Roflumilast Elicits Therapeutic and Neuroprotective Effects in 3-Nitropropionic Acid-Induced Huntington’s Disease-Like Neurodegeneration in Rats by Mitigating NLRP3 Inflammasome-Mediated Pyroptosis, Ferroptosis, and Glial Activation
Mohamed Taha, Dalia Salah, Kareem Abdou, Mahmoud A. Senousy
TL;DR
Roflumilast protects against Huntington’s disease-like brain damage in rats by reducing harmful cell death and inflammation.
Contribution
This study demonstrates roflumilast's therapeutic effects in Huntington’s disease by targeting pyroptosis, ferroptosis, and glial activation.
Findings
Roflumilast improved motor and behavioral deficits in 3-NP-induced HD-like rats.
The drug reduced NLRP3 inflammasome activation and markers of pyroptosis and ferroptosis.
Roflumilast attenuated astrogliosis, microglial activation, and neuroinflammation in the striatum.
Abstract
Huntington’s disease (HD) pathogenesis involves diverse cellular mechanisms, yet the contributions of pyroptosis and ferroptosis remain elusive. Roflumilast, a phosphodiesterase-4 (PDE-4) inhibitor, has shown neuroprotective effects, but its precise mechanisms are yet to be elucidated. We evaluated the potential neuroprotective and therapeutic effects of roflumilast in 3-nitropropionic acid (3-NP)-induced HD-like neurodegeneration, focusing on pyroptotic and ferroptotic cell death signaling. Adult male Wistar rats were assigned to five groups: normal control (saline + 0.5% carboxymethyl cellulose), roflumilast-control (1 mg/kg/day, p.o. for 21 days), 3-NP (20 mg/kg/day, i.p. for seven days), roflumilast-prophylactic (1 mg/kg/day, p.o. for 21 days prior to 3-NP), and roflumilast-treatment (1 mg/kg/day, p.o. for 21 days post-3-NP). Behavioral outcomes of the open-field, rotarod, and grip…
Genes, proteins, chemicals, diseases, species, mutations and cell lines named across the full text — each resolved to its canonical identifier and authoritative record.
Click any figure to enlarge with its caption.
Figure 10
Figure 11
Figure 1
Figure 2
Figure 3
Figure 4
Figure 5
Figure 6
Figure 7
Figure 8
Figure 9
Figure 12- —Cairo University
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
TopicsInflammasome and immune disorders · Tryptophan and brain disorders · Genetic Neurodegenerative Diseases
Introduction
Huntington’s disease (HD) is a progressive, vicious, neurological genetic disorder. Its estimated pooled prevalence among populations is 4.88 per 100,000 people worldwide [1]. HD is characterized by cognitive, mental, and motor impairments, with Huntington’s chorea as one of its characteristic motor signs, followed by hypokinetic symptoms as the disease progresses [2, 3]. Pathologically, HD is characterized by striking striatal neurodegeneration, particularly of the striatal medium spiny neurons [2, 3]. The treatments available only aim to control HD symptoms, with no current treatment to prevent its progression [2–4], opening the path for discovering new drugs for HD management.
The pathophysiology of neuronal cell death in HD can be explained based on several theories, including mutant huntingtin (mHtt) protein accumulation (expanded CAG repeats), transcriptional dysregulation, excitotoxicity, mitochondrial dysfunction, and oxidative stress [2, 3]. One well-known animal model of HD is the 3-nitropropionic acid (3-NP) model. 3-NP induces neurotoxicity by irreversibly inhibiting the succinate dehydrogenase enzyme (complex II) in mitochondria, causing mitochondrial dysfunction and consequent oxidative injury [5]. The resulting increase in reactive oxygen species (ROS) production, along with compromised antioxidant defenses, leads to striatal neurodegeneration [5]. The appearance of striatal lesions alongside the motor impediments following 3-NP intake resembles the HD phenotype, highlighting its preclinical value [5]. This evidence supports that mitochondrial respiratory chain inhibition plays a major role in HD pathophysiology.
Striatal neurons rely on an afferent supply of brain-derived neurotrophic factor (BDNF) from the cerebral cortex and hippocampus [6]. BDNF is entangled in neuronal differentiation, development, survival, and synaptic plasticity [6]. 3-NP induces neuronal cell death by disrupting mitochondrial membrane potential and temporarily reducing cAMP response element-binding protein (CREB) activity and its target BDNF [7]. Indeed, phosphorylated CREB (p-CREB) binds to the cAMP-response element on DNA to induce BDNF, antiapoptotic genes, detoxifying enzymes, and mitochondrial biogenesis genes [8, 9].
One of the hallmarks of clinical and experimental HD is neuroinflammation [10, 11]. It includes the activation of the innate immune response, mediated by astrocytes and microglia, as a consequence of neuronal injury resulting from cellular insults such as oxidative stress [11, 12]. The HD striatum has been found to include proinflammatory cytokines, such as tumor necrosis factor-α, interleukin (IL)−1β, IL-6, and IL-8, indicating microglial activation [11]. Notably, 3-NP administration triggers neuroinflammation by its ability to induce ROS production [13].
Pyroptosis and ferroptosis are regulated cell death forms recently linked to neuronal loss in HD [14, 15]. Pyroptosis is a distinctive pattern of regulated cell death, owing to its pro-inflammatory and lytic properties. It is usually initiated by inflammasomes, while the execution involves caspases and gasdermins [14, 16]. The inflammasome complex is a multimeric protein assembly essential to innate immunity, produced and activated in neuronal and glial cells. Three elements make up this complex: a sensing receptor, a linker, and procaspase. The sensor module is the NACHT, LRR, and PYD domains-containing protein 3 (NLRP3) protein, which acts as a pattern recognition receptor (PRR), detecting exogenous pathogen-associated molecular patterns (PAMPs) (e.g., pathogens, viruses, and toxins) [14, 16, 17]. Additionally, NLRP3 identifies endogenous damage-associated molecular patterns (DAMPs) that are generated by stressed or injured cells [17]. Upon the assembly of NLRP3 inflammasomes, procaspases are turned into active caspases, which drive the maturation of proinflammatory cytokines such as IL-18 and IL-1β and cleave gasdermins to produce their N-terminal fragments. These fragments form cell membrane pores, resulting in water influx, cell rupture, and inflammatory cytokine release, including IL-18 and IL-1β [14, 16, 17]. The canonical pathway, triggered by PAMPs or DAMPs, mainly involves the activation of caspase-1, which cleaves and activates gasdermin D (GSDMD), while the non-canonical pathway involves caspases 4/5/11 activation via direct binding to lipopolysaccharide [16, 17]. Marked NLRP3 and caspase-1 activation was detected in HD mouse models, shedding light on the pyroptosis function in HD and offering potential novel targets for therapy [14, 18].
