Regulation of Macrophage Cell Surface GAPDH Alters LL-37 Internalization and Downstream Effects in the Cell

Infectious diseases like tuberculosis (TB) have bedeviled humankind for centuries. According to the World Health Organization report, worldwide, 10 million new cases of TB with 1.5 million deaths occur annually. Efforts to control it are hindered by the rapid emergence of multiple drug-resistant strains resulting in relatively ineffectual treatment [1]. Due to the alarming increase in the number of drug-resistant cases, there is a shift in the treatment paradigm toward the prospect of modulating the host immune response as an adjunctive therapy. Host-directed therapy includes agents that enhance host immunity to maximize bacterial killing while reducing inflammatory tissue damage and enhancing the activity of the antibiotics [2]. Earlier, it has been demonstrated that vitamin 1,25-D3 [3] and phenylbutyrate activate the antibacterial autophagy mechanism resulting in the elimination of intracellular Mycobacterium tuberculosis (M.tb) in macrophages via induction of cathelicidin LL-37 peptide. LL-37 is the sole member of the human cathelicidin family of cationic host defense peptides (CHDP) which possess potent antimicrobial properties. It represents the cleaved C-terminal portion of the antimicrobial protein, hCAP-18 [4]. Released at the site of infection, by macrophages and neutrophils as natural host defense molecules [5], these CHDPs manifest their protective effect by initiating innate immune defense functions like autophagy, apoptosis, inflammasome activation, reduction of pro-inflammatory cytokine TNF-α, and upregulation of chemokines like MCP-1, involved in infection clearance [6‒9]. Studies demonstrating an increase in LL-37 expression in M.tb infected cell culture models and mice indicate it to be an important mechanism deployed by host cells to decrease the load of intracellular pathogens [10‒12]. Exogenous treatment with LL-37 selectively modulates host immune responses leading to the resolution of pathogen-induced inflammation while maintaining anti-infective immunity by upregulating autophagy [9, 13, 14]. In TB, autophagy has been associated with bactericidal and anti-inflammatory functions. It relies on the ability of the host macrophages to eliminate intracellular pathogens through maturation of mycobacterial phagosomes into phagolysosomes and inhibition of inflammatory responses [15]. Macrophage polarization also influences the anti-infective nature of macrophages with M1 type demonstrating enhanced LL-37 secretion in contrast to the M2 type [16].

Human macrophages internalize LL-37 peptide which then interacts with cytosolic molecules like glyceraldehyde-3-phosphate dehydrogenase (GAPDH) or p62/SQSTM1, to stimulate a variety of signal transduction pathways key to the innate immune response [17‒19]. The mechanism by which LL-37 exerts its function, by gaining access to and interacting with these cytosolic receptors, is not clearly understood. What is clear is that internalization of LL-37 into cells is a critical prerequisite for its multiple downstream functions. Certain downstream effects of LL-37 were previously demonstrated to occur after interaction with a variety of putative cell surface receptors such as formyl peptide receptor like-1 receptor, purinergic receptor 7 (P2X7), G protein-coupled receptors (GPCRs), and chemokine receptor 2. However, the receptor(s) through which LL-37 exerts its effects on macrophages remain unidentified, and the anti-inflammatory function of LL-37 has been shown to be independent of any of the above known surface receptors [20]. LL-37-mediated autophagy activation in M.tb was shown to involve P2X7 along with regulation of calcium influx into the cells [4]. However, it was shown that during M.tb infection in macrophages, the expression of surface P2X7 remains unaltered [9, 21, 22]. Also, LL-37 induces Ca²⁺ mobilization in cells by a process that involves P2X7. This regulates cell migration. Calcium regulation is thus one of the many signaling cascades activated by LL-37’s interaction with the surface of cell. It is important to identify all of the players on the plasma membrane responsible for internalization of LL-37 and subsequent cellular responses into cells during infection/inflammation. This is vital for understanding subsequent interactions of the peptides with intracellular effector molecules [23].

During initial stages of M.tb infection, host cell metabolic reprogramming occurs and the Warburg effect is induced, with upregulation of the glycolytic gene (GAPDH) and a shift toward glycolytic pathways and hypoxia to decrease intracellular bacillary load [24, 25]. This is a consequence of enhanced secretion of β-defensin-2 and activation of the VDR pathway [26]. The multifunctional glycolytic enzyme GAPDH performs numerous pleiotropic functions depending on its subcellular localization and posttranslational modifications [27, 28]. It also regulates autophagy, cell death, and intracellular membrane trafficking [29, 30]. Apart from being a cytosolic protein, GAPDH is known to be recruited onto the exterior of the plasma membrane, where it functions as a non-canonical receptor for endocytosis of transferrin and the antimicrobial protein lactoferrin [31, 32]. It is also secreted into the extracellular milieu for the same purpose [33, 34]. The anti-inflammatory function of exogenously added GAPDH has been previously reported for its therapeutic potential against lipopolysaccharide-induced sepsis [35]. In the present study, we have attempted to determine the process by which cathelicidin LL-37 is internalized into infected cells via its interaction with extracellular GAPDH. We hypothesized that enhancement of surface localization of GAPDH molecules should also result in increased LL-37 internalization and promote protective responses in cells. We were able to demonstrate the internalization and endocytic trafficking of GAPDH-LL-37 complex.

