AR Deciency Inhibits LPS-induced M1 Response in Macrophages by Activating Autophagy

Macrophage M1 polarization mediates inammatory responses and tissue damage. Recently, aldose reductase (AR) has been shown to play a critical role in of M1 polarization in macrophages. However, the underlying mechanisms are unknown. Here, we demonstrated, for the rst time, that AR deciency repressed the induction of inducible nitric oxide synthase in lipopolysaccharide (LPS)-stimulated macrophages via activation of autophagy. This suppression was related to a defect in the inhibitor of nuclear factor κB (NF-κB) kinase (IKK) complex in the classical NF-κB pathway. However, the mRNA levels of the IKKβ and IKKγ were not reduced in LPS-treated AR knockout (KO) macrophages, indicating that their proteins were downregulated at the post-transcriptional level. We discovered that LPS stimuli induced the recruitment of more beclin1 and increased autophagosome formation in AR-decient macrophages. Blocking autophagy by 3-methyladenine and ammonium chloride treatment restored IKKβ and IKKγ protein levels and increased nitric oxide synthase production in LPS-stimulated AR-decient macrophages. More assembled IKKβ and IKKγ undergo ubiquitination and recruit the autophagic adaptor p62 in LPS-induced AR KO macrophages, promoting their delivery to autophagosomes and lysosomes. Collectively, these ndings suggest that AR deciency involves in the regulation of NF-κB signaling, and extends the role of selective autophagy in ﬁ ne-tuned M1 macrophage polarization. qRT-PCR using the following primers, specic for mouse AR: forward primer, 5′-ACGGCTATGGAACAACTA-3′ and reverse primer, 5′-TGTGGCAGTATTCAATCAG-3′; mouse IKKβ: forward primer, 5′-GGAGCCTGGGAAATGAAAGAA-3′ and reverse primer, 5′-GCCAGAGCCCTACCTGATTG-3′; mouse IKKγ: forward primer, 5′-AAGCACCCCTGGAAGAACC-3′ and reverse primer, 5′-CCTGCTCTGAAGGCAGATGTA-3′; and mouse β-actin: forward primer, 5′-CGTGCGTGACATCAAAGAGAA-3′ and reverse primer, 5′-GCTCGTTGCCAATAGTGATGA-3′. Gene expression was normalized using β-actin as an internal control, and fold changes were calculated.


Introduction
Macrophages demonstrate signi cant plasticity and can modify their phenotype and function in response to their microenvironment [1]. Macrophages are roughly categorized into two different subsets, namely, in ammatory M1 and anti-in ammatory M2 macrophages [2]. In response to lipopolysaccharide (LPS) treatment either alone or in combination with proin ammatory cytokines, such as interferon-γ, macrophages undergo M1 polarization, characterized by the expression of pro-in ammatory cytokines and cytotoxic mediators (reactive oxygen and nitrogen species), as well as increased phagocytic and antigen-presenting capacity [3].
Aldose reductase (AR), a rate-limiting enzyme in the polyol pathway that catalyzes the reduction of glucose to sorbitol in the presence of reduced nicotinamide adenine dinucleotide phosphate, has emerged as a molecular target in multiple in ammatory diseases [4]. Ravindranath et al. demonstrated that transgenic mice overexpressing AR show a more pronounced in ammatory response in a cecal ligation and puncture model [5]. In addition, AR inhibition suppresses in ammatory disorders or immune responses in several other models [6][7][8][9][10]. It was reported that M1-polarized human monocyte-derived macrophages expressed signi cantly higher levels of AR mRNA and AR protein compared with M2polarized macrophages in vitro [11]. Accumulating evidence implicates the classical nuclear factor-κB (NF-κB) signaling pathway in the modulation of M1 macrophage polarization [12][13][14][15]. Previously, Zhang et al. reported that the expression of AR is upregulated after spinal cord injury in wild type (WT) mice, while phosphorylated NF-κB is downregulated, and the number of M1-like macrophages is decreased after spinal cord injury in the AR knockout (KO) mice [16]. However, the exact mechanisms involved are unknown.
Recently, it has been reported that AR de ciency or inhibition enhances autophagy in mouse cardiac myocytes under pathological cardiac hypertrophy or fasting conditions [17,18]. Autophagy is an essential cell-intrinsic mechanism that affords protection against starvation; moreover, it represents a quality-control system that can deliver damaged organelles, worn-out or misfolded proteins, and invading microorganisms from the cytoplasm to the lysosomes for degradation [19]. Defects in autophagy are linked to many human diseases and to the function of cells of the immune response [20]. Experimentally, Toll-like receptor 4 (TLR4) is a sensor for autophagy associated with innate immunity in macrophages [21].
Herein, we demonstrate that LPS treatment leads to a defect at the IKK complex level and the production of more autophagosomes in AR-de cient macrophages. These ndings prompted us to investigate the inherent relationships between these processes.