Ferroptotic cell death mainly involves intracellular iron accumulation, excessive lipid peroxidation, and reduced glutathione (GSH) depletion. Intracellular iron buildup causes free radicals to develop, which in turn produce lipid peroxides that damage the plasma membrane, ultimately resulting in cell death [19, 20]. Glutathione peroxidase 4 (GPx4) has a vital anti-ferroptotic role, since it is the sole enzyme that can reduce lipid peroxides into lipid alcohols. The detoxification effect of GPx4 requires GSH, which is depleted in ferroptosis due to increased oxidative stress [19]. Evidence suggests a key role of ferroptosis in HD. Toxic iron levels were observed in the neurons of HD animal models and magnetic resonance images of HD patients’ brains [15, 21]. Moreover, increased lipid peroxidation and decreased GSH levels have been noted in HD brains [15], further highlighting ferroptosis inhibition as a therapeutic strategy in HD.
Roflumilast is an FDA-approved drug for severe chronic obstructive pulmonary disease treatment [22]. It acts by selectively inhibiting the phosphodiesterase-4 (PDE-4) enzyme, with milder emetic action than its predecessor, rolipram [22]. PDE-4 is abundant in the brain and controls cAMP levels by degrading it to 5’-AMP [23]. PDE-4 inhibitors prevent this breakdown, sustaining the activation of protein kinase A (PKA) [24]. PKA is the major kinase responsible for CREB phosphorylation and activation to induce target genes, including BDNF [24]. Indeed, prior evidence showed that the prototypic PDE-4 inhibitor rolipram ameliorated intracerebral hemorrhage by increasing cAMP [24]. Roflumilast, a second-generation, lipophilic, and highly permeable PDE-4 inhibitor, has been shown to have a neuroprotective role and might be a new therapeutic option [25–29]. Roflumilast improved the cognitive function in APP/PS1 mice [25] and elicited neuroprotection in several CNS animal models, including quinolinic acid-induced HD [27], cerebral ischemia/reperfusion injury in juvenile rats [28], and rotenone-induced Parkinson’s disease (PD) [29], through its anti-inflammatory and antioxidant effects. Thereby, the repositioning of roflumilast in HD treatment could be a novel strategy, and the full mechanistic profile of its neuroprotective effect, as well as its appropriate regimen (prophylaxis or treatment), remains to be investigated.
In this study, the neuroprotective and therapeutic effects of roflumilast prophylaxis and treatment in rats with 3-NP-induced HD-like neurodegeneration were investigated along with its possible modulatory effects on pyroptotic and ferroptotic neuronal cell deaths, synaptic plasticity, astrogliosis, microglial activation, and neuroinflammation.
Materials and Methods
Drugs and Chemicals
Roflumilast (Westabreath^®^, manufactured by Western Pharmaceutical Industries, Egypt) was suspended in a 0.5% solution of sodium carboxymethyl cellulose (CMC, International Corporation for Scientific and Medical Supplies, Cairo, Egypt) before use. 3-NP (Sigma-Aldrich, St. Louis, MO, USA) was prepared in saline solution prior to use. All chemicals and reagents used were of high analytical grade and were carefully selected to ensure the highest quality and consistency throughout the experiment.
Animals
Adult male rats of the Wistar albino strain, aged 8–10 weeks, with a body weight ranging from 200 to 250 g, were acquired and homed at the animal house of the Faculty of Pharmacy, Cairo University, Cairo, Egypt. For the purpose of acclimatization, rats were originally housed for a week. In the facility, standard polypropylene cages were used to house the rats (five per cage) under a 12/12-hour light/dark cycle, a humidity of 60% ± 10%, and a constant temperature of 25 ± 2 °C. They were also given water and a standard pellet diet (EL Nasr Pharmaceutical Chemicals Co., Cairo, Egypt) ad libitum.
The experimental protocol stuck to the Guide for the Care and Use of Laboratory Animals protocol (NIH Publication No. 85 − 23, revised 2011) executed by the Ethics Committee of Animal Experimentation at the Faculty of Pharmacy, Cairo University, Cairo, Egypt (approval number: BC3524). Every attempt has been performed to decrease the number of rats used in the research.
Experimental Design
Overall, 80 rats were assigned to five groups at random. Each group contained 15 rats, with the exception of the 3-NP group, which had 20 rats to make up for the high mortality caused by 3-NP. To note, the mortality rate in the 3-NP group was 25% during the study period based on our previous work [30]. As depicted in Fig. 1, the experimental design was constructed as follows:
Group 1 (Normal control group): Rats were intraperitoneally injected with saline solution for 7 days and then received CMC (0.5%) orally for 21 days.
Group 2 (roflumilast-control group): Rats were intraperitoneally injected with saline solution for 7 days and then received roflumilast (1 mg/kg/day) suspended in CMC orally for 21 consecutive days [27–29].
Group 3 (3-NP group): Rats were administered 0.5% CMC orally for 21 days, followed by intraperitoneal administration of 3-NP (20 mg/kg/day) dissolved in saline solution for 7 consecutive days [30].
Group 4 (roflumilast-prophylaxis group): Rats were administered roflumilast (1 mg/kg/day) suspended in CMC orally for 21 days [27–29], and then starting from the 22nd day, they were intraperitoneally injected with 3-NP (20 mg/kg/day) dissolved in saline solution for 7 days, as in group 3.
Group 5 (roflumilast-treatment group): Rats were administered 3-NP injections (20 mg/kg/day) dissolved in saline solution intraperitoneally for 7 days, as in group 3, and starting from the 8th day, they received roflumilast (1 mg/kg/day) suspended in CMC orally for 21 days [27–29].
Fig. 1. Experimental design timeline. CMC Carboxymethyl cellulose, 3-NP 3-nitropropionic acid, ROF roflumilast
The 3-NP intraperitoneal subchronic administration dosage, route, and regimen were based on a previous study to phenotypically induce HD-like symptoms [30]. Roflumilast oral dose (1 mg/kg/day) and regimen were according to previous reports in CNS animal models [27–29]. Specifically, the roflumilast daily dose 1 mg/kg was superior to 0.5 mg/kg across biochemical and histological markers tested in cerebral ischemia/reperfusion injury in juvenile rats [28]. Moreover, upon comparing 0.3 and 1 mg/kg doses for 4 weeks, the higher dose of roflumilast showed aprofound neuroprotective effect in a PD rat model [29]. However, in a study comparing roflumilast doses, 1 versus 2 mg/kg/day for 21 days, both doses showed comparable neuroprotective efficacy in the quinolinic acid-induced rat model of HD [27]. Thereby, the 1 mg/kg/day dose choice was based on selecting a dose having higher effectiveness with lower side effects. The duration of roflumilast administration (21 days) was selected based on another chemical HD model [27]. The behavioral, biochemical, and histopathological alterations in rats were then demonstrated.
Behavioral and Functional Assessment
Behavioral tests were performed 24 h after the last dose, with animal handling, timing, testing conditions, and scoring criteria were standardized across all groups. The least stressful behavioral test was performed first, followed by the more stressful test, with a 30-minute gap between each test. All rats within groups were evaluated for their motor capability using 3 tests beginning with the open-field, then the rotarod, and finally the grip strength test. Every behavioral test was conducted under dim white light during the rat’s light cycle in a sound-insulated environment.
Open-Field Test (OFT)
The spontaneous locomotor activity of rats was evaluated using the OFT. A square wooden box (80 × 80 × 40 cm) with a floor divided by black lines into a 4 × 4 grid of 16 equal squares was used to conduct the test. After being placed in the open-field centre, each rat was left to freely move in the box, and its locomotor behavior was recorded for 3 min using a video camera. After removing each rat, the box’s floor and walls were cleaned using ethanol to get rid of any residual odor. Each rat was assessed for the rearing and ambulation frequencies (vertical movement and the number of line crossings, respectively) and the latency time (time passed with no movement) [31].