J774 (mouse macrophage cell line) was obtained from ECACC and cultured in DMEM high glucose with 10% fetal bovine serum. THP-1 cells were obtained from the National Centre for Cell Sciences (NCCS), Pune, India. Cells were cultured in RPMI-1640 supplemented with 10% FCS. Peritoneal macrophages were isolated exactly as described previously from C57BL/6 mice [34, 36]. THP-1 cells were activated for 24 h with 25 ng/mL of phorbol 12-myristate 13-acetate (PMA), followed by 24 h of rest before further stimulation or infection before use [37]. To determine the signaling pathway for LL-37 induction or its internalization, cells were treated with 100 nm KN-62 (1-(N,O-bis(5-Isoquinolinesulfonyl)-N-methyl-l-tyrosyl)-4-phenylpiperazine), a P2RX7 receptor inhibitor for 4 h (Calbiochem), p38 mitogen-activated protein kinase (MAPK) inhibitor SB0502880 (Sigma) for 1 h, cholesterol synthesis inhibitor mevinolin and methyl-β-cyclodextrin (to disrupt membrane rafts) (Sigma) for 4 h, bafilomycin 100 nm alone for 2 h or before treatment with LL-37 (AnaSpec #AS-61302) or scrambled LL-37 (AnaSpec #AS-63708).

M.tb H37Rv and M.tb H37Rv bacilli expressing GFP and mCherry have been described previously [38, 39]. The M.tb strains were cultured in Middlebrook 7H9 broth with 0.2% glycerol, 10% OADC, and 0.1% Tween 80. Fluorescent constructs were maintained under selection pressure of hygromycin (50 μg/mL). PMA-activated THP-1 cells were infected with either M.tb H37Rv or H37Rv expressing fluorescent markers at 1:5 multiplicity of infection. Subsequently, cells were allowed to phagocytose M.tb for 4 h, after which non-phagocytosed bacteria were removed by rigorous washing. This was followed by treatment with 50 μg/mL of gentamicin solution for 1 h to remove extracellular bacteria. The cells were incubated for 24 h in a humidified chamber (37°C, 5% CO₂) with complete media. A parallel set of uninfected cells was also maintained to serve as control. For in vivo infection experiments, C57BL/6 mice were injected i.p. with 1 × 10⁷ bacilli in 2 mL PBS. Forty-eight hours postinfection, the peritoneal macrophages were harvested and examined by flow cytometry to confirm the presence of intracellular bacilli.

For cellular GAPDH knockdown, J774 and THP-1 cells were transfected with either mouse or human GAPDH short hairpin RNA (shRNA) lentiviral particles (Sigma-Aldrich) as per the manufacturer’s instructions. As control, separate sets of cells were transfected with pLKO.1-puro non-target shRNA control lentiviral particles (Sigma-Aldrich). Stably transfected cells were selected and cultured in medium supplemented with puromycin (selection pressure of 7 μg/mL was utilized in the case of J774 cells and 12.5 μg/mL for THP-1 cells). Culture medium of GAPDH knockdown cells was supplemented with sodium pyruvate to compensate for energy production [40]. Knockdown was confirmed by Western blot as described below.

The rat GAPDH-EGFP/pcDNA3 vector and pmCherry-N1 vector were received as gifts from Prof. S. Sealfon, Mount Sinai School of Medicine, and Dr. J. Lippincott Schwartz, NIH, respectively. The GAPDH gene was cloned in pmCherry-N1 at HindIII/BamHI site to express GAPDH-mCherry fusion gene. The GAPDH-mCherry-N1 insert was cloned in retroviral pLNCX2 vector at HindIII/NotI site to create pLNCX2-GAPDH-mCherry plasmid. Similarly, mCherry-N1 insert was cloned at HindIII/NotI site to create pLNCX2-mCherry plasmid. These were then utilized for transfection of PT-67 cells and production of retroviral particles which were then used to transfect THP-cells as described previously (http://www.takara.co.kr/file/manual/pdf/PT3132-1.pdf). Stably transfected cells overexpressing either GAPDH-mCherry or only mCherry were selected and maintained using G418 (400 μg/mL).

Intracellular calcium levels were evaluated in PMA-activated THP-1 cells which were either kept untreated or incubated with LL-37 for 12 h. Subsequently, cells were washed and incubated with 5 mm fluo-3AM (Thermo Fisher Scientific) for 30 min at 37°C in PBS followed by incubation in PBS supplemented with 0.1% bovine serum albumin (BSA) for 30 min at 37°C. Cellular fluorescence was visually compared by confocal microscopy and quantified by flow cytometry. Intracellular ATP levels were quantified in THP-1 cells utilizing a BacTiter-Glo ATP assay kit (Promega #G8230). A crystal violet assay was used to normalize cell numbers.

GAPDH was measured in culture lysates of peritoneal macrophages infected intraperitoneally with M.tb H37Rv after 48 h. Cell lysates were filtered through a 0.22-μm filter and then assayed using a KDalert^TM GAPDH Assay Kit (Ambion) as per the manufacturer’s instructions. Controls were set in parallel wherein no infection was given. Results were normalized to protein concentration of the cell lysate and presented as GAPDH activity/mg protein of cell lysate ± SD. J774 cells were treated with 100 ng/mL of recombinant mouse IFNγ (PeproTech #315-05). After 24 h, GAPDH activity was measured as described above.

For co-localization studies of cell surface GAPDH and LL-37, 5 × 10⁵ J774 macrophages were cultured in confocal imaging dishes for 48 h. Cells were blocked with 2% BSA in SFM at 4°C for 30 min. Subsequently, cells were stained with rabbit anti-GAPDH (Millipore Sigma, Burlington, MA, USA) in fluorescence-activated cell sorting (FACS) buffer for 1 h at 4°C. After washing the cells extensively, cells were stained subsequently with anti-rabbit Alexa Fluor 647 (Thermo Fisher Scientific), and 10 μg LL-37-FAM (labeled peptide) was allowed to bind onto the cell surface for 45 min at 4°C. Finally, cells were washed thoroughly, fixed in 4% paraformaldehyde, and imaged with a Nikon A1R confocal microscope using a 60X oil immersion 1.49 NA objective with aperture set at 1 Airy unit for co-localization studies.

