: 5′- GCGGCATACCCTCAAAAACC -3′
Mouse (prime-time assays)
hprt1	Mm.PT.39a.22214828
il6	Mm.PT.58.10005566
ido1	Mm.PT.58.29540170
kmo	Mm.PT.58.9602069
kynu	Mm.PT.58.41853210
qprt	Mm.PT.58.7691745.gs
IAV
M2 gene	Sens: 5′- AAGACCAATCCTGTCACCTCTGA -3′
	Antisens: 5′- CAAAGCGTCTACGCTGCAGTCC -3′
	Probe: 5′[FAM] TTTGTGTTCACGCTCACCGTCCC [TAM] -3′

Table 2.

Sequence of Primers Used for Human qPCR.

Gene	Sequences
Human (SYBR)
hprt1	Sens: 5′- TGCAGACTTTGCTTTCCTTGGTCAGG -3′
	Antisens: 5′- CCAACACTTCGTGGGGTCCTTTTCA -3′
cxcl1	Sens: 5′- TCACCCCAAGAACATCCAAAGT -3′
	Antisens: 5′- TTGAGGCAAGCTTTCCGC -3′
icam-1	Sens: 5′- CAAGGGGAGGTCACCCGCGAGGTG -3′
	Antisens: 5′- TGCAGTGCCCATTATGACTG -3′
ifnb	Sens: 5′- TCTCCTGTTGTGCTTCTCCAC -3′
	Antisens: 5′- GGCAGTATTCAAGCCTCCCAT -3′
tlr3	Sens: 5′- GGTCCCAAGCCTTCAACGA -3′
	Antisens: 5′- GGTGAAGGAGAGCTATCCACATTT -3′
tlr7	Sens: 5′- AAGAATTTGTCTCTTCAGTGTCCA -3′
	Antisens: 5′- ATCTTGGCACCTCTCATGCT -3′
tnfa	Sens: 5′- AGCCCTGGTATGAGCCCATCTA -3′
	Antisens: 5′- CGGCAAAGTCGAGATAGTCGG -3′
Human (prime-time assay)
hprt1	Hs.PT.58v.45621572
grin1	Hs.PT.58.39141804
grin2d	Hs.PT.58.2807609
il6	Hs.PT.58.40226675
il8	Hs.PT.58.39926886.g
il10	Hs.PT.58.2807216
ido1	Hs.PT.58.924731
kmo	Hs.PT.58.21410295
kynu	Hs.PT.58.665413
qprt	Hs.PT.58.142969

Determination of KP Metabolites

An analytical protocol based on liquid chromatography coupled with tandem mass spectrometry (LC-MS/MS), with slight modifications to previously published methods, ⁶⁰ was developed to measure tryptophan (Try) and its metabolites (Kyn and QA). Lung samples were ground with a Precellys^® homogenizer in 300 µL of a 50:50 mixture of PBS and acetonitrile. Then, 50 µL of the prepared lung or serum sample was mixed with 50 µL of acetonitrile (for protein precipitation) containing deuterated compounds at 50 µM as an internal standard, and the mixture was centrifuged at 11 800 rpm at 4°C for 10 minutes. Supernatant (50 µL) was added to deionized water (600 µL). Then, 10 µL of this mixture was injected onto a UPLC-MS/MS system (Acquity TQ-XS Detector, Waters, Milford, MA, USA), which was equipped with a C18-XB column (1.7 µm, 100 Å, 150 mm × 2.1 mm, Phenomenex^®). Ions of each analyzed compound were detected in positive ion mode using multiple reaction monitoring. MassLynx software (Waters) was used for data acquisition and processing.

Statistical Analysis

Data are expressed as the mean ± SEM. The results were statistically analyzed using Prism software (GraphPad, version 9, San Diego, CA, USA): the Mann–Whitney test was used where specified in the figure legends, and the two-way ANOVA analysis was used where specified in the figure legends, followed by Tukey’s multiple comparison test. Differences were considered significant when P < .05; P = non-significant (ns). A significant virus effect (compared to non-infected mice) is indicated by the following symbols: *P < .05, **P < .01. ***P < .001, and ****P < .0001.