Materials And Methods
Isolation and culture of BMMs WT and AR KO C57BL/6J mice were bred at the Laboratory Animal Center of Fourth Military Medical University. Bone marrow cells were isolated from adult mice and processed as described previously [22].
Then the cells were cultured in DMEM (Gibco, Carlabad, CA, USA) supplemented with 0.001% βmercaptoethanol, 1% penicillin/streptomycin, 1% HEPES, 10% FBS (Gibco, Carlabad), and 20% sL929 supernatant from sL929 cells, which secrete macrophage colony-stimulating factor (M-CSF) required for the promotion of hematopoietic stem cells differentiation into macrophages. After 7 days of incubation with conditioned medium, the oating cells were removed, and the viable cells attached to at-bottomed 6-or 24-well plastic culture plates or 75 cm 2 culture asks (Nunc, Roskilde, Denmark) were used to obtain BMMs. Next, the culture media were changed to complete medium, without sL929 cell culture supernatant; the cells were further cultured for 3 days to restore BMMs to their resting state. Flow cytometry analysis of surface antigen showed that 98% of the cultured cells expressed F4/80, a speci c marker of macrophages (Fig. 1a).
The reagents were used at the indicated concentrations.

Flow cytometry
BMMs were incubated in blocking solution (rat serum, 20 min at 4℃) and were subsequently stained with FITC-conjugated anti-F4/80 (1:100, AbD Serotec, Hercules, CA, USA) in the dark for 20 min at 4°C. Acquisitions were performed on a Millipore ow cytometer (Guawa 6HT). Subsequent data analyses were completed using FlowJo software version 7.6.2 (Tree Star, Ashland, OR, USA). Results are expressed as % of positive cells.

Immuno uorescence
Cells were cultured to ~70-80% con uence using glass coverslips in 24-well plates. After treatment, primary BMMs were xed with 4% paraformaldehyde in phosphate buffer saline (PBS, pH = 7.4) for 30 min and then washed thrice with PBS. Non-speci c antibody binding was blocked by incubating the cells for 1 h at room temperature (20-25°C) in PBS containing 5% bovine serum albumin and 0.3% Triton-100, following which the cells were stained with the following primary antibodies overnight at 4°C: mouse anti-AR Laboratories, West Grove, PA, USA). Diamidinophenylindole (DAPI, 1 μg/ml, Sigma-Aldrich) was applied to visualize the nuclei. Rhodamine-phalloidin (1:250, Life Technologies-Invitrogen, Carlsbad, USA) was used to stain the actin cytoskeleton of the cells. The cells were observed using a laser confocal microscope (FluoView FV1000 MPE, Olympus Corporation, Tokyo, Japan). We applied the JACoP plugin of Image J (National Institutes Health, Bethesda, MD, USA) to perform colocalization analysis of any image pair, and the colocalization rate of the green and red signals was evaluated using Manders' overlap coefficient [23].