Interaction of LL-37 peptide and cell surface GAPDH internalized into endocytic compartment was confirmed by acceptor photobleaching förster resonance energy transfer (FRET) assay. For this, J774 cells were blocked with FACS block for 30 min at 4°C. Next, they were incubated with rabbit anti-GAPDH antibody (1 µg) in SFM for 1 h at 4°C and washed 3 times with SFM. Subsequently, cells were incubated with anti-rabbit Alexa 568 (acceptor) and 10 µg of LL-37-FAM (donor) (AnaSpec, USA) for 1 h at 4°C and then shifted to 37°C for 15 min. Cells were washed with chilled SFM and fixed with 4% paraformaldehyde. Control was also set in parallel as above where instead of LL-37-FAM, rabbit IgG-FITC (unrelated protein) was added. FRET experiments were carried out using the acceptor photobleaching method, and FRET efficiency was calculated using the following formula: FRET efficiency = (donor intensity [postbleach]-donor intensity [prebleach])/donor intensity (postbleach) [38].

The affinity of GAPDH-LL-37 interaction was determined using microscale thermophoresis (MST) as described previously [36]. Briefly, human GAPDH was labeled with fluorescent dye NT-647 using Monolith NT Protein Labeling Kit RED-NHS (NanoTemper Technologies, Munich, Germany) as per the manufacturer’s instructions. The concentration of the fluorescent-labeled GAPDH was kept constant, and the concentration of the titrant, i.e., LL-37, was varied from 10 µm to 10 pm. Serial dilutions of LL-37 were mixed with labeled GAPDH in a 1:1 ratio, and the mixture was loaded into Monolith NT.115 capillaries. The analysis was carried out using a Monolith NT.115 instrument (NanoTemper Technologies). The human cathelicidin LL-37 (LLGDFFRKSKEKIGKEFKRIVQRIKDFLRNLVPRTES) and labeled LL-37-FAM (5-FAM-LC-LLGDFFRKSKEKIGKEFKRIVQRIKDFLRNLVPRTES) were purchased from AnaSpec (#AS-61302 and #AS-63694, respectively). LL-37-scrambled peptide (GLKLRFEFSKIKGEFLKTPEVRFRDIKLKDNRISVQR) was also purchased from AnaSpec (#AS-63708). Mouse cathelicidin-related antimicrobial peptide (mCRAMP, GLLRKGGEKIGEKLKKIGQKIKNFFQKLVPQPEQ, # crb1000256h) was obtained from DISCOVERY Peptides.

THP-1 and J774 macrophage cells were either infected with M.tb H37Rv and/or were treated with IFNγ in the presence of various inhibitors as described above. At given time points, cells were harvested and aliquots of 5 × 10⁵ cells were washed with PBS. Cells were either processed for surface binding (at 4°C) or internalization (at 37°C) of 10 μg LL-37-FAM (AnaSpec #AS-63694) for 30 min. Cells processed for internalization were additionally treated with 0.1% pronase at 4°C to remove any residual surface-bound molecules. Cells were fixed using 4% paraformaldehyde or resuspended in PBS and analyzed by flow cytometry. P2X7 and GAPDH surface expression levels were evaluated by flow cytometry exactly as described previously [27]. Briefly, 5 × 10⁵ cells were washed with FACS buffer (20 mm HEPES [pH 7.4], 150 mm NaCl, 1 mm CaCl₂, 1 mm MgCl₂, 5 mm KCl supplemented with 5% FCS) and blocked with FACS block (FACS buffer further supplemented with 5% normal goat serum and normal human serum). Subsequently, cells were stained with anti-P2X7 antibody (#SCBT-SC-514962) or rabbit anti-GAPDH (Sigma #G9545) or isotype control (rabbit IgG, Invitrogen #10500 C) antibody. The antibodies were subsequently detected by either anti-mouse Alexa Fluor 647 (Invitrogen # A28181) or anti-rabbit Alexa Fluor 647 (Invitrogen #A-21244) and analyzed using FACSVerse flow cytometer (BD Biosciences). Signals from dead cells were excluded by gating out the cells that stained positive with 7-aminoactinomycin D. Gating strategy is included in online supplementary Figures (for all online suppl. material, see https://doi.org/10.1159/000530083).

PMA-activated THP-1 cells were infected with M.tb H37Rv (multiplicity of infection: 1:5) and treated with exogenous LL-37 (50 μg/mL). Infected cells were then analyzed for bacterial survival assay. At 0, 24, and 48 h post-addition of LL-37, macrophages were lysed with 0.05% SDS, and serial dilutions were plated onto Middlebrook 7H11 agar (Becton Dickinson, Difco, 212203 Difco). Colony-forming unit (CFU) counts were determined after 3 weeks of incubation at 37°C.