Rotarod Test
The rats’ motor coordination was evaluated using rotarod apparatus. The apparatus is a metallic rod of 3 cm in diameter, 120 cm in length, and rotates with a constant speed of 20 rpm. Each animal was trained on the apparatus on three separate days for 5 min. For each rat, the latency of fall was recorded in test duration of 5 min [32].
Grip Strength Test
The neuromuscular functions of rats were measured when they grasped the grip bar of the grip strength meter (Ugo Basile, Italy) using their forelimbs. Rats were then pulled straight back in a slow manner until they released their grip, and the force was measured for each rat using the meter. For each rat, the test was performed 3 times and the average value was used in statistical comparisons [33].
Biochemical Analysis
24 h after the performance of behavioral assessment, each group was subdivided into three sets of rats. They were decapitated while under light anaesthesia (3% isoflurane inhalation). The brain was promptly separated from each rat and rinsed with ice-cold saline. For the purposes of histological and immunohistochemical analyses, the entire brains of the first set (n = 3) were preserved in 10% (v/v) formalin for 24 h. From the second set of rats (n = 6), the striata were dissected from the brain of each rat and homogenized in phosphate-buffered saline (PBS) for use in colorimetric and enzyme-linked immunosorbent assay (ELISA). Reverse transcriptase-quantitative polymerase chain reaction (RT-qPCR) and western immunoblotting were performed on the striata of the third set (n = 6).
Colorimetric Assays
10% striatal tissue homogenates were prepared in PBS. Malondialdehyde (MDA) and GSH levels were biochemically assayed using colorimetric kits procured from Biodignostics, Egypt (Cat#: MD 2529 and GR 2511), as previously described [34, 35]. Neuronal iron levels were measured using the Iron Colorimetric Assay kit (BioVision, USA, Cat#: K390-100) per the vendor’s instructions. We used the Bradford method to quantitatively assess the protein concentration of the tissue homogenates [36], and the findings are presented per milligram protein.
ELISA Measurements
The striatal protein levels of total CREB (t-CREB), NLRP3, caspase-1, and GPx4 were assessed using rat ELISA kits (MyBioSource, China, Cat#: MBS2504589, MBS7255410, MBS765838, and MBS069787) based on the manufacturer’s recommendations. Rat ELISA kits supplied by Cloud-Clone Corp., USA, were employed to measure the striatal BDNF and IL-1β protein levels (Cat#: SEA011Ra and SEA563Ra), while the p-CREB levels were measured using the Duoset rat ELISA kit (Cat#: DYC2510), following the vendor’s instructions. The results are shown per milligram protein.
RT-qPCR Assay
The RT-qPCR technique was utilized for the assay of PDE-4 expression. Briefly, for RNA extraction, the striatal tissue lysate was treated with Direct-zol RNA Miniprep Plus (Cat#: R2072, Zymo Research Corp., USA). Using a Beckman dual spectrophotometer (USA), the amount and quality of RNA were evaluated. Following the manufacturer’s instructions, the SuperScript IV One-Step RT-PCR kit (Cat#: 12594100, Thermo Fisher Scientific, Waltham, MA, USA) was used to perform reverse transcription and real-time PCR in a single step using the Step One Applied Biosystem apparatus (Foster City, CA, USA). The primers’ sequences used were GAPDH: F: 5′-TGACAACTCCCTCAAGATTGTCA-3’, R: 5′-GGCATGGACTGTGGTCATGA-3’ and PDE-4: F: 5′-CGTCAGTGCTGGACAGTC-3’, R: 5′-CCAGCGTACTCCGACACACA-3’. Data were given in cycle threshold (Ct) following the qPCR experiments. Using the delta-delta Ct (ΔΔCt) approach, the relative quantification of PDE-4 was calculated after normalization with GAPDH as a housekeeping gene.
Western Blotting
Protein extraction and purification were conducted using TriFast (Peqlab, VWR Company, USA) per the manufacturer’s instructions. Following the Bradford method, the protein fraction concentration was measured. The PageRuler™ Unstained Broad Range Protein Ladder (Thermo Scientific™, Cat#: 26630) was employed. After the gel was polymerized, 30 µg of proteins were loaded on SDS-PAGE and electrophoresed. The gel was then transferred to a Hybond*™* nylon membrane (GE Healthcare) and then incubated for 1 h at room temperature in the blocking solution (2–5% non-fat dry milk in blotting buffer, pH = 7.4). β-actin was employed as a housekeeping protein. Anti-GSDMD antibody [EPR20859] (ab219800, dilution 1:1000) and anti-β-actin antibody [mAbcam 8226]-loading control (ab8226, dilution 1:1000) were used for western blot detection. Overnight incubation at 4 °C was performed in a primary antibody solution. After washing, the membrane was incubated with the HRP-conjugated secondary antibody at room temperature for 1 h. The gel documentation system (Geldoc-it, UVP, England) was applied for data analysis using Totallab analysis software v1.0.1 (www.totallab.com).
Histopathological Examination
Formalin-preserved intact brains of three rats in each group were embedded in paraffin blocks after being dehydrated in a graded alcohol series and cleared with xylene. After cutting the samples into 4-µm thickness by rotary microtome, they were stained with hematoxylin and eosin (H&E) to demonstrate the general morphology and with Nissl toluidine blue dye to demonstrate the intact striatal neuron count. Histopathological examination was performed by an expert histopathologist blinded to the animals’ grouping employing a full HD microscopic imaging system and Leica Application Software (Leica Microsystems GmbH, Wetzlar, Germany). For each sample, at least five representative, non-overlapping fields per tissue section were randomly selected and examined.
Immunohistochemical Examination
Sections of paraffin-embedded tissue, each 5 μm thick, were cut out on adhesive slides, deparaffinized, and then rehydrated with distilled water. After heat-induced epitope retrieval, tissue sections were incubated for 1 h at room temperature with primary anti-glial fibrillary acidic protein antibody (anti-GFAP, abx000622, Abbexa, Cambridge, UK, dilution 1:200) and anti-ionized calcium binding adaptor molecule-1 antibody (anti-Iba-1, 10904-1-A, Proteintech, Germany, dilution 1:400). Following washing, peroxidase blocking was performed, and then, per the manufacturer’s instructions, detection was done using an HRP-labeled secondary detection kit (BioSB, USA). For quality control, negative control slides prepared without incubating with the primary antibody were examined. For each group, the mean area percentage of positive staining in the striatum was used to quantitatively measure positive expression. A skilled histopathologist conducted the examination using a full HD microscope with Leica Application Software (Leica Microsystems GmbH, Wetzlar, Germany) while blinded to the identities of the groups. Five representative, non-overlapping fields per section were examined.
Functional Enrichment Analysis
Construction of the protein-protein interaction (PPI) network of the studied genes and functional enrichment analysis of gene ontology (GO) biological process and Reactome pathways were performed using the STRING database (https://string-db.org/).
