Abstract

   Cadmium is a non-essential element and neurotoxin that causes
   neuroinflammation, which leads to neurodegenerative diseases and brain
   cancer. To date, there are no specific or effective therapeutic agents
   to control inflammation and alleviate cadmium-induced progressive
   destruction of brain cells. Fluoroquinolones (FQs), widely used
   antimicrobials with effective blood-brain barrier penetration, show
   promise in being repurposed as anti-inflammatory drugs. Therefore, we
   aimed to test the efficacy of repurposed FQs for the treatment of
   cadmium-induced inflammation using cultures of U-87 MG human astrocytes
   and primary human astrocytes. Both FQs abrogated cadmium-induced
   interleukin (IL)-6 and IL-8 release from human astrocytes in a
   concentration and time-dependent manner, although levofloxacin had a
   stronger inhibitory effect than moxifloxacin. The downregulation of
   inflammatory cytokine release occurred with a concomitant reduction in
   cadmium-induced elevations in p65 nuclear factor-κB (NF-κB) and
   extracellular signal-regulated kinases (ERKs) 1/2 phosphorylation.
   Additionally, levofloxacin treatment significantly alleviated
   cadmium-induced activation of phosphorylated NF-κB translocation and
   toll-like receptor (TLR)-4/signal transducer and activator of
   transcription (STAT) 3 signaling. Transcriptome analysis revealed that
   modulation of inflammation-related pathways was the most enriched after
   FQ treatment. Our data suggest that FQs, particularly levofloxacin,
   attenuate the inflammatory process mediated by cadmium in human
   astrocytes. These effects may be mediated, at least in part, by
   inhibition of immune pathways regulated by TLR4, STAT3, ERK MAPK, and
   NF-κB.

Introduction

   Cadmium is a non-essential element classified as a human carcinogen by
   the International Agency for Cancer Research and the US Environmental
   Protection Agency; cadmium intoxication is becoming a global health
   problem [[34]1]. Prolonged exposure to cadmium is strongly correlated
   with a wide range of neurological issues, such as malignancy, as well
   as long-term motor and cognitive impairment in children [[35]2, [36]3]
   and mining workers [[37]4, [38]5]. Astrocytes, glial cells of the
   central nervous system (CNS), are the brain’s most abundant and
   proliferative cell type. They play an essential role in the defense of
   neurons against neurotoxic insults; therefore, they are a primary
   target in mitigating heavy metal neurotoxicity. Although there are no
   cadmium-specific transport channels, cadmium enters cells via transport
   by other essential element channels (Ca^2+, Fe^2+, and Zn^2+) [[39]6].
   Cadmium also penetrates the brain via the blood–brain barrier (BBB) and
   olfactory pathways; once cadmium has entered the brain, it causes
   irreversible damage due to its very long biological half-life (10 to 30
   years) [[40]7]. Excessive accumulation of cadmium in astrocytes
   triggers long-lasting inflammatory responses. We previously showed that
   although interleukin (IL)-6 and IL-8 are barely detectable in healthy
   astrocytes, they are rapidly induced 10- to 100-fold in response to
   cadmium [[41]8], with even 1 μM cadmium, about 50 times below the safe
   level of cadmium intake recommended by the United States’ Food and Drug
   Administration (45 μM), eliciting an inflammatory response.

   Dysregulation of immune and inflammatory responses in the CNS
   microenvironment is another way by which cadmium exerts neurotoxic
   effects; this dysregulation may contribute to the development or
   progression of brain cancer and glioblastoma (GBM), the most common and
   aggressive malignant brain tumor in adults [[42]9]. IL-6 and IL-8
   likely play pivotal roles in microenvironment-associated
   chemotherapeutic resistance in GBM [[43]10]; therefore, using
   anti-inflammatory drugs in addition to cytotoxic drugs may result in
   more successful GBM treatments. Unfortunately, there are no proven
   effective treatments for chronic cadmium intoxication. One of the major
   difficulties in treating cadmium poisoning in the brain is the presence
   of the BBB, because most chelator agents, such as
   ethylenediaminetetraacetic acid (EDTA), cannot cross the BBB under
   physiological conditions [[44]11].

   Fluoroquinolones (FQs) are broad-spectrum quinolone antibiotics
   commonly used in clinical practice for urinary tract infections; FQs
   have a well-established pharmacokinetic and safety profile with a wide
   therapeutic index and minimal serious side effects [[45]12]. FQs also
   exert pleiotropic immunomodulatory effects. A previous experiment
   indicated that levofloxacin could be repurposed as a therapy to treat
   inflammation resulting from chronic bacterial infections such as
   pneumonia and cystic fibrosis [[46]13]. Several studies in immune cells
   i.e., human peripheral blood monocytes and THP-1 macrophages, show that
   FQs possess immunomodulatory properties and affect the levels of
   multiple molecules, including inducible nitric oxide synthase,
   cyclooxygenase-2, and other pro-inflammatory mediators [[47]14–[48]17].
   However, to date, the anti-neuroinflammatory activities of FQs have not
   been explored. Moxifloxacin, levofloxacin, and ciprofloxacin are
   considered a preferred choice for the treatment of bacterial meningitis
   and tuberculous meningitis due to their excellent lipid solubility,
   compared to norfloxacin, ofloxacin, and gemifloxacin [[49]18]. The
   therapeutic concentrations in cerebrospinal fluid (CSF) of these three
   drugs can be achieved after oral or intravenous administration. The
   median area under the curve (AUC) CSF/AUC serum ratios of moxifloxacin,
   levofloxacin, and ciprofloxacin were 0.53, 0.44, and 0.24, respectively
   [[50]19, [51]20]. However, some studies have shown that ciprofloxacin,
   a second-generation FQ, has been associated with a higher risk of
   tendonitis compared with newer-generation FQs such as moxifloxacin and
   levofloxacin [[52]21]. Moreover, ciprofloxacin is classified as a weak
   base and needs an acidic environment to dissolve [[53]22]. The study of
   ciprofloxacin in cell culture models is challenging because acidic pH
   may affect the stability and efficacy of the drug; we, therefore, did
   not include ciprofloxacin in this study.