Transmission electron microscopy
Cells were grown to ~80% con uence in 75 cm 2 culture asks. After stimulation, BMMs were harvested and centrifuged at 1,500 g for 10 min. The pellets were xed in a mixture of 4% paraformaldehyde and 0.05% glutaraldehyde, post-xed in 0.5% osmium tetroxide, dehydrated in graded ethanol series and propylene oxide, and embedded in epoxy resin. Ultrathin sections were obtained using an ultramicrotome (EM UC6, Leica Microsystems, Baden-Württemberg, Germany); they were then mounted on mesh grids (6-8 sections/grid) and counter-stained with uranyl acetate and lead citrate. The sections were viewed using a transmission electron microscope (JEM-1230, JEOL, Tokyo, Japan) equipped with a Gatan digital camera.

Statistical analysis
All data are expressed as the mean ± SD. Statistically significant differences between the mean values were determined by two-tailed Student's t-test. Differences at P < 0.05 were considered statistically signi cant. *P < 0.05; **P < 0.01.

LPS stimulation upregulates AR expression in BMMs
It has been reported that AR can regulate in ammatory signals and immune responses in several animal disease models [5][6][7][8]. Since in ammatory signals and immune responses are tightly correlated with macrophage polarization, we investigated changes in the expression of AR upon LPS stimulation in BMMs from WT C57BL/6 mice.
We used the immuno uorescence staining to label AR in BMMs. The number of AR-immunoreactive pro les in BMMs was signi cantly increased after treatment with LPS for 24 h (Fig. 1b). We next used qRT-PCR and western blotting to detect the changes of AR mRNA and protein in LPS-stimulated BMMs.
The transcription and expression of AR were both signi cantly increased following treatment for 24 h (Fig. 1c, d). Collectively, these results indicate that AR expression is triggered upon stimulation of BMMs with LPS.
AR de ciency suppresses the M1 response and NF-κB activation at the IKK complex level in LPSstimulated BMMs In response to LPS treatment, macrophages undergo M1 polarization, characterized by the expression of inducible nitric oxide synthase (iNOS) [25]. Activation of the canonical NF-κB signaling pathway in response to LPS was demonstrated by monitoring iNOS expression, through immunoblot analysis, and nitric oxide (NO) production-important downstream products of LPS signaling in macrophages [26].
We hypothesized the existence of a relationship between AR and the NF-κB p65 signaling pathway. To test this hypothesis, we carried out experiments on BMMs from WT or AR KO mice. Stimulation of BMMs with 500 ng/ml LPS increased the levels of iNOS in a time-dependent manner during the rst 32 h (Fig. 2a). Expression of iNOS was peaked at 16 h after LPS treatment (Fig. 2a). The induction of iNOS upon LPS stimulation decreased markedly when the AR gene was knocked out (Fig. 2a). These results suggested that the LPS-stimulated induction of iNOS was closely associated with AR.
For most inducers of the classical NF-κB signaling pathway, IKKα is not required for the phosphorylation of IκBα; however, genetic studies have shown that NF-κB p65 signaling requires IKKβ and IKKγ to phosphorylate IκBα [27]. The canonical NF-κB pathway uses several important components, including TLR4, IKKβ, IKKγ, IκBα, and p65 [28][29][30]. Following 16 h of LPS exposure, we detected lower levels of IKKβ, IKKγ, phospho-IκBα (p-IκBα), phospho-p65 (p-p65), and iNOS proteins in AR KO BMMs than in control cells (Fig. 2b). We used qRT-PCR to measure the mRNA levels of IKKβ and IKKγ in the total RNA obtained from the BMMs after 16 h of treatment with or without LPS. The results showed that their mRNA levels were not affected in the LPS-treated AR KO BMMs (Fig. 2c). The above data indicated that the suppression of the M1 response in AR-de cient macrophages after LPS stimulation was due to the posttranscriptional degradation of IKKβ and IKKγ. To determine the temporal patterns of change in the IKKβ and IKKγ protein levels, we stimulated BMMs from WT or AR KO mice with LPS for 32 h and immunoblotted the lysates for IKKβ and IKKγ. The time-course experiments revealed that LPS induced a gradual reduction of IKKβ and IKKγ proteins in BMMs from AR KO mice (Fig. 2d).