Statistical Analysis
Using the G*Power software 3.1.9.7, the group size was calculated following a power analysis (power = 0.8, α = 0.05). The results are presented as mean ± standard deviation (SD) or standard error of the mean (SEM). To ascertain if the values in each group had a normal distribution, the Shapiro-Wilk test was employed. To examine for outliers, Dixon’s Q test was utilized (outliers were identified at P < 0.05 and removed). Data were compared using one-way ANOVA with Tukey’s post-hoc test, except for the rotarod latency of fall, where the Kruskal-Wallis and Dunn’s multiple comparison tests were used. The Tukey’s HSD post-hoc analysis was used to generate multiplicity adjusted P-values for pairwise comparisons. Effect sizes between groups were calculated for behavioral tests and presented as partial eta-squared (η^2^). Analysis of correlations between parameters was performed using Pearson correlation. The GraphPad Prism software 8.4.2 (GraphPad Software Inc., San Diego, USA) was employed to conduct the statistical analyses. A two-tailed P-value < 0.05 was deemed statistically significant.
Results
Roflumilast Improves Behavioral and Functional Abnormalities in Rats
The OFT showed the derangements of locomotor activity of rats induced by 3-NP. As depicted in Fig. 2A and B, the ambulation and rearing frequencies of rats in the 3-NP group were markedly decreased by 70.80% and 80.41%, respectively, compared to the normal control group. On the other hand, the ambulation and rearing frequencies were significantly improved by the administration of roflumilast in the prophylaxis group, increasing them by 207.82% and 305.04%, respectively, compared to the 3-NP group. In the treatment group, roflumilast increased ambulation and rearing frequencies by 186.33% and 278.23%, respectively, compared to the 3-NP group.
Fig. 2. Effect of roflumilast on behavioral and motor abnormalities in the 3-NP-injected rats. A OFT ambulation frequency [F (4, 66) = 33.93, P < 0.0001], η^2^ = 0.673 (large), B OFT rearing frequency [F (4, 66) = 35.92, P < 0.0001], η^2^ = 0.685 (large), C OFT latency time [F (4, 65) = 75.65, P < 0.0001], η^2^ = 0.823 (large), D Representative track plots of the OFT for each group, E Rotarod latency of fall, Kruskal-Wallis statistic H = 40.87, P < 0.0001, η^2^ = 0.55 (large), and F Grip strength [F (4, 66) = 28.89, P < 0.0001], η^2^ = 0.637 (large). Data are expressed as mean ± SD, n = 15. ANOVA followed by Tukey’s post-hoc test was used for OFT and grip strength data analyses. Analysis of the rotarod test results were done using the Kruskal-Wallis test followed by Dunn’s test. ^a^significant difference from the normal control group, ^b^significant difference from the 3-NP group. Significance level was set at P < 0.05. 3-NP, 3-nitropropionic acid; ROF, roflumilast
In the 3-NP group, the latency time (Fig. 2C) was significantly increased by 2.72-fold compared to the normal control group, while it was normalized by roflumilast administration in both the prophylaxis and treatment groups, rescuing the locomotor behavior of rats. Figure 2D portrays representative track plots of rats from different groups in the OFT.
The rotarod latency of fall (Fig. 2E) exhibited a marked decline (almost 8-fold) in the 3-NP-intoxicated rats as compared to the normal control group. Remarkably, when compared to the 3-NP group, it was significantly higher in the roflumilast prophylaxis and treatment groups by 595.06% and 573.42%, respectively, showing a marked enhancement in the rats’ motor coordination.
The grip strength test disseminated the 3-NP negative effects on the rats’ neuromuscular function (Fig. 2F). This was indicated by a significant decrease in the grip strength by 37.83% compared to the normal control rats. The roflumilast prophylaxis and treatment groups exhibited a significant increment in the grip strength at the point of normalization, demonstrating its positive impact on the rats’ neuromuscular function.
To note, there were no statistical differences in the assessed behavioral parameters between the roflumilast-control group and the normal control group (P > 0.05). In addition, the two groups were not statistically different in all the tested biochemical, histopathological, and immunohistochemical investigations. These findings suggest the safety of roflumilast and the absence of oxidative or inflammatory perturbations associated with its intake under normal physiologic conditions. To enhance conciseness, the roflumilast-control group versus normal control group comparison will not be further discussed.
Roflumilast Attenuates Striatal PDE-4 Expression and Augments p-CREB/BDNF Levels in 3-NP-Injected Rats
As displayed in Fig. 3A, the striatal expression of PDE-4 was significantly increased by approximately sixfold post-3-NP injection compared with the normal control values (P < 0.0001). The administration of roflumilast significantly reduced striatal PDE-4 expression by 29.36% and 55.39% in the prophylaxis and treatment groups, respectively, compared to the 3-NP group (P < 0.0001 for each). Interestingly, the roflumilast-treated rats showcased a more prominent reduction of PDE-4 expression compared to the prophylaxis group (P < 0.0001).
Fig. 3. Effect of roflumilast on PDE-4 expression, CREB, and BDNF levels in the striatum of 3-NP-injected rats. A PDE-4 [F (4, 24) = 269.3, P < 0.0001], B Total CREB [F (4, 25) = 2.424, P = 0.075], C p-CREB [F (4, 25) = 2258, P < 0.0001], D p-CREB/t-CREB [F (4, 25) = 1072, P < 0.0001], E BDNF [F (4, 25) = 1005, P < 0.0001]. Data are expressed as mean ± SD, n = 6. ANOVA followed by Tukey’s post-hoc test was used for data analysis. ^a^significant difference from the normal control group, ^b^significant difference from the 3-NP group, ^c^significant difference from the ROF-prophylaxis group. Significance level was set at P < 0.05. BDNF, brain-derived neurotrophic factor; p-CREB, phosphorylated-cAMP response element-binding protein; t-CREB, total CREB; 3-NP, 3-nitropropionic acid; PDE-4, phosphodiesterase-4; ROF, roflumilast
As depicted in Fig. 3B–E, 3-NP substantially affected the synaptic plasticity of rats. This was evidenced by the significant reduction in the striatal levels of p-CREB, p-CREB/t-CREB ratio, and BDNF by 77.88%, 78.49%, and 69.66%, respectively, in the 3-NP group compared to the normal control group (P < 0.0001 for each). Roflumilast significantly augmented the p-CREB, p-CREB/t-CREB, and BDNF levels in both the prophylaxis and treatment groups in comparison with the 3-NP alone group. These findings indicate its neuroprotective and neurorestorative actions, with more prominent effects appeared in the treatment group (P < 0.0001). Notably, p-CREB, p-CREB/t-CREB, and BDNF levels were elevated by 87.68%, 80.70%, 77.34%, respectively, in the roflumilast-prophylaxis group, while they were significantly increased by 210.70%, 201.50%, 181.60%, respectively, in the roflumilast-treatment group compared to the 3-NP alone group (P < 0.0001 for each). These results show the marked neurorestorative effects of roflumilast. To note, t-CREB levels were not significantly different across all groups (P = 0.075).