   The neuroprotective effects of FQs are intriguing in preclinical
   studies; however, clinical evidence is still being developed.
   Moxifloxacin, levofloxacin, and ciprofloxacin are the most studied FQs
   for potential neuroprotective effects. Two published case reports
   elicited benefit outcomes of levofloxacin administration in patients
   diagnosed with Parkinson’s disease [[54]23] and progressive
   supranuclear palsy (PSP) [[55]24]. Another phase 2a proof-of-concept
   study in patients with amyotrophic lateral sclerosis (ALS) who received
   three daily doses of combined ciprofloxacin and celecoxib (PrimeC) for
   one year confirmed the safety and efficacy of the drug for ALS
   treatment [[56]25]. Ex vivo CSF and serum samples from meningitis
   patients given a single oral dose of 400 mg moxifloxacin influence the
   production of pro- and anti-inflammatory cytokines in
   lipopolysaccharide-activated human monocytes [[57]14]. However, there
   is a need for more robust clinical trials to understand their
   mechanisms of action better.

   In the present study, we explored the mechanisms underlying the
   anti-neuroinflammatory effect of FQs during cadmium poisoning. We
   demonstrate that moxifloxacin and levofloxacin inhibit cadmium-induced
   secretion of IL-6 and IL-8 in human astrocytes and that FQs selectively
   reduce cadmium-induced nuclear factor-κB (NF-κB), extracellular
   signal-regulated kinases (ERKs), toll-like receptor (TLR)-4, and signal
   transducer and activator of transcription (STAT) 3 expression or
   activation. Our results suggest that inhibition of these pathways may
   mediate the anti-neuroinflammatory effect of FQs.

Materials and methods

Human astrocyte cultures

   Primary human astrocytes (PHAs) were provided by Prof. Christopher
   Power (Department of Medicine, University of Alberta, Canada). PHAs
   were obtained from 15–19 week aborted fetuses over a five-year period
   May 2018 to May 2023 with written consent under protocol 27660 approved
   by the University of Alberta Human Research Ethics Board (Biomedical)
   and were prepared as reported previously [[58]26]. PHAs were grown in
   Minimum Essential Medium (MEM) containing 10% fetal bovine serum, 1%
   L-glutamine, 1% sodium pyruvate, and 1% penicillin–streptomycin (Gibco,
   NY, USA). The human astrocytoma U-87 MG cell line (U-87 MG) was
   obtained from American Type Culture Collection (VA, USA) and cultured
   in MEM supplemented with 10% fetal bovine serum, 1% sodium pyruvate,
   and 1% penicillin–streptomycin. Cells were grown at 37°C in 5% CO[2]
   fully humidified air and subcultured twice weekly. Cells were used from
   the fifth to the tenth passage. All biological reagents used in cell
   culture are guaranteed mycoplasma-free.

Preparation of cadmium chloride, moxifloxacin, and levofloxacin

   Cadmium chloride (CdCl[2]), moxifloxacin, and levofloxacin were
   purchased from Sigma Aldrich (MO, USA). Cadmium chloride was dissolved
   in sterile water at a concentration of 1000 mM and stored at −20°C.
   Moxifloxacin and levofloxacin were freshly dissolved in sterile water
   at a concentration of 10 mM and 20 mM, respectively. In the
   experiments, stock solutions were diluted in serum-free MEM (MEM plus
   1% sodium pyruvate) to reach the indicated concentrations.

MTT cell viability assay

   PHAs and human astrocytoma U-87 MG cells were plated at 1.5 × 10^4
   cells/well in a 96-well plates in complete MEM for 24 h followed by
   exposure to moxifloxacin, levofloxacin, cadmium, or combinations of
   moxifloxacin or levofloxacin with cadmium, in serum-free MEM for 24 h
   before assessment of cell viability by MTT assays. MTT solution was
   added to each well to a final concentration of 0.5 mg/mL. Formazan
   crystals were dissolved in dimethylsulfoxide and measured
   spectrophotometrically at 570 and 630 nm using a microplate reader
   (Thermo Fisher Scientific, Vantaa, Finland). Viability in treated cells
   was calculated as a percentage of that in mock-treated with culture
   medium.

Measurement of intracellular cadmium

   Human astrocytoma U-87 MG cells were plated at 1.5 × 10^6 cells in 100
   mm culture dishes in complete MEM for 24 h followed by exposure to
   cadmium, or combinations of moxifloxacin or levofloxacin with cadmium,
   in serum-free MEM for 24 h. Cells were then washed twice with
   phosphate-buffered saline (PBS) containing 10 mM disodium EDTA and
   twice with PBS without EDTA, trypsinized with 0.1% trypsin, and
   centrifuged at 3,000 × g for 10 min. Cell pellets were digested with
   concentrated nitric acid (65% w/w, 1 mL) at 90°C three times before
   digesting with concentrated nitric acid (65% w/w, 2 mL) overnight at
   room temperature. Afterward, the digested samples were diluted to a
   final volume of 20 mL and cadmium contents were measured using a flame
   furnace atomic absorption spectrophotometer (PinAAcle 900T, Perkin
   Elmer, MA, USA). The amount of intracellular cadmium content was
   expressed as μg of cadmium/g of wet weight protein.