Following the treatments with either LL-37 peptide or the various inhibitors described above, cells were washed. and their lysates were prepared in 100 μL of RIPA cell lysis buffer (Gbiosciences) supplemented with protease inhibitor cocktail (Calbiochem). Protein concentration was estimated by the Bradford method. Cell membrane fractions, whole cell lysate, or Co-IP samples were resolved using 10% SDS-PAGE, transblotted onto nitrocellulose membrane, and blocked using 2% BSA. Blots were then probed with primary antibodies at 4°C overnight. Bound antibodies were then detected using anti-mouse (Sigma A4416) or anti-rabbit peroxidase antibody (Sigma A6154) and developed using LuminataTM Forte Western HRP Substrate (Merck Millipore WBLUF0500). As a loading control, anti β-actin (Sigma A2228) was used for cell lysate and lamin B1 (Abcam 133741) for isolated organelle samples. Early endosomes, late endosomes, and lysosomes were identified using anti-Rab5, Rab7, and LAMP-1 (Abcam #ab13253, ab137029, ab24170) respectively. For analysis of autophagosome machinery components, ATG5 (Cell Signaling Technology #2630S), LC3 (Sigma #L7543), p62 (BD Biosciences #610833) antibodies were utilized. For preparation of membrane fraction, J774 cells were allowed to bind LL-37 (50 μg/mL) at 4°C. Cells were subsequently washed with PBS containing 2% BSA and resuspended in 1 mL of membrane homogenization buffer. Membrane fractions were prepared as described previously [41]. Briefly, cells were homogenized by passing through a 26G needle, and the nuclear fraction was then separated by centrifugation at 800 g for 10 min. Postnuclear supernatant was centrifuged at 100,000 g for 1 h at 4°C and the resultant pellet (membrane fraction) was dissolved in PBS containing 2% triton-X-100 for further experiments.

Endosomal fractions were prepared by the sucrose density gradient method as described previously [42]. Briefly, after LL-37 internalization, cells were washed twice with SFM and resuspended in 1 mL of homogenization buffer for endosome fraction preparation. Postnuclear supernatant was prepared and loaded onto sucrose gradients of 54%, 40%, and 30% and centrifuged at 100,000 g for 1 h at 40°C. The buffy coat between supernatant and 40% sucrose layers was collected as endosomes. The fractions were washed in PBS at 100,000 g and dissolved in 2% lysis buffer for further experiments [43]. Co-IP was carried out as described below.

Membrane fraction from LL-37-bound J774 cells and endosomes from either LL-37 or scrambled peptide internalized cells were prepared as described above. From these fractions, LL-37 was immunoprecipitated using monoclonal rabbit anti-LL-37 antibody (AnaSpec) that had been previously immobilized onto anti-rabbit IgG-magnabeads (Pierce). For negative control isotype, control rabbit IgG antibody was immobilized on anti-rabbit IgG magnabeads. All Co-IP samples were boiled in SDS sample buffer, and the bound proteins were resolved on 12% SDS-PAGE, transferred to nitrocellulose membrane, and probed with either anti-GAPDH antibody of anti-LL-37 antibody. As control, similar set of experiments was performed where control cells which were allowed to internalize scrambled peptides were used.

For confocal microscopy, THP-1 GAPDH knockdown and empty vector macrophages were PMA activated. The cells were allowed to adhere for 12 h at 37°C and were infected with M.tb H37Rv-GFP as described above. Subsequently, cells were treated with LL-37 for 24 h in complete media. Following treatment, 1 h prior to the termination of the experiment, complete media containing LysoTracker dye at a concentration of 200 nm was added to the cells for staining of acidified lysosomes. Cells were fixed with 4% paraformaldehyde (Sigma) for 20 min, followed by 3 washes with PBS. DAPI (Invitrogen) staining was done at room temperature for 20 min to stain the nucleus of cells before imaging in a Nikon A1R confocal microscope.

24 h postinfection with M.tb H37Rv, total cellular mRNA of the THP-1 cells was isolated using TRIzol Reagent (Life Technologies) as per the manufacturer’s instruction. Synthesis of cDNA was carried out using RevertAid First Strand cDNA Synthesis Kit (Thermo Scientific) as per the manufacturer’s instruction. For real-time PCR analysis, the reaction mixture containing cDNA template, primers, and Maxima SYBR Green qPCR Master Mix (Thermo Scientific) was run in ABI Fast Real-time PCR System (Applied Biosystems, Foster City, CA, USA). Fold changes in mRNA levels were determined after normalization to internal control β-actin levels and analyzed using the 2C (T) method. Primers utilized are listed in online supplementary Table S1.

Nitric oxide concentration was estimated in culture supernatants by using the Griess reagent [44]. Then, 50 μL of the supernatants were incubated with 50 μL/well of Griess reagent (1.0% sulfanilamide, 0.1% naphthalenediamine dihydrochloride, and 2.5% H₃PO₄) in a 96-well assay plate at room temperature for 15 min. The absorbance at 550 nm was recorded using a microplate reader (BioTek Instruments, Inc.). The absorbance was changed to micromolar (μM) NO by a standard curve to estimate known concentrations of NaNO₂.

All experiments were repeated at least three times and statistical analysis was performed using unpaired Student’s t test. All images were analyzed using ImageJ software and graphs were plotted using GraphPad Prism software.