Roflumilast Mitigates Striatal NLRP3, Caspase-1, and GSDMD Pyroptotic Markers and IL-1β Proinflammatory Cytokine in 3-NP-Injected Rats
Pyroptosis was examined by measuring the NLRP3 inflammasome and caspase-1 using ELISA, and GSDMD, the final effector of pyroptosis, using western blotting. The impact of 3-NP on striatal pyroptosis was shown by the significant elevation of NLRP3 (Fig. 4A), caspase-1 (Fig. 4B), and GSDMD (Fig. 4C and D) levels in the 3-NP group by 501.63%, 366.83%, and 306.45%, respectively, compared with the normal control values (P < 0.0001 for each). Roflumilast, in both prophylaxis and treatment regimens, significantly attenuated the three tested pyroptotic markers, indicating pyroptosis inhibition. Compared to the 3-NP group, the levels of NLRP3, caspase-1, and GSDMD were decreased by 42.67%, 56.83%, 55.56%, respectively, in the roflumilast-treatment group relative to a decline of 24.96%, 29.61% and 34.20%, respectively, in the prophylaxis group (P < 0.0001 for each). These findings highlight the more profound effect of roflumilast treatment.
Fig. 4. Effect of roflumilast on the pyroptotic indicators NLRP3, caspase-1, and GSDMD and the proinflammatory cytokine IL-1β in the striatum of 3-NP-injected rats. A NLRP3 [F (4, 24) = 322.7, P < 0.0001], B Caspase-1 [F (4, 25) = 931.3, P < 0.0001], C GSDMD/β-actin [F (4, 24) = 252.0, P < 0.0001], D Representative western blot of GSDMD protein expression level, and E IL-1β [F (4, 24) = 1314, P < 0.0001]. Data are expressed as mean ± SD, n = 6. ANOVA followed by Tukey’s post-hoc test was used for data analysis. ^a^significant difference from the normal control group, ^b^significant difference from the 3-NP group, ^c^significant difference from the ROF-prophylaxis group. Significance level was set at P < 0.05. GSDMD, gasdermin D; IL-1β, interleukin-1β; NLRP3, NACHT, LRR, and PYD domains-containing protein 3; 3-NP, 3-nitropropionic acid; ROF, roflumilast
Concerning the inflammatory response, IL-1β was examined to indicate the striatum’s innate immune response to the neurotoxicity invoked by 3-NP. Striatal IL-1β levels were markedly increased by 746.38% in the 3-NP group compared with levels in the normal control group (P < 0.0001), indicating significant neuroinflammation (Fig. 4E). Roflumilast administration significantly ameliorated the neuroinflammatory status, as shown by the significant reduction of IL-1β in both prophylaxis and treatment groups by 21.74% and 61.39%, respectively, compared to the 3-NP group (P < 0.0001 for each). This finding significantly demonstrates the more profound neurorestorative impact of roflumilast treatment over that of prophylaxis (P < 0.0001).
Roflumilast Modulates Striatal iron, MDA, GPx4, and GSH Ferroptotic Markers in 3-NP-injected Rats
Ferroptosis was evaluated by measuring iron, lipid peroxidation, GPx4, and GSH levels. As displayed in Fig. 5A, 3-NP induced a state of ferroptosis through the elevation of striatal iron levels by almost threefold compared with levels in the normal control group (P < 0.0001). Subsequently, this was followed by a substantial elevation of striatal MDA levels (Fig. 5B) by 5.52-fold compared to the control group (P < 0.0001), signifying a marked lipid peroxidation. Consequently, the antioxidant defensive mechanisms were depleted in the striatum of 3-NP-intoxicated rats, as evidenced by a significant decline in GPx4 (Fig. 5C) and GSH (Fig. 5D) levels reaching 18.38% and 20.92% of the normal control values (P < 0.0001 for each).
Fig. 5. Effect of roflumilast on the ferroptotic indicators iron, MDA, GPx4, and GSH in the striatum of 3-NP injected rats. A Iron [F (4, 24) = 329.1, P < 0.0001], B MDA [F (4, 25) = 403.9, P < 0.0001], C GPx4 [F (4, 25) = 811.2, P < 0.0001], D GSH [F (4, 24) = 386.7, P < 0.0001]. Data are expressed as mean ± SD, n = 6. ANOVA followed by Tukey’s post-hoc test was used for data analysis. ^a^significant difference from the normal control group, ^b^significant difference from the 3-NP group, ^c^significant difference from the ROF-prophylaxis group. Significance level was set at P < 0.05. GPx4, glutathione peroxidase 4; GSH, reduced glutathione; MDA, malondialdehyde; 3-NP, 3-nitropropionic acid; ROF, roflumilast
Upon administration of roflumilast in either prophylaxis or treatment, the ferroptotic state was reduced, indicating its anti-ferroptotic action. The results supported this claim, showing a significant decline in striatal iron levels by 26.60% and 41.97% in the roflumilast prophylaxis and treatment groups, respectively, compared to the 3-NP group (P < 0.0001 for each). In parallel, MDA levels were reduced by 32.56% and 50.32% with both regimens, respectively, compared with the 3-NP group levels. Meanwhile, roflumilast prophylaxis averted the 3-NP actions on striatal GPx4 and GSH levels, showing a 3.91-fold and a 2.82-fold elevation of GPx4 and GSH levels, respectively, compared to the 3-NP group (P < 0.0001 for each). Similarly, treatment with roflumilast raised the GPx4 and GSH levels by 173.03% and 233.96%, respectively, in comparison with the 3-NP group levels (P < 0.0001 for each).
Roflumilast Improves Histopathological Derangements and Increases Neuronal Survival in the Striatum of 3-NP-Injected Rats
Figure 6 represents the photomicrographs of the H&E stained sections of the rat striatum. Rats in the normal control group demonstrated normal morphological structure of the striatum, showing intact neurons (Fig. 6A). Rats in the roflumilast-control group reserved the normal morphology of the striatum as well, confirming the drug safety (Fig. 6B). However, the 3-NP group showed excessive vacuolation and degeneration of the neurons in the striatum with diffused gliosis in several examined sections associated with perivascular edema (Fig. 6C). Both the roflumilast prophylaxis (Fig. 6D) and treatment (Fig. 6E) groups showed marked improvement, characterized by mild gliosis and fewer degenerated neurons, with apparently normal neurons in several examined animals. Comparable results were observed in both intervention groups.
Fig. 6. Comparative photomicrographs of rat striatum using H&E stain. A Control group showing normal striatum. B ROF-control group showing normal striatum. C 3-NP group showing numerous vacuolated and degenerated neurons in the striatum with perivascular edema (arrow). D ROF-prophylaxis showing mild gliosis in the striatum. E ROF-treatment showing mild gliosis in the striatum (arrow). 3-nitropropionic acid; ROF, roflumilast
The Nissl-stained sections of the striatum (Fig. 7) show the detectable intact neuron count, indicating neuronal survival. The normal control and roflumilast-control groups showed normal, intact, lightly stained neurons in the striatum (Fig. 7A and B, respectively). However, the 3-NP group showed numerous dark-stained neurons, signifying neuronal death (Fig. 7C). Prophylaxis (Fig. 7D) and treatment (Fig. 7E) groups showed fewer dark neurons and more intact neurons when compared to the 3-NP group. As depicted in Fig. 7F, the neuronal survival rate was significantly decreased in the 3-NP group by 70.37% compared to the normal control group, presenting the loss of intact neurons. The neuronal survival rate was significantly increased in both the roflumilast prophylaxis and treatment groups by 209.72% and 211.11%, respectively, compared with that in the 3-NP group.