AR de ciency enhances autophagosome formation and maturation in LPS-stimulated BMMs
To determine whether autophagosome formation occurred because of AR de ciency, we treated BMMs from WT or AR KO mice with or without LPS. Beclin1 is a key component of the class III phosphatidylinositol 3-kinase (PI3K) complex, which initiates autophagosome formation [31]. Incubation of BMMs with 500 ng/ml LPS for 16 h led to the recruitment of beclin1, as shown in the change in beclin1 staining from a diffuse to punctate pattern (Fig. 3a). We then conducted a detailed quantitative analysis of LPS-stimulated beclin1 aggregation. LPS induced a signi cant increase in the number of beclin1 + pro les in AR KO cells (Fig. 3a). The increase in beclin1 levels was also con rmed by immunoblotting (Fig. 3b).
In macroautophagy, cells form double-membraned vesicles, known as autophagosomes, around a portion of the cytoplasm [32]. Two forms of LC3, LC3 and LC3 , are produced post-transcriptionally [33]. LC3I is cytosolic, whereas LC3 is membrane-bound and mainly enriched in the autophagosomal vacuoles [34]. The autophagic vacuoles can be imaged using confocal microscopy by virtue of immuno uorescence staining for LC3, which delineates them in the cytoplasm [35]. We then conducted a detailed quantitative analysis of LPS-stimulated autophagy. Without LPS treatment, AR KO macrophages exhibited more LC3 + vacuoles compared to WT cells (Fig. 3c). After LPS induction, there was a signi cant increase in the number of autophagosomes in AR KO BMMs (Fig. 3c). The increase in LC3 was con rmed by immunoblotting with an LC3 antibody (Fig. 3d). To observe the temporal pattern of autophagy within 32 h, we immunoblotted lysates from LPS-stimulated WT or AR KO BMMs for LC3. The time-course experiment showed that LPS treatment signi cantly increased LC3 protein abundance in AR KO cells compared with WT controls within 32 h (Fig. 3e).
As their maturation proceeds, autophagosomes fuse with lysosomes, resulting in the degradation of their contents [36]. LAMP1 is widely used as a lysosome marker. The merging of LC3 with LAMP1 indicates the presence of mature autophagosomes (autolysosomes). Immunoblotting showed that LAMP1 was signi cantly upregulated in LPS-stimulated AR KO cells (Fig. 3f). In AR KO BMMs, LPS promoted the maturation of autophagosomes as evidenced by an increase in the colocalization of LC3 with LAMP1 (Fig. 3g).
Transmission electron microscope was also used to examine the effects of LPS on autophagy in BMMs from WT or AR KO mice. The number of double-membrane vacuoles, a typical feature of autophagosomes, was markedly increased in non-stimulated AR KO macrophages compared with that in WT cells; however, the fusion of primary lysosome and autophagosome, a typical feature of autolysosomes, was rare in AR KO and WT macrophages without LPS stimulation (Fig. 3h). Furthermore, it was observed that LPS treatment induced more autophagosomes and autolysosomes in AR-de cient cells compared to controls (Fig. 3h).
Collectively, these results indicate that AR de ciency promotes autophagosome biogenesis in nonstimulated macrophages, and further strengthens biogenesis and maturation in LPS-activated BMMs.