Enzyme-linked immunosorbent assay (ELISA)

   Human astrocytoma U-87 MG cells were plated at 1.2 × 10^5 cells/well in
   12-well plates in complete MEM for 24 h followed by exposure to
   moxifloxacin, levofloxacin, or cadmium, or combinations of moxifloxacin
   or levofloxacin with cadmium, in serum-free MEM for 6 or 24 h. Spent
   media from U-87 MG cells was then collected and the amount of human
   IL-6 and IL-8 was determined using hIL-6 and hIL-8 ELISA kits according
   to the manufacturer’s instructions (eBioscience, CA, USA). Briefly,
   samples containing IL-6 and IL-8 were added to ELISA plates precoated
   with purified human IL-6 and IL-8 antibodies. IL-6 and IL-8 proteins
   linked with IL-6 and IL-8 antibodies were labeled with a
   biotin-conjugated antibody, avidin-HRP complex. This complex was then
   detected with the 3,3′, 5,5′-tetramethylbenzidine peroxidase solution
   and measured spectrophotometrically at 450 and 570 nm using a
   microplate reader (Thermo Fisher Scientific, Vantaa, Finland). The
   concentrations of the samples were calculated from the standard curve
   and expressed as pg/mL.

Western blotting

   PHAs and human astrocytoma U-87 MG cells were plated at 5 × 10^5 cells
   in 60 mm culture dishes in complete MEM for 24 h followed by exposure
   to cadmium, or combinations of moxifloxacin or levofloxacin with
   cadmium, in serum-free MEM for 1 h for detection of NF-κB and ERK or 3h
   for detection of TLR4 and STAT3. Cells were then washed with cold PBS
   and exposed to lysis buffer (50 mM Tris-base, 150 mM NaCl, 0.5% Nonidet
   P-40 in addition to protease inhibitor cocktail). Cell debris was
   removed by centrifugation and the supernatants were collected as
   whole-cell lysates. Total protein was quantified using Bio-Rad DC
   Protein assay reagents (Bio-Rad, Hercules, CA, USA). Proteins (10 μg)
   were separated using 12% SDS-PAGE and electroblotted in 20% methanol,
   25 mM Tris, and 192 mM glycine onto a nitrocellulose membrane. The
   membrane was blocked with 5% nonfat dry milk in 25 mM Tris-HCl, 150 mM
   NaCl, and 0.1% Tween 20, followed by incubation overnight at 4°C with
   anti-ERK1/2 (Cat# 4695 diluted 1:1000—Cell Signaling, MA, USA),
   anti-p-ERK1/2 (Cat# 3179 diluted 1:1000—Cell Signaling), anti-p65 NF-κB
   (Cat# AB16502 diluted 1:1000—Abcam), anti-p-p65 NF-κB (Cat# AB86299
   diluted 1:1000—Abcam, Cambridge, UK), anti-STAT3 (Cat# 12640s diluted
   1:1000—Cell Signaling, MA, USA), anti-p-STAT3 (Cat# 9145s diluted
   1:2000—Cell Signaling, MA, USA), anti-TLR4 (Cat# 48–2300 diluted
   1:200—Invitrogen, MA, USA), or β-actin (Cat# 4070 diluted 1:3000—Cell
   Signaling, MA, USA), or GAPDH (Cat# 5174 diluted 1:3000—Cell Signaling,
   MA, USA). Afterward, membranes were incubated at room temperature for 2
   h with monoclonal anti-rabbit IgG-HRP (Cat# 111-035-003 diluted
   1:5000—Jackson ImmonoResearch, PA, USA) and developed using an enhanced
   chemiluminescence system kit (Thermo Fisher Scientific, MA, USA).
   Membranes were imaged with Image Studio software version 5.2 (LI-COR,
   Lincoln, NE, USA). Protein levels were normalized to β-actin or GAPDH.

Immunofluorescence analysis of NF-κB (p65) localization

   Human astrocytoma U-87 MG cells were plated at 1.5 × 10^4 cells in an
   8-well cell culture chamber slide in complete MEM for 24 h followed by
   exposure to cadmium, or combinations of moxifloxacin or levofloxacin
   with cadmium, in serum-free MEM for 1 h for detection of NF-κB nuclear
   translocation. Cells were then washed with cold PBS, fixed with cold 4%
   paraformaldehyde, permeabilized with 0.1% Triton X-100 (Sigma, MO,
   USA), and blocked with 1% Bovine serum albumin (Himedia, Maharashtra,
   India). Subsequently, cell samples were incubated overnight at 4°C with
   a primary antibody specific to p-p65 NF-κB (Cat# AB86299 diluted
   1:500—Abcam, Cambridge, UK), followed by incubation at room temperature
   for 2 h with monoclonal Alexa fluor 488 goat anti-rabbit IgG-HRP (Cat#
   A11008 diluted 1:500—Thermo Invitrogen, CA, USA). Fluoroshield mounting
   medium with DAPI (Cat#104139—Abcam, Cambridge, UK) was counterstained
   to locate the nucleus. Images were observed by immunofluorescent
   microscopy (EVOS FL Auto 2 Cell Imaging System—Thermo Invitrogen, CA,
   USA). Data shown representative of three independent experiments.
   Original magnification, ×40. The fluorescence intensity of cytoplasmic
   and nuclear p65 subunit was calculated using ImageJ software and
   results are presented as a percentage of nucleic over cytoplasmic
   positive cells to p65 NF-κB.

RNA preparation and RNA-seq analysis

   Human astrocytoma U-87 MG cells were plated at 5 × 10^5 cells in 60 mm
   culture dishes in complete MEM for 24 h followed by exposure to
   cadmium, or combinations of moxifloxacin or levofloxacin with cadmium,
   in serum-free MEM for 3 h. Cells were then washed with PBS and total
   RNA was isolated from cells using Monarch Total RNA Miniprep Kit with
   on-column DNase treatment according to the manufacturer’s instructions
   (New England Biolabs, MA, USA). An equal amount of RNA was pooled from
   four independent experiments. The quality and quantity of RNA were
   analyzed using capillary electrophoresis on the Bioanalyzer 2100
   (Agilent Technologies, California, United States). The extracted RNA
   was stored at −80°C until sequencing. RNA-seq libraries were
   constructed using the TruSeq mRNA LT Sample Prep Kit using Illumina
   HiSeq 2500 (Illumina, CA, USA), according to the manufacturer’s
   instructions, by Macrogen (Seoul, South Korea). RNA-seq analysis was
   performed by BMKGENE Multi-omics Sequencing Services (Beijing, China).
   Briefly, high-quality reads were aligned with the human genome hg38
   (Homo_sapiens.GRCh38_release95.genome.fa) using HISAT2. Transcripts
   from the preprocessed RNA-seq alignment assembly were identified, and
   their expression levels were measured using Stringtie with Gencode V32
   annotation. The sequencing data of the following groups were combined
   for analysis using Stringtie Merge: (1) mock-treated vs.
   cadmium-treated; (2) cadmium-treated and moxifloxacin vs.
   cadmium-treated; (3) cadmium-treated and levofloxacin vs.
   cadmium-treated; and (4) combinations of moxifloxacin with
   cadmium-treated group vs. combinations of levofloxacin with
   cadmium-treated group. DESeq2 was then used to compare the combined
   transcripts to determine the differential expression levels of given
   transcripts.