Fig. 7. Comparative photomicrographs of rat striatum using the Nissl toluidine blue stain. A Control group showing normal striatum. B ROF-control group showing normal striatum. C 3-NP group showing numerous dark neurons in the striatum. D ROF-prophylaxis showing few dark neurons in the striatum. E ROF-treatment showing few dark neurons with apparently normal striatum. F The chart represents neuronal survival rate in the striatum [F (4, 20) = 1017, P < 0.0001], presented as mean ± SEM and analyzed using ANOVA followed by Tukey’s post-hoc test. ^a^significant difference from the normal control group, ^b^significant difference from the 3-NP group. Significant difference was set at P ˂ 0.05. 3-nitropropionic acid; ROF, roflumilast
Roflumilast Obliterates Striatal Astrogliosis and Microglial Activation in 3-NP-Injected Rats
GFAP was evaluated as a marker of astrocyte activation. Figure 8 portrays the immunostaining of GFAP (as GFAP area %) in the studied groups. The control and roflumilast-control rats showed low GFAP expression in the striatum (Fig. 8A and B, respectively). However, significant elevation of GFAP immunostaining was observed in the 3-NP group, indicating astrocyte activation (Fig. 8C). Both the roflumilast prophylaxis (Fig. 8D) and treatment (Fig. 8E) groups showed a noticeably lower GFAP expression compared to the 3-NP group. Quantitative analysis shown in Fig. 8F revealed that the 3-NP-intoxicated rats exhibited marked astrogliosis, as evident by an elevated GFAP area % by 381.89% compared to the normal control group. Remarkably, roflumilast mitigated the astrocyte activation, as shown by the significant decrease in GFAP area % in both the prophylactic and treatment groups by 47.4% and 59.6%, respectively, compared to the 3-NP group, demonstrating the more profound impact of roflumilast treatment.
Fig. 8. Comparative photomicrographs of rat striatum using GFAP immunostaining. A Control group showing lower GFAP expression in the striatum. B ROF-control group showing lower GFAP expression in the striatum. C 3-NP group showing higher GFAP expression in the striatum. D ROF-prophylaxis showing moderate GFAP expression in the striatum. E ROF-treatment showing moderate GFAP expression in the striatum. F The chart represents quantification of GFAP expression as area percentage [F (4, 20) = 199.3, P < 0.0001] presented as mean ± SEM and analyzed using ANOVA followed by Tukey’s post-hoc test. ^a^significant difference from the normal control group, ^b^significant difference from the 3-NP group, ^c^significant difference from the ROF-prophylaxis group. Significant difference was set at P ˂ 0.05. GFAP, glial fibrillary acidic protein; 3-nitropropionic acid; ROF, roflumilast
Figure 9 shows the immunohistochemical staining of the microglial activation marker Iba-1 in the five studied groups. The control and roflumilast-control rats showed low Iba-1 immunostaining in the striatum (Fig. 9A and B, respectively). The 3-NP-injected rats exhibited reactive gliosis, as evidenced by the escalation of Iba-1 immunostaining (Fig. 9C). However, a significant decrease in Iba-1 expression was observed in the roflumilast prophylaxis (Fig. 9D) and treatment (Fig. 9E) groups compared to the 3-NP group. Quantitatively, the Iba-1 area % was elevated by almost fivefold in the 3-NP-injected rats compared with the values of the normal control rats. Meanwhile, roflumilast comparably obliterated microglial activation in both the prophylaxis and treatment groups by moderating the Iba-1 area % by 46.61% and 51.30%, respectively, against the 3-NP group (Fig. 9F).
Fig. 9. Comparative photomicrographs of rat striatum using Iba-1 immunostaining. A Control group showing lower Iba-1 expression in the striatum. B ROF-control group showing lower Iba-1 expression in the striatum. C 3-NP group showing higher Iba-1 expression in the striatum. D ROF-prophylaxis showing moderate Iba-1 expression in the striatum. E ROF-treatment showing moderate Iba-1 expression in the striatum. F The chart represents quantification of Iba-1 expression as area percentage [F (4, 20) = 57.84, P < 0.0001] presented as mean ± SEM and analyzed using ANOVA followed by Tukey’s post-hoc test. ^a^significant difference from the normal control group, ^b^significant difference from the 3-NP group. Significant difference is considered at P ˂ 0.05. Iba-1, ionized calcium binding adaptor molecule-1; 3-nitropropionic acid; ROF, roflumilast
Correlations Exist Between Functional Output and Cell Death Markers
As a secondary outcome, we conducted an exploratory correlation analysis between biochemical markers with each other and with behavioral outcomes using data values of different parameters analyzed from the same rats. We found significant correlations between the behavioral/functional outcomes and pyroptotic, ferroptotic, and synaptic plasticity markers (P < 0.05) (Fig. 10). The ambulation and rearing frequencies in the OFT, latency of fall in the rotarod, and grip strength were negatively correlated with striatal NLRP3, caspase-1, IL-1β, iron, and MDA, while positively correlated with GPx4, GSH, BDNF, and p-CREB. In the contrary, the OFT latency time showed the reverse pattern of correlations.
Fig. 10. Correlation between behavioral/functional output and pyroptotic, ferroptotic, and synaptic plasticity markers in the striatum of 3-NP-injected rats. A correlation heatmap with a red-green scale generated using https://www.chiplot.online/correlation_heatmap.html. Values of different parameters were taken from the same rats (6 rats/group, total n = 30). Correlations were computed using Pearson correlation. The Pearson correlation coefficient is presented inside the cells of the heatmap. Positive and negative correlations are represented by red and green colors, respectively. The correlation coefficients range from − 1 to + 1, with values close to ± 1 indicating strong relationships, while values close to zero indicate weak relationships. P-values less than 0.05 were considered statistically significant. AF, ambulation frequency; BDNF, brain-derived neurotrophic factor; p-CREB, phosphorylated-cAMP response element-binding protein; GPx4, glutathione peroxidase 4; GSH, reduced glutathione; IL-1beta, interleukin-1β; LF, latency of fall; LT, latency time; MDA, malondialdehyde; NLRP3, NACHT, LRR, and PYD domains-containing protein 3; OFT, open-field test; RF, rearing frequency
Notably, there were strong correlations between pyroptotic and ferroptotic markers (P < 0.0001). Specifically, striatal NLRP3, caspase-1, and IL-1β were positively correlated with iron and MDA and negatively correlated with GPx4 and GSH. The Pearson correlation coefficients are given in Fig. 10.
Visualization of the PPI Network of the Studied Genes
In a secondary analysis, we displayed the intricate molecular interactions of the differentially expressed target genes through constructing their PPI network using the STRING database. As depicted in Fig. 11, functional enrichment analysis revealed that the PPI network (PDE-4B, CREB/BDNF, NLRP3/caspase-1/GSDMD, IL-1β, and GPx4) is involved in IL-1 processing and signaling, pyroptosis, innate immune response, cytokine production, glial cell activation, chronic inflammatory response, and negative regulation of neural precursor cell proliferation. Based on the observed modulatory effects of roflumilast, these exploratory PPI results might suggest its multi-pathway modulating potential.