Antimicrobial peptides, like LL-37, are released at sites of infection-induced inflammation. Upon internalization into macrophages via cell surface receptors, they kill intracellular pathogenic bacilli like M.tb and also regulate the inflammatory milieu. P2X7 on macrophage surface has been shown to mediate the actions of LL-37. However, its role in M.tb infection has remained unclear. Treatment with agonists does not cause any inhibition of LL-37-induced anti-mycobacterial activity; moreover, M.tb infection does not alter P2X7 levels on cells [9]. In agreement with these earlier reports, we too found that, in M.tb infected cells, the levels of surface P2X7 remain unaltered (Fig. 1a). Though internalization of LL-37 into macrophages has been reported to be independent of the P2X7 receptor, the production of anti-inflammatory cytokines has been found to be dependent on its interaction with cytosolic GAPDH [20]. Previous work from our group had revealed that, in response to diverse cellular stresses, macrophages and other cells recruit GAPDH onto their surface to function as a receptor for diverse ligands [27, 45]. We, therefore, explored the role of this macrophage surface-recruited form of GAPDH as a putative receptor for LL-37. A significant increase in total GAPDH activity (∼50%) and enhanced surface recruitment of GAPDH (∼45%) was observed in peritoneal macrophages harvested from M.tb H37Rv infected mice (Fig. 1b, c). This correlated with an increase in cell surface LL-37 binding and uptake by cells (Fig. 1d, e). To further confirm the role of GAPDH in binding and internalization of LL-37, we genetically modified GAPDH levels in J774 target cell line using ShRNA-mediated knockdown. After confirming a significant decrease of total (∼70%) as well as cell surface (∼50%) GAPDH expression with no depletion of cellular ATP levels (online suppl. Fig. S1A–C; Fig. 1f), we evaluated the binding and uptake of LL-37 in these cells. A significant decrease in LL-37 binding and internalization was observed in the GAPDH K/D cells (Fig. 1g, h). Having confirmed the role of GAPDH in LL-37 binding and uptake, we then evaluated these responses in cells pretreated with IFNγ, a cytokine which is known to modulate cellular metabolism and also initiate antimicrobial activity in macrophages [46]. Treatment of cells with IFNγ for 24 h resulted in enhanced GAPDH activity and ∼50% increase in cell surface GAPDH localization (Fig. 1i, j). This correlated with a significant increase in LL-37 binding and uptake by these cells (Fig. 1k, l). Kinetic analysis of LL-37 internalization revealed a time-dependent increase in peptide uptake which failed to occur in cells where GAPDH had been knocked down (online suppl. Fig. S1D). No necrosis or apoptosis was noticed as a consequence of the knockdown, and in our previously published work, we have shown that the induction of apoptosis in GAPDH KD and wild-type cells is comparable when cells are induced with actinomycin D [47]. We also confirmed that ShRNA-mediated knockdown of GAPDH does not affect the GAPDH KD cell’s ability to respond to external stimulus. Both empty vector and GAPDH KD THP-1 cells could equally well carry out endocytosis of BSA and also uniformly responded to lipopolysaccharide treatment by upregulation of TLR-4 membrane expression levels (online suppl. Fig. S1E, F).

To investigate the receptor responsible for LL-37 internalization, the previously identified receptor for LL-37 was revisited. As shown previously, treatment with IFNγ enhanced GAPDH recruitment on the cell surface of WT cells but failed to do so in GAPDH K/D cells. However, surface expression of P2X7 continued to be elevated irrespective of GAPDH K/D (Fig. 2a, b). In contrast, LL-37 uptake did not correlate with P2X7 regulation, as treatment of GAPDH K/D cells with IFNγ failed to demonstrate any significant increase in LL-37 uptake (Fig. 2c). The latter data suggest P2X7-independent internalization of LL-37. To determine if a feedback loop exists between LL-37 and GAPDH for activating its internalization, we examined the importance of Ca++ which had previously been implicated in the LL-37-induced response. On the same line, GAPDH recruitment on the cell surface is a Ca++-dependent process [43]. To explore if LL-37 induces a calcium-mediated recruitment of GAPDH to the plasma membrane surface to facilitate its internalization, we first checked GAPDH expression on surface of cells treated with the peptide and found a marked rise in intracellular Ca++ which was absent in case of GAPDH KD cells (Fig. 2d, e). Next, we treated cells with KN-62, a tyrosine derivative which is known to inhibit Ca++/calmodulin-dependent kinase type II and exhibit selective P2X7 receptor-blocking properties [48, 49]. Treatment of cells with KN-62 did lead to an inhibition of LL-37 uptake into cells; however, this was also accompanied with a decrease in cell surface GAPDH expression as well as a significant fall in intracellular Ca++ levels (Fig. 2f–h). These results indicate that the internalization of LL-37 peptide into cells occurs via the surface GAPDH receptor which is recruited to the surface by P2X7-induced mobilization of Ca++.

Cationic amphipathic peptides like LL-37 are known for their membrane binding activity, and their internalization into cells has been observed to occur via receptors in lipid raft regions of the plasma membrane [50]. Previous work has demonstrated the recruitment of GAPDH into detergent soluble as well as lipid-raft fractions of the plasma membrane [27, 31]. Confocal microscopy demonstrated co-localization of LL-37 with GAPDH in the plasma membrane (Fig. 2i). A Co-IP assay confirmed their interaction (Fig. 2j). To ascertain if GAPDH that is present in lipid raft regions on the cell surface facilitates LL-37 binding, the lipid rafts were disrupted utilizing mevinolin and methyl-β-cyclodextrin. Disruption was confirmed by a significant decrease in cholera toxin (a lipid raft marker) binding to cells (Fig. 2k). We observed a depletion of both GAPDH and LL-37 signals on treated cells as compared to control untreated cells (Fig. 2l). Flow cytometry analysis of raft disrupted cells shows a significant ∼2-fold decrease in surface GAPDH levels and concomitant ∼2-fold decrease in FAM-LL-37 association with membrane and its subsequent internalization which correlated with a significant decrease in LL-37 binding and internalization when cells were pretreated with lipid raft disrupting agents (Fig. 2m–o). In comparison to control cells, subsets of cells treated with IFNγ, also show initial higher accumulation of FAM-LL-37 at their membrane, followed by increase in LL-37 uptake into the cytoplasm. Taken together, the data confirm that LL-37 is internalized inside the cells via its association with GAPDH. This association and subsequent internalization are modulated by changes in the inflammatory milieu, like increased secretion of IFNγ or during cellular infection with bacteria.