Mapping identification of differentially expressed genes and functional
enrichment analysis

   AnnotateDESeq2 was used to map transcripts with annotated gene IDs to
   generate a list of differentially expressed genes (DEGs).
   Down-regulated genes were defined as those with a log2FC < −1 and P
   value < 0.05 and up-regulated genes were defined as those with a log2FC
   > 1 and P value < 0.05. To provide insights into the biological
   function of enriched DEGs, DEGs in the following groups were subjected
   to Gene Ontology (GO) term functional enrichment analysis and Kyoto
   Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis
   using the g:Profiler tool: (1) mock-treated vs. cadmium-treated; (2)
   cadmium-treated vs. moxifloxacin and cadmium-treated; (3)
   cadmium-treated groups vs. levofloxacin and cadmium-treated; and (4)
   combinations of moxifloxacin with cadmium-treated group vs.
   combinations of levofloxacin with cadmium-treated group. Pathway
   mapping was conducted using the Homo sapiens Orthology database as a
   reference.

Statistical analyses

   Statistical analyses were performed using GraphPad Prism software
   (GraphPad Software, version 10, CA, USA). All results were expressed as
   means ± standard error of mean (SEM) from at least three independent
   experiments performed in triplicate. Multiple comparisons were
   performed by a one-way analysis of variance (ANOVA) followed by Tukey’s
   post-hoc tests. Student’s t-tests were used to detect significant
   differences between two groups. A P value < 0.05 was considered
   statistically significant.

Results

Cellular toxicity of FQs on U-87 MG human astrocytes and PHAs

   To investigate the anti-inflammatory effects of FQs independent of cell
   growth inhibition, their potential cytotoxicity on U-87 MG human
   astrocytes and PHAs was assessed. Cells were incubated with either
   moxifloxacin or levofloxacin at various concentrations (10, 100, 500,
   1000 μM) for 24 h and cell viability was measured using an MTT assay.
   As shown in [59]Fig 1A, there were no significant toxic effects of FQs
   on cells at concentrations between 10 and 500 μM. When cells were
   exposed to both cadmium and FQs, cytotoxic effects of FQs were not seen
   ([60]Fig 1B). Consistent with the results in U-87 MG human astrocytes,
   FQ administration at all concentrations did not induce significant cell
   death in PHAs ([61]Fig 1C). In general, intracellular cadmium
   accumulation in human U-87 MG astrocytes decreased with increasing
   moxifloxacin and levofloxacin concentrations. At 500 μM, moxifloxacin
   had a strong chelating effect in the presence of cadmium ([62]Fig 1D).
   To employ therapeutically relevant concentrations and avoid cytotoxic
   effects, FQ concentrations of 10 and 100 μM were selected for the
   following experiments.

Fig 1. Effect of moxifloxacin and levofloxacin on cell viability of U-87 MG
human astrocytes and primary human astrocytes.

   [63]Fig 1
   [64]Open in a new tab

   A U-87 MG cells were treated with various concentrations of either
   moxifloxacin or levofloxacin (0, 10, 100, 500, and 1000 μM) for 24 h
   and cell viability was determined by MTT cell viability assay. B, C
   U-87 MG cells and primary human astrocytes were cotreated with cadmium
   (1–10 μM) and either moxifloxacin or levofloxacin (10–500 μM) for 24 h
   and cell viability was determined by MTT cell viability assay. The cell
   viability results are expressed as the percentages of mock-treated
   cells. D U-87 MG cells were cotreated with cadmium (10 μM) and either
   moxifloxacin or levofloxacin (100 and 500 μM) for 24 h and the
   intracellular cadmium level was determined by FAAS technique. 10 μM of
   BAPTA and 20 of μM EDTA were used as positive controls. Data represents
   mean±SEM of at least three independent experiments. *P<0.05 and
   **P<0.01 compared between cadmium-treated cells and mock-treated cells;
   ^#P<0.05 and ^##P<0.01 compared between cells exposed to cadmium and
   moxifloxacin or cadmium and levofloxacin compared with cadmium alone.
   Cd, Cadmium; MFX, Moxifloxacin; LFX, Levofloxacin.

FQs reduce the production of proinflammatory cytokines in cadmium-activated
U-87 MG human astrocytes

   To confirm the anti-inflammatory role of FQs in cadmium activation,
   protein levels of IL-6 and IL-8 in the cell culture medium of U-87 MG
   cells treated with 1 and 10 μM of cadmium, either in the presence or
   absence of 10 and 100 μM FQs, were determined by ELISA. As shown in
   [65]Fig 2A, 2B, cadmium dose-dependently induced IL-6 and IL-8 release
   from U-87 MG cells, while FQs alone did not have any effect on cytokine
   production. The cadmium-induced increase in IL-6 and IL-8 was
   ameliorated by FQ treatment in a concentration and time-dependent
   manner; however, moxifloxacin administered at 10 μM had no significant
   effects on either IL-6 or IL-8 secretion (both P > 0.05). Overall,
   moxifloxacin had a smaller inhibitory effect on both IL-6 and IL-8
   release than levofloxacin.