Fig. 11. Protein-protein interaction network of the studied genes. A PPI network constructed using STRING database, the PPI enrichment P-value = 1.43e-08, number of proteins = 8, number of nodes = 8, number of edges (representing protein-protein associations) = 15, and average local clustering coefficient = 0.789. B Reactome pathways enrichment. C Results of the GO biological process. The available results or the first 11 results at maximum are visualized according to groups of similarity = 1, highest strength, and FDR. BDNF, brain-derived neurotrophic factor; Casp1, caspase 1; CREB, cAMP response element-binding protein; FDR, false discovery rate; GPx4, glutathione peroxidase 4; GSDMD, gasdermin D; IL-1b, interleukin-1β; NLRP3, NACHT, LRR, and PYD domains-containing protein 3; PDE-4b, phosphodiesterase-4B
Discussion
This study evaluated the possibility of using the PDE-4 inhibitor roflumilast as a new therapeutic agent for HD in the 3-NP experimental rat model. Roflumilast has a high safety profile [37]. It also shows advantageous pharmacokinetics such as selective inhibition of PDE-4B and PDE-4D subtypes and efficient blood-brain barrier (BBB) penetration. These characteristics, which increase neuroplasticity and decrease microglial activation, make roflumilast a promising candidate for repurposing in neurological diseases [38–40].
To our knowledge, this is the first study to demonstrate the dual prophylactic and therapeutic effects of roflumilast in an experimental model of HD via concurrent modulation of pyroptosis, ferroptosis, and neuroinflammation. These findings highlight the multifaceted neuroprotective potential of roflumilast and offer new insights into its role in targeting emerging pathogenic mechanisms in HD. Specifically, this study showed the positive impact of roflumilast on neuronal cell death by attenuating both ferroptosis and pyroptosis in HD-like neurodegeneration. These changes were reflected in the improvement of the behavioral and functional impediments inflicted by 3-NP. Notably, in an exploratory analysis, we found highly significant correlations between the pyroptotic and ferroptotic indicators and the results of behavioral tests.
Using a battery of behavioral tests, we showed that roflumilast improved the behavioral and motor signs characteristic of HD-like neurodegeneration. Furthermore, roflumilast, administered prophylactically or therapeutically, ameliorated the histopathological and immunohistochemical changes caused by 3-NP. Moreover, the biochemical analyses performed indicated that roflumilast lessened PDE-4 gene expression and augmented p-CREB/BDNF levels in the rats’ striatum. In parallel, mitigation of pyroptotic and ferroptotic markers was shown in both roflumilast-prophylactic and treated rats, providing evidence of its anti-pyroptotic and anti-ferroptotic effects. Roflumilast also exerted an anti-neuroinflammatory effect by alleviating astrocyte and microglial activation and decreasing IL-1β levels in the striatum. Further in-depth investigations are required to validate the roflumilast dual modulation of pyroptosis and ferroptosis.
The 3-NP mitochondrial neurotoxin model is valuable in drug discovery and development in HD preclinical studies [5, 41]. Mitochondrial dysfunction is greatly associated with HD pathogenesis. Indeed, mHtt interferes with the mitochondrial processes, disrupting the respiratory chain and promoting excessive ROS production and consequent neuronal degeneration in HD patients [42, 43]. Moreover, metabolic energy deficits have been documented in HD patients, and decreased activities of the respiratory chain complexes II, III, and IV have been shown in postmortem brain tissue samples of HD [43]. A credible model of HD motor and cognitive symptoms is also provided by 3-NP administration, which results in a specific degeneration of GABAergic medium-sized spiny neurones with excitotoxic-like lesions of the striatum [5, 41]. Here, repeated administration of 3-NP significantly deteriorated the locomotion performance, the motor coordination, and the grasping power of rats, as previously reported [30, 44]. Our results revealed that roflumilast administration, either as a preventative agent or a therapy, dramatically improved the behavioral and motor function of 3-NP-injected rats.
In this study, striatal PDE-4 upregulation was observed in the 3-NP-intoxicated rats. Mechanistically, PDE-4 elevation reduces cAMP levels in the striatum, with a consequent inactivation of PKA/p-CREB/BDNF axis, ultimately disrupting the synaptic plasticity [27]. Besides PDE-4 inhibition, our findings disclosed that roflumilast attenuated striatal PDE-4 mRNA expression. These results agree with previous reports that roflumilast treatment attenuated striatal PDE-4 mRNA expression in quinolinic acid-induced HD [27]. An additional finding of our study is that roflumilast mitigated PDE-4 expression in both therapeutic and prophylactic regimens in the 3-NP model, with more profound therapeutic effects.
Moreover, the 3-NP-intoxicated rats showed a decrease in synaptic plasticity, as indicated by the decline in striatal BDNF and p-CREB, mimicking the pathological alterations associated with HD [6]. These findings mirror those of an earlier report [7]. BDNF is a neurotrophic protein induced by the transcriptional factor p-CREB to enhance neuronal survival and maintain synaptic plasticity [45]. This neuroprotective action is mainly achieved via BDNF binding to tropomyosin receptor kinase B (TrkB). This interaction leads to the activation of multiple signaling pathways that ultimately promote neuronal proliferation, plasticity, and survival [46]. In our study, roflumilast administration exerted neuroprotective and restorative effects, evidenced by the augmented p-CREB/BDNF axis in both regimens. These findings agree with earlier research demonstrating the neuroprotective properties of roflumilast in other brain models by increasing synaptic plasticity [25, 28, 47]. These results could be attributed to the fact that roflumilast-mediated PDE-4 inhibition enhances the cellular levels of cAMP, which subsequently activate the PKA/p-CREB/BDNF axis [27]. Indeed, roflumilast preferentially decreases PDE-4B and 4D isoforms with an increase in p-CREB/BDNF expression [26]. Notably, PDE-4D is primarily involved in neuronal plasticity [23, 39]. Interestingly, the increase in p-CREB/BDNF in our study was more prominent with roflumilast treatment, following the more significant PDE-4 downregulation relative to prophylaxis.
Although the inflammasome is a part of the innate immune response in neuronal and glial cells, its activation could have devastating consequences, culminating in neurodegenerative diseases [14, 16]. NLRP3 inflammasome activation by DAMPs, such as the 3-NP-trigered ROS, mediates caspase-1 activation [17]. The latter causes the activation of pro-IL-18 and pro-IL-1β into their mature forms. Caspase-1 also activates GSDMD via cleaving the inhibitory C-terminal fragment, while releasing the active N-terminal fragment, which forms pores in the cell membrane, allowing the release of mature IL-18 and IL-1β. These pores drive water influx, leading to cell swelling, lytic cell rupture, and death [14, 16, 17].