Confocal microscopy revealed that the GAPDH-LL-37 complex traffics into intracellular vesicles. We observed co-localization of both peptide and GAPDH in the same endocytic punctae (Fig. 3a). Interaction between internalized LL-37 and the co-localized GAPDH in endosomes was confirmed by FRET (Fig. 3b, c) and co-immunoprecipitation assays (Fig. 3d, SIG). Affinity of their interaction was evaluated using MST analysis. As negative control, a scrambled peptide with altered residues which does not show downstream effects like the LL-37 peptide was used. KD was determined to be 315 ± 40 nm (Fig. 3e, inset), and we also utilized the mCRAMP, as an additional interaction pair and found an interaction with nM affinity for GAPDH (Kd 472 nm) (Fig. 3f).

Metabolic polarization status of macrophages determines their ability to restrict the intracellular M.tb burden. M1 macrophages with a pro-inflammatory phenotype show enhanced glycolytic activity as compared to M2 macrophage which present with an anti-inflammatory signature. Though M1 macrophages are better equipped to control bacterial burden, an excessive inflammatory milieu is detrimental to the host organism [51]. Therefore, a proper balance of pro- and anti-inflammatory cytokines has to be maintained. Exogenous LL-37 decreases TNF-α and IL-17 pro-inflammatory cytokines while inducing IL-10, transforming growth factor β and chemokines like MCP-1. This aids in clearance of infection and suppression of excessive inflammation [9, 52]. Considering the previous results, we next sought to determine if internalization of LL-37 is regulated by the polarization state of macrophages. For this, we polarized J774 macrophages into M1 and M2 phenotypes as described previously [53]. Polarization was confirmed by evaluation of specific cell surface markers and NO release (online suppl. Fig. S1H–L). We observed that M1 macrophages presented ∼2-fold elevated expression of GAPDH as compared to the M2 cell type which showed a significant ∼6-fold reduction in GAPDH levels on their surface (online suppl. Fig. S1M). Correspondingly, M1 macrophages also demonstrated significantly enhanced LL-37 binding and internalization (online suppl. Fig. S1N, O). These results also strongly imply a role for GAPDH in regulating cellular phenotypic immune responses of LL-37 in infected cells.

Recently, the upregulation of GAPDH has been shown to have functional significance in host responses against invading bacteria via non-glycolytic functions like redox sensing and autophagy [54, 55]. We utilized GAPDH K/D cells to determine if interaction between GAPDH and LL-37 has a role for induction of autophagy in M.tb infected cells. LCB-II lipidation is the principal biomarker to measure autophagic flux. Western blot assays demonstrated that LL-37 treatment of infected GAPDH K/D cells failed to bring about any increase in the conversion of LC3-I to LC3-II lipidation (Fig. 4a, b). Treatment of cells with LL-37 in the presence of bafilomycin A1 increased LC3-I to LC3-II conversion when compared to bafilomycin A1 alone in empty vector control cells but not in GAPDH K/D cells. In addition, LL-37 induced a decrease in p62/SQSTM1 (an ubiquitin-binding protein and an indicator of autophagic degradation), demonstrating the activation of autophagy in cells. Cells with GAPDH knocked down were unable to show this decrease in p62 after treatment with LL-37 (Fig. 4c, d). Corresponding to this, we also observed LL-37-induced upregulation of autophagy genes ATG5 and ATG7 at the mRNA level. However, no regulation of autophagy genes was observed in GAPDH K/D cells treated with LL-37 (Fig. 4e, f). Further investigating the role of GAPDH in LL-37-induced autophagy, we explored whether LL-37 modulates M.tb trafficking to lysosomes and affects their intracellular survival. We observed an enhanced intracellular localization of M.tb H37Rv-GFP with LysoTracker^® Red stained acidic compartments in empty vector infected cells as compared to cells where GAPDH had been knocked down (Fig. 4g, h). Localization of M.tb into lysosomes is associated with decrease in their intracellular survival. Accordingly, we observed a significant decrease in cellular M.tb CFU load of empty vector-infected cells treated with LL-37 as compared to GAPDH K/D cells (Fig. 4i). These observations suggest that multifunctional GAPDH is a key player in the LL-37-mediated induction of autophagy. To verify that these results are primarily due to LL-37 interaction with GAPDH, we overexpressed GAPDH in cells and observed that LL-37 was able to further enhance autophagy and decrease M.tb survival (online suppl. Fig. S2A–I).

Earlier reports indicate a downregulation of cellular p38 MAP kinase pathways upon M.tb infection [56]. In agreement with previous reports, we also observed that LL-37 induces p-p38 MAPK phosphorylation in infected cells; on the other hand, this effect was abrogated in GAPDH K/D cells, which ultimately stalled its autophagy functions resulting in compromised clearing of the intracellular pathogen (Fig. 4j–l). These findings demonstrate a role for GAPDH in LL-37-mediated activation of the p38 MAPK pathway and control of M.tb survival in infected cells.

Upon infection, immune cells respond to cues from both the host and pathogen, maintaining a balance between pro- and anti-inflammatory signals through a continuous feedback [57]. Pro-inflammatory cytokines secreted by the host activate macrophages for clearance of pathogen, but this can also induce excessive inflammation, which is detrimental to the host. Host defense peptide, LL-37 (cathelicidin), is a primary effector component of the innate immune system that facilitates the clearance of infection and resolution of inflammation [8].