Fig 2. Effect of moxifloxacin and levofloxacin on cadmium-induced
upregulation of IL-6 and IL-8 in U-87 MG human astrocytes.

   [66]Fig 2
   [67]Open in a new tab

   A, B U87-MG cells were exposed to cadmium (1 and 10 μM) in the absence
   or the presence of 10 and 100 μM moxifloxacin or levofloxacin for 6 and
   24 h and the protein levels of IL-6 and IL-8 in the culture medium were
   determined by ELISA. Data represents mean±SEM of at least three
   independent experiments. *P<0.05 and **P<0.01 compared between
   cadmium-treated cells and mock-treated cells; ^#P<0.05 and ^##P<0.01
   compared between cells exposed to cadmium and moxifloxacin or cadmium
   and levofloxacin compared with cadmium alone. Cd, Cadmium; MFX,
   Moxifloxacin; LFX, Levofloxacin.

FQs suppress the phosphorylation of ERK and NF-κB in cadmium-activated U-87
MG human astrocytes and PHAs

   As mitogen-activated protein kinase (MAPK) and NF-κB are major
   regulators of the inflammatory response induced by cadmium, we
   hypothesized that the anti-inflammatory activity of FQs might result
   from blocking activation of these pathways. To ensure the robustness of
   our findings and confirm the observed effects are not confined to a
   single cell line, we designed the study to incorporate two distinct
   cells. Western blotting analysis of U-87 MG human astrocytes and PHAs
   revealed a similar therapeutic response to the drugs. While
   cadmium-treated cells had significantly higher levels of ERK1/2 and p65
   NF-κB phosphorylation than mock-treated cells (P < 0.05), this increase
   was alleviated by treatment with FQs at 100 μM, but not 10 μM ([68]Fig
   3). In particular, the phosphorylation levels of p65 NF-κB were
   significantly lower in cells treated with cadmium and levofloxacin than
   in those treated with cadmium and moxifloxacin (P < 0.05) ([69]Fig 3C,
   3D). Furthermore, the immunocytochemical analysis confirmed the Western
   blotting findings that cadmium induced translocation of p65 from the
   cytoplasm to the nucleus, and levofloxacin appeared to have a strong
   potent effect on blocking the translocation ([70]Fig 4).

Fig 3. Effect of moxifloxacin and levofloxacin on cadmium-induced ERK and p65
NF-κB phosphorylation in U-87 MG human astrocytes and primary human
astrocytes.

   [71]Fig 3
   [72]Open in a new tab

   A, B Representative blots of phosphorylated and total forms of ERK1/2
   in U87-MG cells and primary human astrocytes exposed to cadmium (10 μM)
   in the absence or the presence of 10 and 100 μM moxifloxacin or
   levofloxacin for 1 h. C, D Representative blots of phosphorylated and
   total forms of p65 NF-κB in U87-MG cells and primary human astrocytes
   exposed to cadmium (10 μM) in the absence or the presence of 10 and 100
   μM moxifloxacin or levofloxacin for 1 h. Data represents mean±SEM of at
   least three independent experiments. *P<0.05 compared between
   cadmium-treated cells and mock-treated cells; ^#P<0.05 and ^##P<0.01
   compared between cells exposed to cadmium and moxifloxacin or cadmium
   and levofloxacin compared with cadmium alone. Cd, Cadmium; MFX,
   Moxifloxacin; LFX, Levofloxacin. The original blots are presented in
   [73]S1 Raw images.

Fig 4. Effect of moxifloxacin and levofloxacin on cadmium-induced p65 NF-κB
nuclear translocation in U-87 MG human astrocytes.

   [74]Fig 4
   [75]Open in a new tab

   A, B Representative images of immunofluorescence staining of p65 NF-κB
   and percentage of nucleic over cytoplasmic positive cells to p65 NF-κB
   in U87-MG cells exposed to cadmium (10 μM) in the absence or the
   presence of 10 and 100 μM moxifloxacin or levofloxacin for 30 min. The
   nuclei of astrocytes were labeled by DAPI (blue) and the NF-κB protein
   was labeled by anti-p65 antibody (green). Cells were photographed under
   confocal microscopy (scale bar 100 μm, magnification × 40). Data
   represents mean±SEM of at least three independent experiments. *P<0.05
   compared between cadmium-treated cells and mock-treated cells; ^#P<0.05
   and ^##P<0.01 compared between cells exposed to cadmium and
   moxifloxacin or cadmium and levofloxacin compared with cadmium alone.
   Cd, Cadmium; MFX, Moxifloxacin; LFX, Levofloxacin.

FQs suppress the expression of TLR4 and phosphorylation of STAT3 in
cadmium-activated U-87 MG human astrocytes and PHAs

   Previous reports demonstrate that the TLR4 pathway mediates the
   inhibitory effect of FQs on pro-inflammatory factor expression and that
   FQ is a TLR4 antagonist [[76]27]. Therefore, we determined the effect
   of TLR4 activation on the inhibitory effects of FQs on inflammatory
   responses. As shown in [77]Fig 5A, 5B, both U-87 MG human astrocytes
   and PHAs exhibited a marked increase in TLR4 levels in cadmium-treated
   cells (P < 0.05), a response that was notably diminished in the
   presence of levofloxacin. This indicates that the ameliorating effects
   of levofloxacin treatment on cadmium toxicity may stem from its
   regulation of the TLR4 signaling pathway. Because cross-talk between
   TLR4 and STAT3 regulates cytokine-dependent inflammation, we next
   investigated whether STAT3 was also involved in the anti-inflammatory
   effects of FQs. As shown in [78]Fig 5C, 5D, consistent with the results
   of TLR4, levofloxacin inhibited the STAT3 phosphorylation induced by
   cadmium. By contrast, moxifloxacin had no inhibitory effect on
   cadmium-induced TLR4 expression or STAT3 activation in human
   astrocytes.

Fig 5. Effect of moxifloxacin and levofloxacin on cadmium-induced TLR4 and
STAT3 pathway in U-87 MG human astrocytes and primary human astrocytes.