AR de ciency mediates autophagic degradation of IKKβ and IKKγ in LPS-treated BMMs
The addition of the PI3K inhibitor 3-MA has been shown to block autophagosome formation [37]. Our results showed that 3-MA (3 mM) was able to block LPS-stimulated autophagy as detected by LC3 immuno uorescence (Fig. 4a) and immunoblotting (Fig. 4b). NH 4 Cl is a commonly used lysosome inhibitor. It localizes to acidic vesicles and impairs lysosomal acidi cation and protease activity, resulting in the accumulation of autophagosomes by impairing their fusion with lysosomes [38]. Incubation of the cells with 10 mM NH 4 Cl for 16 h resulted in the accumulation of LC3 + vacuoles (Fig. 4c) and a large increase in LC3 levels (Fig. 4d) in LPS-stimulated macrophages.
To investigate the mechanisms underlying the observed downregulation of key components of the IKK complex and enhanced autophagy, we analyzed the protein levels of IKKβ and IKKγ by immunoblotting in LPS-stimulated BMMs after inhibiting autophagy using 3-MA. Our results revealed that the protein levels of IKKβ and IKKγ were rescued following 3-MA administration in LPS-treated AR-de cient cells (Fig. 4c). The induction of iNOS in LPS-treated AR KO macrophages was also signi cantly increased after 3-MA treatment (Fig. 4c). This was further supported by our data showing that the protein levels of IKKβ and IKKγ were restored as re ected by the increased levels of iNOS in LPS-stimulated AR-de cient cells upon addition of NH 4 Cl (Fig. 4d). These results suggest that IKKβ and IKKγ were loaded into autophagosomes and subsequently degraded in lysosomes, which limited iNOS induction and M1 polarization in LPSactivated AR-de cient BMMs.

AR de ciency facilitates IKKβ and IKKγ to colocalize with autophagosomes and lysosomes in LPSstimulated BMMs
In the initial phase of autophagy, beclin1 serves as a platform to recruit autophagy activators and inhibitors [39]. We immunostained BMMs for IKKβ, IKKγ and beclin1 after 1 h of LPS incubation. We showed that more intracellular structures were labeled by IKKβ and IKKγ colocalized with beclin1 + foci in AR KO BMMs after exposure to LPS (Fig. 5a, b). These results indicate that in the initial induction of autophagy, IKKβ and IKKγ are recruited by beclin1.
To verify that IKKβ entered the autophagosomes, we used immunocytochemistry to label IKKβ and autophagic vacuoles. We observed autophagosome formation 1 h after treatment with LPS. Autophagosome induction was poor and few overlapping IKKβ-labeled structures and LC3-labeled puncta were observed in resting macrophages (Fig. 5c). Upon LPS stimulation, a higher number of IKKβ + structures colocalized with LC3 + vacuoles in AR KO macrophages compared to WT macrophages (Fig. 5c). Similarly, AR knockout increased LC3 staining and its colocalization with IKKγ in LPS-treated BMMs (Fig. 5d).
Next, we detected higher levels of colocalization between the lysosome marker LAMP1 and IKKβ and IKKγ in LPS-induced AR KO BMMs (Fig. 5e, f).
These results indicate that after recruitment by Beclin1, IKKβ and IKKγ can be engulfed by autophagosomes and probably be degraded after fusion with lysosomes.
AR de ciency, polyubiquitination and p62 co-mediate the autophagic degradation of IKKβ and IKKγ in LPS-stimulated BMMS Autophagic adaptors represent a mechanism through which intracellular targets are delivered to autophagosomes [40]. An adaptor protein, p62, recognizes polyubiquitinated targets and binds to the ubiquitin-like autophagosome membrane LC3 in the autophagic degradation pathway [41]. It has been shown that TLR4-mediated autophagy is a p62-dependent type of selective autophagy in macrophages [42]. K63-linked polyubiquitination has also been associated with the formation and autophagic degradation of protein inclusions [43]. Interestingly, the preferential binding of p62 to K63-linked chains constitute a further level of regulation during selective autophagy [44].
Confocal microscopy of BMMs immunostained for PloyUb (linkage-speci c K63) and IKKβ showed that in the absence of LPS stimulation, ubiquitin resided mainly in the cytosol where it appeared as small PolyUb + dots, whereas IKKβ showed an occasional granular pattern (Fig. 6a). Treatment of BMMs for 1 h with LPS resulted in the formation of large IKKβ aggregates colocalized with PolyUb (Fig. 6a). Confocal microscopy experiments further suggested that in stimulated WT cells, only a small fraction of IKKβ was colocalized with PolyUb + ; contrastingly, in LPS-treated AR KO cells, the overlaps between IKKβ and PolyUb + dots were markedly increased (Fig. 6a). Similarly, we found that LPS stimulation induced the recruitment of IKKγ to PolyUb + structures, and the colocalization of IKKγ and PolyUb was markedly increased in AR KO cells (Fig. 6b).
Next, we treated primary macrophages with or without LPS and immunostained the cells for p62 and IKKβ. We noted that only a small fraction of p62 was associated with IKKβ in non-stimulated and LPStreated WT cells, but this association was greatly increased in LPS-stimulated AR-de cient BMMs (Fig. 6c). Consistently, LPS-induced interaction between IKKγ and p62 was promoted by the knockout of AR in BMMs (Fig. 6d). These data indicated that IKKβ and IKKγ aggregates were polyubiquitinated and they could be targeted by p62 to the autophagy pathway in LPS-induced AR-de cient macrophages.