Intracellular uptake of LL-37 is important for its biological functions [58]. It has been shown that it interacts with receptors on the cell surface, such as formyl peptide receptor 2 (FPR2), which belongs to a class of GPCRs and mediates the chemotaxis of immune cells [51, 59]. The utilization of GPCR by cathelicidin has not been observed in many studies, implicating alternative mechanisms and receptors being involved in its internalization. However, for its anti-inflammatory functions, LL-37 was shown to get internalized into the cytoplasm and interact with intracellular protein partners like GAPDH and p62, which were shown to contribute to LL-37-mediated immune responses by modulation of nuclear factor-κB (NF-κB), p38, and JNK MAPK, MKP1, and phosphoinositide 3-kinase signaling pathways. Interaction of the peptide with macrophages leads to enhanced killing and clearance of invading bacteria [60, 61]. Internalization of peptide by host cells is crucial to exerting these downstream responses [23]. Network analysis has revealed a complex interaction between LL-37 and at least 16 different proteins and receptors [62].

Previous studies by our group have shown that GAPDH, which is an intracellular receptor for LL-37, is also recruited on the externalized plasma membrane surface in response to cellular stress like hypoxia and infection, where it moonlights as a receptor for various ligands [36, 43, 45, 47, 63]. With this perspective, in the present study, we provided evidence that cell surface-associated GAPDH functions as a putative receptor for LL-37 internalization. Downstream effects of LL-37 in the cell were shown to be dependent on the cellular milieu that affects its interaction with specific interacting protein partners or receptors to influence its downstream effects [20, 64]. We utilized M.tb as a cellular macrophage infection model and Th1 cytokine, IFNγ treatment to induce classical macrophage activation. In both model systems, we observed that infection or IFNγ treatment altered GAPDH surface levels. This resulted in enhanced binding and internalization of LL-37 into macrophages. Previous studies have also shown that M.tb infection in macrophages augments host defense by an increase in glycolysis and upregulation of GAPDH [65]. These studies were confirmed using shRNA-mediated GAPDH knockdown in cells which not only decreased the cytosolic pool of GAPDH but also led to a (>50%) reduction in cell surface recruited-form of GAPDH resulting in the reduction in LL-37 binding and uptake into cells. No loss of viability was observed in GAPDH knockdown cells. The addition of pyruvate in the medium compensates for the glycolytic functions and ATP levels of GAPDH KD cells [66]. Substantial evidence exists that LL-37 reprograms the immune cells by maintaining both homeostasis and a delicate balance of inflammatory responses [67]. Thus, the LL-37 peptide has both anti-inflammatory and pro-inflammatory effect-inducing capabilities and is, therefore, regarded as a multifunctional immunomodulatory peptide. LL-37-mediated pro-inflammatory activity but not anti-inflammatory activity mediated by GPCRs in THP-1 macrophages has been reported [59]. LL-37 has also been demonstrated to function in association with the P2X7. However, treatment of cells with KN-62, an agonist for the P2X7 does not inhibit LL-37-induced anti-mycobacterial effects [9]. Rather, IFNγ-dependent upregulation of P2X7 [68] showed no consecutive increase in LL-37 uptake in GAPDH KD cells. It is thus likely that LL-37 uptake is not synergistic to P2X7 expression. In agreement with earlier investigators [23], we too observed a decrease in LL-37 internalization on treatment with KN-62, but it was accompanied by a decrease in surface GAPDH which could be due to a decrease in intracellular Ca++ levels. As previously demonstrated by our previous work, GAPDH recruitment on the cell surface is a Ca++-dependent process [43]. This suggests a cross-talk and the existence of feedback between multiple receptors utilized by LL-37 for its downstream function, which will be delineated in our future studies.

LL-37 has the potential to affect multiple signaling events via its interaction with plasma membrane domains of receptors. Several mechanistic studies have demonstrated the mechanisms for LL-37 intracellular uptake, endocytic mobilization, and interaction with several receptors for mediating its downstream effects [69]. The presence of cholesterol-rich regions on the mammalian cell membrane is one of the factors involved in LL-37 internalization [70]. Interestingly, many reports indicate that endocytosis of LL-37 may originate at lipid raft regions [71]. Consistent with this, we found that GAPDH-dependent LL-37 internalization into the cells occurs via lipid rafts and cholesterol-rich regions of the membrane which provide a favorable microenvironment for the attachment and uptake of LL-37. We demonstrated that cell surface-localized GAPDH interacts with LL-37 on the membrane. Once internalized into infected cells and cells expressing the inflammatory phenotype, LL-37 has been shown to localize and traffic into endosomes and lysosomes and cross the cellular membranes to initiate protective immune and antimicrobial responses. In our studies, we too observed that the GAPDH-LL-37 complex traffics into the same endocytic compartments upon being internalized in cells.

Under non-inflammatory conditions, LL-37 is expressed at high levels in the lung (1-10 μg/mL). This rises to 20 μg/mL during inflammation, wherein it reduces TNF-α levels by NF-Kβ(P50/65) transcriptional regulation. This ability to dampen inflammatory conditions and enhance the resolution of inflammation by regulating macrophage signaling is the immunoregulatory property of this peptide [13]. The anti-inflammatory properties of LL-37 were identified through its antagonistic actions against IFNγ, TNF-α, IL-4, and IL-12 responses in several types of cells [72]. During the early stage of microbial infection, it was shown to upregulate chemokines like MCP-1, involved in infection clearance, thus maintaining cellular homeostasis [6, 7]. When added exogenously to the infected cell, LL-37 suppressed excess inflammation [9]. To understand this regulation of differential actions of LL-37 in different immunological environments, we explored the LL-37 internalization kinetics in M1-classical/M2-alternatively activated polarized subsets. We observed that IFNγ-polarized M1 macrophages show enhanced binding and cellular uptake of LL-37 when compared to the IL-4 polarized M2 subset. This is corroborated by the higher expression of GAPDH on the cell surface of M1 macrophages. Thus, this GAPDH-selective internalization of the peptide to resolve inflammatory conditions in the M1 subset is a unique property of the peptide.