   [79]Fig 5
   [80]Open in a new tab

   A, B Representative blots of TLR4 in U87-MG cells and primary human
   astrocytes exposed to cadmium (10 μM) in the absence or the presence of
   10 and 100 μM moxifloxacin or levofloxacin for 3 h. C, D Representative
   blots of phosphorylated and total forms of STAT3 in U87-MG cells and
   primary human astrocytes exposed to cadmium (10 μM) in the absence or
   the presence of 10 and 100 μM moxifloxacin or levofloxacin for 3 h.
   Data represents mean±SEM of at least three independent experiments.
   *P<0.05 compared between cadmium-treated cells and mock-treated cells;
   ^#P<0.05 and ^##P<0.01 compared between cells exposed to cadmium and
   moxifloxacin or cadmium and levofloxacin compared with cadmium alone.
   Cd, Cadmium; MFX, Moxifloxacin; LFX, Levofloxacin. The original blots
   are presented in [81]S1 Raw images.

FQ alters transcriptome profiles of cadmium-activated U-87 MG human
astrocytes

   To comprehensively understand the effects of FQs on cadmium-induced
   inflammatory responses in human astrocytes, and to study the underlying
   mechanisms, we performed whole-transcriptome sequencing (RNA-seq)
   analysis. Gene expression from cadmium-treated cells was compared with
   that from mock-treated, cadmium and moxifloxacin-treated, and cadmium
   and levofloxacin-treated cells. Comparative transcriptome analyses
   identified 2012 DEGs, including 1268 upregulated and 744 downregulated
   genes in cadmium-treated cells relative to mock-treated cells ([82]Fig
   6A). We next clustered the DEGs and predicted the functional canonical
   pathway networks associated with these changes in gene expression using
   Gene Set Enrichment Analysis. Results of KEGG enrichment analysis
   showed the top-ranked upregulated canonical pathways were “TNF (tumor
   necrosis factor) signaling pathway,” “pathways in cancer,” “IL-17
   signaling pathway,” “MAPK signaling pathway,” “NF-κB signaling
   pathway,” “cytokine–cytokine receptor interaction,” “mineral
   absorption,” “JAK (Janus kinase)-STAT signaling pathway,” and “TLR
   signaling pathway” ([83]Fig 6B and [84]S1 Table). Of note, the enriched
   KEGG pathways were involved in cadmium-induced neuropathology.

Fig 6. Comparative transcriptomics analysis for differentially expressed
genes (DEGs) in cells treated with and without cadmium.

   [85]Fig 6
   [86]Open in a new tab

   A Volcano plot of up-regulated (p < 0.05, log2FC > 1) and
   down-regulated (p < 0.05, log2FC < −1) DEGs. B KEGG enrichment scatter
   plot of up-regulated DEGs. The dot size indicates the amount of DEG
   enriched in the pathway. The darker the color, the more significant the
   enrichment result. Cd, Cadmium; FC, Fold change.

   When comparing the expression of transcripts in U-87 MG cells modulated
   by FQs in response to cadmium, a heatmap showed that changes were
   similar in response to both moxifloxacin and levofloxacin treatment
   ([87]Fig 7A). Across the two FQs, patterns of DEGs were similarly
   enriched in the following KEGG pathways: “antigen processing and
   presentation,” “purine metabolism,” “PI3K (phosphoinositide
   3-kinase)-AKT signaling pathway,” “MAPK signaling pathway,” and
   “inflammatory mediator regulation of TRP (transient receptor potential)
   channels” ([88]Fig 7B, 7C and [89]S2 and [90]S3 Tables). Nonetheless,
   some notable differences were observed. KEGG analysis revealed that
   genes associated with “mineral absorption” and “oxidative stress”
   pathways were downregulated in moxifloxacin-treated cells, whereas gene
   signatures in “cytokine–cytokine receptor interaction,” “calcium
   signaling pathway,” “cAMP signaling pathway,” “NF-κB signaling
   pathway,” and “JAK-STAT signaling pathway” were downregulated in
   levofloxacin-treated cells ([91]S4 Table). Results of GO analysis were
   closely associated with KEGG pathway analysis. As shown in [92]Fig 7D
   and [93]S5 Table, enriched GO terms were mainly related to “cytokine
   receptor binding,” “zinc ion binding,” “protein kinase binding,” and
   “phosphotransferase activity,” as well as “MAPK cascade” and “pattern
   recognition receptor (PRR) signaling pathway.”

Fig 7. Comparative transcriptomics analysis for differentially expressed
genes (DEGs) in cadmium-activated cells after treatment with moxifloxacin or
levofloxacin.

   [94]Fig 7
   [95]Open in a new tab

   A Heat map and hierarchical clustering of all DEGs in moxifloxacin
   (Cd10+M100) vs. levofloxacin (Cd10+L100) treatment. Blue indicates
   downregulated expression, and red indicates upregulated expression
   compared with a reference expression level. B, C KEGG enrichment
   scatter plot among the most up-regulated and down-regulated DEGs in
   cadmium-treated cells (Cd10) vs. combinations of moxifloxacin with
   cadmium-treated cells (Cd10+M100) and cadmium-treated cells (Cd10) vs.
   combinations of levofloxacin with cadmium-treated cells (Cd10+L100).
   The bar chart indicates the percentage of annotated genes. D Enriched
   GO terms corresponding to the most up-regulated and down-regulated DEGs
   in the biological process (BP), cellular component (CC), and molecular
   function (MF). The dot size indicates the amount of DEG enriched in the
   pathway. The darker the color, the more significant the enrichment
   result. Cd, Cadmium; MFX, Moxifloxacin; LFX, Levofloxacin.