Discussion
Our results showed that AR de ciency led to the suppression of the M1 phenotype in macrophages, limiting the pro-in ammatory activity through the degradation of key component of the IKK complex. Knockout of AR triggered autophagosome formation; exposure of AR KO macrophages to LPS markedly enhanced autophagosome biogenesis and maturation. Autophagy was able to capture and degrade IKKβ and IKKγ through polyubiquitination, leading to the recruitment of p62 and LC3. Autophagosome formation in the AR knockout background was functionally important; pharmacological blockade of autophagy pathway increased the protein levels of IKKβ and IKKγ and enhanced the production of iNOS. These data suggested a close relationship among AR, the M1 response, and autophagy.
There are two main NF-κB pathways in the cell. The noncanonical NF-κB signaling pathway depends on IKKα and plays a critical role in the development of lymphoid organs responsible for the generation of B and T lymphocytes [45]. Neither IKKβ nor IKKγ de ciency affects this pathway [45]. However, the canonical pathway, induced by most physiological NF-κB stimuli, depends on IKKβ and IKKγ; the activation of this pathway mainly leads to the phosphorylation of IκBα and nuclear translocation of mostly p65-containing heterodimers [28,46]. However, IKKγ de ciency appears to have a greater impact on the classical NF-κB signaling pathway than IKKβ de ciency [45]. It has been reported that cells lacking IKKγ have no detectable NF-κB response to almost any pro-in ammatory or immune regulatory stimuli [47]. Our results showed that the exposure of AR KO macrophages to LPS impaired IKKβ and IKKγ. This may be attributed to the suppression of the TLR4/NF-κB signaling pathway activation, iNOS induction and M1 polarization in AR-de cient BMMs.
We investigated the mechanism through which certain signals downregulated IKKβ and IKKγ proteins without affecting their mRNA levels in AR KO macrophages after LPS treatment. We speculated that IKKβ and IKKγ protein levels were decreased at the post-transcriptional level in AR KO macrophages upon LPS stimulation. We found that the exposure of AR KO macrophages to LPS markedly enhanced autophagosome formation and maturation. Furthermore, we investigated the relationship between the impairment of the IKK complex and autophagosome formation. We observed massive colocalization of IKKβ, IKKγ, LC3 + autophagic vacuoles, and LAMP1 + lysosomes in BMMs from AR KO mice after LPS stimulation. Furthermore, our nding that AR de ciency led to the polyubiquitination of IKKβ and IKKγ provides a possible answer. The K63-ubiquitinated IKKβ and IKKγ can recruit p62, which can deliver them to the autophagy pathway through its LC3-binding domain. Conversely, pharmacological suppression of autophagy partially restored IKKβ and IKKγ protein levels and iNOS induction.
Thus, manipulation of AR expression or activity may have a therapeutic potential by regulating macrophage polarization for in ammatory diseases.