A potential criticism of our study could be the use of a human peptide (LL-37) to stimulate mouse cells; however, earlier, we confirmed our current findings in human primary monocyte-derived macrophages (online suppl. Fig. S2H, I). We had also analyzed the internalization of mCRAMP (the mouse homolog of LL-37) in mouse origin J774 GAPDH KD cells. We observed that mCRAMP is internalized in J774 empty vector control cells. However, GAPDH KD cells show a significant decrease in uptake (online suppl. Fig. S2J). This is similar to the results as observed when LL-37-FAM was utilized. Our MST data also point toward the capacity of GAPDH in being able to bind both human and mouse orthologs of cathelicidin host defense peptide. As GAPDH is a highly conserved molecule, we speculate that this may be an evolutionarily conserved function.

Once internalized, LL-37 is known to regulate the innate immune functions via induction of autophagy and elimination of intracellular pathogens [3, 4, 9, 23, 51, 73]. Consistent with this, we demonstrate the functional significance of GAPDH-LL-37 association in the biological process of autophagy during M.tb infection. The observed increase in LC3-II/LC3-I ratio showed that LL-37 treatment could induce autophagy in infected cells. However, this effect was abolished in GAPDH knockdown cells. qRT-PCR assays also showed that ATG5 and ATG7 genes involved in autophagosome biogenesis and LC3-I lipidation were significantly upregulated in LL-37-treated cells. This was corroborated by the observed enhanced co-localization of H37Rv-GFP bacteria with LysoTracker-stained acidic compartments and a decrease in CFU load upon LL-37 treatment. However, GAPDH knockdown failed to induce these LL-37-mediated effects. Further, these effects were reversed by overexpressing GAPDH. Importantly, we observed that LL-37-mediated autophagy was dependent on the p38 MAP kinase pathway, and GAPDH KD cells were unable to show upregulation of phosphorylation upon LL-37 treatment. Our observations are in agreement with the earlier study, where GAPDH knockdown completely abolished MAPK p38 phosphorylation in response to LL-37 [20]. Previous studies have shown the Ca++ mobilization pathways are essential for the LL-37-induced autophagy process [4]. We found that GAPDH KD cells do not show an increase in calcium flux upon LL-37 treatment. However, whether abrogated calcium signaling is the reason that GAPDH KD cells fail to show LL-37-dependent autophagy is a question that needs to be explored in future studies. In regards to autophagy, it has been demonstrated that the AMPK-dependent GAPDH phosphorylation is essential for SIRT1 activation and stimulation of autophagy. Inside the nucleus, GAPDH directly interacts with SIRT1, displacing SIRT1’s repressor and increasing SIRT1 deacetylase activity [54]. Therefore, it should be possible to address in future studies whether LL-37 affects GAPDH-SIRT1-dependent autophagy pathways.

Our current results expand the role of GAPDH as a mononuclear cell surface receptor for human cathelicidin LL-37 and demonstrate the significance of GAPDH in the functioning of the CHDP in M.tb infected macrophages. Based on our overall findings, we propose a model suggesting that M.tb-infection alters host glycolytic metabolism resulting in GAPDH exposure on cell surface that causes an increase in binding and uptake of peptide. This further initiates a downstream signaling pathway which involves peptide interaction with GAPDH facilitating mycobacterial phagosomes to fuse with lysosomes, thus ensuring protection from pathogens.

Considering the success of LL-37 in modulating host responses and bacterial killing, many synthetic analogs of LL-37 and other CHDPs are under investigation for future consideration as host-directed therapies for multidrug-resistant bacterial strains. Thus, understanding the steps of LL-37 targeting into host cells would allow examination of the efficiencies of LL-37 variants to mediate downstream effects. The present study thus reveals a strategy by which LL-37 interacts with host cell surface GAPDH as its cell surface receptor and is internalized into the cell to perform downstream antimicrobial functions. Our observation that significantly more GAPDH traffics to the cell surface upon infection and in response to IFNγ could also perhaps be exploited for enhanced delivery of peptide into infected cells. However, further detailed studies in this area would require to be conducted.

Mr. Anil Theophilus and Mr. Randeep Sharma are thanked for technical assistance. This is IMTECH communication No. 008/2022.

All animal handling protocols for obtaining primary cells from mice were as per the Institute of Microbial Technology Experimental Animal Facility protocols, and the experiments/procedures were approved by the Statutory Institutional Animal Ethics Committee of the Institute of Microbial Technology (Reference No. IAEC 13/15 and IAEC/17/02). Blood was obtained from healthy donors with written informed consent. Approval for the study was given by the Statutory Institutional Ethics Committee, project IEC aug2017#8.

The authors have no conflicts of interest to declare.

Financial support of DST, DBT, CSIR, ICMR, and UGC by way of research grants and fellowships is acknowledged.

Chaaya Iyengar Raje and Manoj Raje conceptualized and planned the overall research work and finalized the manuscript. Asmita Dhiman conceived and designed the study, collected the data, performed analysis, and wrote the initial draft of paper. Sharmila Talukdar performed some of the Western blots and RT-PCR experiments and collected and performed analysis of the data. Gaurav Kumar Chaubey helped with M.tb animal experiments and performed and analyzed nitric oxide estimation. Rahul Dilawari and Radheshyam Modanwal maintained the cell lines and assisted in tissue culture and MST analysis experiment. Surbhi Chaudhary and Anil Patidar assisted in various cell culture and microscopy-related experiments. Vishant Boradia, Pradeep Kumbhar, and Chaaya Iyengar Raje carried out the cloning of GAPDH-mCherry vector and created the THP-1 cell line overexpressing the gene.

All data generated or analyzed during this study are included in this published article and its online supplementary material. Further inquiries can be directed to the corresponding author.