Discussion

   The neuroinflammatory effect of cadmium exposure is widely
   acknowledged. Specifically, our previous study revealed that the
   pro-inflammatory mediators IL-6 and IL-8 are released from the brain
   cells after cadmium exposure [[96]8]. Although attenuation of this
   inflammatory response limits cadmium-induced tissue damage, current
   therapies for cadmium intoxication are limited and there is no specific
   treatment targeting cadmium-induced inflammation. Given the health
   benefits of FQs, we aimed to test the efficacy of repurposing FQs as
   anti-inflammatory agents. In this study, we compared the
   anti-inflammatory effects of two classes of FQs in cadmium-treated
   human astrocytes in vitro. We found that FQ administration,
   particularly of levofloxacin, ameliorated cadmium-induced inflammatory
   responses in human astrocytes via reduced activation or expression of
   inflammation-related signaling molecules, including TLR4, STAT3, ERK,
   and NF-κB.

   The attenuating effects of moxifloxacin and levofloxacin on IL-6 and
   IL-8 secretion we observed in cadmium-activated astrocytes were
   achieved at levels close to clinically relevant serum and cerebrospinal
   fluid concentrations. The human serum and cerebrospinal fluid
   concentrations of moxifloxacin (following a typical dose of 400 mg)
   peak at 6.93–13.32 μM and 1.95–2.64 μM, respectively, whereas for
   levofloxacin (following a typical dose of 750 mg), concentrations are
   23.93–36.34 μM and 6.73–11.16 μM, respectively [[97]19]. Comparable to
   our study, previous reports on the anti-inflammatory characteristics of
   these two medications were carried out at drug doses ranging from 10 to
   200 μM [[98]13, [99]15]. Interestingly, our results show that
   moxifloxacin and levofloxacin had a different cytokine inhibition
   pattern, with levofloxacin having superior anti-inflammatory activity.
   By contrast, moxifloxacin has more potent immunomodulation activity,
   which may result from its structure: it has a cyclopropyl moiety at the
   N1 position of the quinolone core structure, a methoxy group (-CHO) on
   C8, and a bulky C7 side chain [[100]28]. However, these differences in
   functionality between levofloxacin and moxifloxacin are not consistent
   in all studies: in primary human bronchial epithelial cells from
   patients undergoing lung resections, 20 μM moxifloxacin but not
   levofloxacin significantly reduced IL-8 release [[101]29]. This
   suggests that although the anti-inflammatory and immunomodulatory
   effects of FQs may differ resulting from their chemical structure,
   which stimulant they treat and the cell model employed may also have an
   effect.

   The inhibition of inflammatory cytokine production by FQs was
   associated with several molecules involved in inflammatory signaling
   pathways. The elevation of inflammatory cytokines and phosphorylation
   of ERK and NF-κB induced by cadmium were downregulated with FQ
   treatment. We previously found that pretreatment with specific MAPK and
   NF-κB inhibitors significantly downregulates IL-6 and IL-8 secretion
   upon cadmium exposure [[102]8]. Consistent with this, downregulation of
   IL-6 and IL-8 secretion in cadmium-activated astrocytes by moxifloxacin
   and levofloxacin seemed to be achieved by inhibiting phosphorylation of
   key intracellular signaling molecules, including ERK and p65 NF-κB.
   Similarly, moxifloxacin inhibits NF-κB and ERK activation, which is
   required for IL-8 transcription, in monocytes and bronchial epithelial
   cells [[103]15, [104]30, [105]31], while levofloxacin decreases levels
   of aerosol MP-376-induced IL-6 and IL-8 by inhibiting NF-κB activation
   in human cystic fibrosis bronchial epithelial cells [[106]13].
   Therefore, the present study extends previous findings that FQs act as
   inhibitors of NF-κB and ERK signaling pathways.

   We found that cadmium upregulated TLR4 expression in human astrocytes
   and that this was suppressed by FQ administration. TLR4 is a
   pathogen-associated molecular pattern and damage-associated molecular
   pattern receptor that plays a crucial role in the innate immune
   response. Overactivation of TLR4 is closely linked with excessive
   cytokine production and consequently, heightened immune and
   inflammatory responses [[107]32, [108]33]. Cadmium is an alternative
   TLR4 ligand, which can mediate damage-associated molecular
   pattern-induced inflammatory responses through PRRs and in turn
   activate inflammatory signaling [[109]34]. Cadmium exposure causes TLR4
   activation-associated cell damage by positively regulating MAPK and
   NF-κB signaling pathways such as ERK1/2 and p38 MAPK in human airway
   epithelial cells [[110]35]; MAPK (ERK, JNK, and p38)/NF-κB in porcine
   and murine testicular cells [[111]36, [112]37]; and MyD88/NF-κB in
   chicken peripheral blood lymphocytes [[113]38]. In addition, cadmium
   activates TLR4 signaling in rat brain cells, leading to the
   upregulation of neuronal inflammatory cytokines [[114]39]. However, it
   is yet to be established whether these effects are caused by a physical
   interaction between TLR4 and cadmium similar to other metal ions:
   nickel and cobalt trigger TLR4 dimerization and cellular activation
   through histidine residues in the TLR4 ectodomain [[115]40].
   Furthermore, we found that cadmium upregulated TLR4 expression and
   resulted in activation of its downstream target STAT3 in human
   astrocytes, which was selectively suppressed by levofloxacin. In
   support of our findings, a previous study found that treatment with
   levofloxacin diminished production of IL-1β and tumor necrosis factor-α
   from lipopolysaccharide-activated rat microglia by blocking
   lipopolysaccharide binding to TLR4–MD-2; this in turn diminished TLR4
   homodimerization and activation [[116]27]. Therefore, TLR4 inhibition
   may be an alternative pathway by which FQs elicit anti-inflammatory
   activities.

   In our study, pathways enriched in cadmium-treated human astrocytes
   were closely related to pathways downregulated after FQ treatment. The
   predominant enriched pathways in cadmium-activated astrocytes were
   those related to IL-17, MAPK, NF-κB, JAK-STAT, and TLR signaling, as
   well as cytokine–cytokine receptor interaction and mineral absorption.
   Our transcriptome analysis data showed that inflammatory pathways were
   especially enriched in cadmium-treated cells and that treatment with
   FQs downregulated genes involved in the IL-17 signaling pathway,
   cytokine–cytokine receptor interaction, and the PRR signaling pathway
   implicated in their anti-inflammatory role. We also observed a strong
   chelating effect of moxifloxacin in the presence of cadmium ions. This
   may be due to its unique structure which contains two main sites (COOH
   and a C = O group) of divalent metal cation chelate formation
   [[117]41]. Taken together, results suggest that FQs may serve as an
   anti-inflammatory agent with chelating actions that could be used in
   managing cadmium-induced neurotoxicity in the future.

   Repurposing FQs for non-infectious diseases such as chronic obstructive
   pulmonary disease (COPD) ([118]NCT04879030) (ciprofloxacin,
   levofloxacin, and moxifloxacin) and Crohn’s disease ([119]NCT03850509)
   (OPS-2071) are still an ongoing clinical trial to carefully evaluated
   the risk of promoting resistance. Our results demonstrated that
   moxifloxacin and levofloxacin exhibited neuroprotective effects at
   concentrations below those typically used for antimicrobial therapy.
   Minimum inhibitory concentration (MIC) for the majority of pathogens
   causing bacterial meningitis (Staphylococcus aureus and Streptococcus
   pneumoniae) and tuberculous meningitis (Mycobacterium tuberculosis) of
   moxifloxacin and levofloxacin are 0.012 to 2 mg/L (0.03–5 μM) and 0.016
   to 8 mg/L (0.04–22 μM), respectively [[120]42]. This implies that
   neuroprotective concentrations of moxifloxacin and levofloxacin might
   not have robust antibacterial activity. In addition, a striking
   concentration-dependent relationship was noticed in the effects of FQs
   on astrocytes, where elevated concentrations heighten the risk of
   neurotoxicity. In agreement with several studies, moxifloxacin and
   levofloxacin can significantly cross BBB [[121]19, [122]20]. This
   characteristic enhances their potential to affect the CNS and could
   lead to both therapeutic benefits and risks of side effects. Therefore,
   further research is needed to ensure the long-term effectiveness and
   safety of repurposed antibacterial agents for neuroprotective effects,
   particularly in vulnerable populations.

   Although the present study presents many insights, it also has some
   limitations. One limitation is that although some experimental studies
   on the anti-inflammatory effect of FQs were conducted, we did not
   clarify whether FQs affect other cell signaling pathways, as reported
   by the RNAseq data, which may also be involved in the anti-inflammatory
   effect of FQs. It would be interesting for future studies to clarify
   the multiple targets of moxifloxacin and levofloxacin that mediate its
   inhibitory effect on pro-inflammatory factor expression. An additional
   limitation is that we were only able to examine the impact of FQs as a
   treatment during inflammatory stimulation. Since most cases of cadmium
   toxicity result in chronically induced inflammation, testing the impact
   of FQs administered after cadmium-stimulated astrocyte inflammation
   would be therapeutically important.

Conclusion

   We found that FQs had an inhibitory effect on cadmium-induced release
   of IL-6 and IL-8 in human astrocytes. Compared to moxifloxacin,
   levofloxacin appeared to have superior anti-inflammatory properties.
   Inhibition of TLR4/STAT3 and/or ERK/NF-κB signaling pathways appears to
   be a key mechanism by which FQs exert their anti-inflammatory effects
   ([123]Fig 8). FQs may be further applied for anti-inflammatory clinical
   use; however, further investigations in preclinical and clinical
   studies are required.

Fig 8. Schematic presentation of possible signaling cascades of
fluoroquinolone in suppression of the cadmium-induced inflammatory response.

   [124]Fig 8
   [125]Open in a new tab

   DAMP, damage-associated molecular pattern; MFX, moxifloxacin; LFX,
   levofloxacin; ERK, extracellular signal-regulated kinases; TLR4,
   toll-like receptor 4; IL, interleukin; IκB-α, inhibitor of κB-α;
   p50/p65, nuclear factor-κB; STAT3, signal transducer and activator of
   transcription 3.

Supporting information

   S1 Table. Functional enrichment analysis of DEGs between mock-treated
   cells (MEM) and cadmium-treated cells (Cd10) by KEGG in human
   astrocytoma U-87 MG cell lines.

   (PDF)
   [126]pone.0317281.s001.pdf^ (111.3KB, pdf)
   S2 Table. Functional enrichment analysis of DEGs between
   cadmium-treated cells (Cd10) and combinations of moxifloxacin with
   cadmium-treated cells (Cd10+MFX100) by KEGG in human astrocytoma U-87
   MG cell lines.

   (PDF)
   [127]pone.0317281.s002.pdf^ (110.9KB, pdf)
   S3 Table. Functional enrichment analysis of DEGs between
   cadmium-treated cells (Cd10) and combinations of levofloxacin with
   cadmium-treated cells (Cd10+LFX100) by KEGG in human astrocytoma U-87
   MG cell lines.

   (PDF)
   [128]pone.0317281.s003.pdf^ (109.4KB, pdf)
   S4 Table. Functional enrichment analysis of DEGs between
   cadmium-activated cells after treatment with moxifloxacin (Cd10+MFX100)
   and after treatment with levofloxacin (Cd10+LFX100) by KEGG in human
   astrocytoma U-87 MG cell lines.

   (PDF)
   [129]pone.0317281.s004.pdf^ (117.6KB, pdf)
   S5 Table. Functional enrichment analysis of DEGs between
   cadmium-activated cells after treatment with moxifloxacin (Cd10+MFX100)
   or levofloxacin (Cd10+LFX100) by GO in human astrocytoma U-87 MG cell
   lines.

   (PDF)
   [130]pone.0317281.s005.pdf^ (117KB, pdf)
   S1 Raw images. Raw western blot images for Figs [131]3 and [132]5.

   (PDF)
   [133]pone.0317281.s006.pdf^ (3.9MB, pdf)

Acknowledgments