In our study, 3-NP administration was associated with NLRP3 inflammasome-mediated pyroptosis in striatal cells, as evident by the augmented NLRP3/caspase-1/GSDMD axis. Similarly, previous studies showed increased NLRP3 and caspase-1 in a transgenic HD mouse model and in the substantia nigra of a PD mouse model [48, 49]. The observed increase in striatal GSDMD following 3-NP-induced neurotoxicity was similar to previous results in a PD mouse model [50], further highlighting the implication of GSDMD and pyroptosis in the pathogenesis of neurodegenerative diseases. Interestingly, roflumilast administration in either prophylaxis or treatment mitigated pyroptosis, as evidenced by the remarkable decline in NLRP3 and caspase-1. However, more significant results were shown in the treatment group. Relatedly, our findings agree with those of a recent study showing the impactful action of roflumilast on alleviating chemotherapy-induced cognitive impairment by attenuating the NLRP3/ASC/caspase-1/GSDMD trajectory [51]. The latter study showed a dose-dependent action of roflumilast. The higher dose (1 mg/kg/day) demonstrated superior neuroprotective effects [51], further supporting the dose selection in our study. Altogether, these results highlight that roflumilast neuroprotection is associated with inhibiting pyroptotic cell death.
The substantial increase in IL-1β in the 3-NP-intoxicated rats suggests its release by activated astrocytes and microglia secondary to the 3-NP-triggered ROS production. In addition, the pronounced NLRP3 inflammasome activation in the microglia increases the conversion of pro-IL-1β into mature IL-1β via caspase-1, further exacerbating the IL-1β release and aggravating the neuroinflammatory milieu. This proceeding mimics the neuroinflammation process that contributes to HD pathology [10, 11]. However, roflumilast lowered the proinflammatory IL-1β levels in the striatum, confirming its favourable anti-inflammatory effect, as previously reported [25, 27, 28, 52]. One explanation is the ability of roflumilast to selectively inhibit the PDE-4B isoform, which is involved in inflammation and microglial activation [23, 26, 40]. This resulted in the observed amelioration of astrocyte and microglial activation in the striatum by roflumilast. A second explanation is the anti-pyroptotic effect of roflumilast observed in the current study, leading to the amelioration of mature IL-β1 production and release. Interestingly, roflumilast had a more profound impact on lessening IL-1β in the treatment regimen relative to prophylaxis. Moreover, roflumilast treatment or prophylaxis improved the histophathologic and immunohistochemical changes against 3-NP. Notably, treatment and prophylaxis with roflumilast exerted similar beneficial effects in terms of moderating the histopathological changes, increasing the striatal neuron survival rate, and reducing Iba-1 immunostaining (microgliosis). However, the treatment regimen was more beneficial in lowering the GFAP immunostaining (astrogliosis) than prophylaxis.
Here, we further attempted to explore the ferroptosis role in 3-NP-induced HD-like aberrations and the repercussion of roflumilast on this cell death mechanism. Ferroptosis is instigated by intracellular iron accumulation, leading to increased ROS production, oxidative stress, and neuronal cell damage. This causes a depletion of GSH and GPx4 and a consequent accumulation of lipid peroxides and their toxic derivative MDA [15, 19, 20]. Notably, the brain is increasingly liable to oxidative injury due to its high lipid content, oxygen demand, and less robust antioxidant defenses. MDA is regarded as a trustworthy indicator of oxidative stress-induced lipid peroxidation, a hallmark of HD pathogenesis [53]. In this study, the latter mechanism was supported by showing a 3-fold elevation of iron levels, along with increased lipid peroxidation and decreased GSH and GPx4 in the striatal neurons following 3-NP administration. Our results mirror previous documents of mitochondrial iron buildup in the brain of transgenic HD mouse models and its contribution to HD pathogenesis, supporting this notion by showing a potential protective effect of iron chelators [54, 55]. Furthermore, similar findings of increased striatal MDA and reduced GSH levels were previously reported in quinolinic acid-induced HD, along with the antioxidant effect of roflumilast, showing MDA reduction and GSH restoration [27]. However, the latter study only tackled oxidative stress and didnot question iron-driven ferroptosis or GPx4. A novel finding of our study is that roflumilast exerted an anti-ferroptotic action in both prophylaxis and treatment regimens. This effect was evident by reducing striatal iron and restoring GPx4 and GSH with a simultaneous reduction of MDA.
Intriguingly, in a secondary analysis, we showed strong statistical correlations between striatal pyroptotic and ferroptotic indicators, generating the hypothesis of interconnecting neuronal cell death pathways, with additional verification required. In addition, we highlighted the clustering of differentially expressed target genes, including PDE-4, CREB/BDNF, NLRP3/caspase-1/GSDMD, IL-1β, and GPx4 in a PPI network enriched in innate immune system, glial cell activation, IL-1 signaling, and chronic inflammatory response. These exploratory findings spotlight that the neuroprotective impact of roflumilast might be through multi-target potential affecting this PPI network. However, these results do not fully support dual ferroptosis-pyroptosis inhibition as a primary mode of roflumilast action, with further validation required at the cellular and molecular levels.
The more pronounced impact of roflumilast treatment over prophylaxis, particularly on the biochemical investigations, might be explained in part based on the hypothesis that 3-NP could induce disturbance in the BBB’s integrity through ROS generation [56], leading to more drug passage. Notably, oxidative stress inflicts damage to the BBB, as shown in neurodegenerative diseases [57]. Furthermore, previous studies showed that neurotoxins such as quinolinic acid cause damage and disruption to the BBB integrity [58]. Hence, the disruption induced by 3-NP could possibly alter the BBB permeability. This can plausibly allow more facilitated drug passage to the brain in the treatment group, which received 3-NP prior to roflumilast, in contrast to the prophylactic group that received 3-NP after roflumilast administration.
Few studies have addressed the neuroprotective impact of roflumilast on different CNS models; however, they have some methodological limitations. These include a single treatment regimen [25, 28, 29, 51, 52] and limited behavioral assessment [28, 29, 51]. Some exhibited a lack of mechanistic depth, with overemphasis on one pathway, such as acute inflammatory and injury markers [25, 52] or NLRP3 inhibition [28, 51], overlooking other relevant pathways. While recent studies investigated the neuroprotective potential of natural compounds in 3-NP-induced neurodegeneration and other neurodegenerative disorders targeting NLRP3, neuroinflammation, and oxidative stress [59–61], our study targeted both pyroptosis and ferroptosis using a pharmacological agent as a complementary perspective. Our study is superior in being comprehensive regarding comparing different regimens (treatment vs. prophylaxis), assessing a battery of behavioral tests, and evaluating two neuronal cell death trajectories, along with divergent markers of neurodegeneration. Although therapeutic strategies targeting ferroptosis using synthetic ferroptosis inhibitors (liproxstatin-1, ferrostatin-1), iron chelators (deferiprone, deferoxamine), or antioxidants (e.g., coenzyme Q10) are promising in preclinical and clinical HD models [15, 54, 55, 62], our findings pave the way for using roflumilast as a potential therapeutic candidate for HD through modulating pyroptosis and ferroptosis. Future studies are needed to investigate the additional benefits of roflumilast over existing therapies or its added value in combination therapy of HD.
Nevertheless, we acknowledge certain limitations in our study. These include the fact that 3-NP-induced HD animal models do not represent the genetic alteration associated with HD pathology or the full disease picture. This explains the lack of assessment of genetic HD markers, such as mHtt expression, which may limit direct translation to clinical HD pathology. Hence, future work employing genetic HD models and assessing genetic disease markers would reinforce the clinical relevance of our observations. Another limitation is that blinding was not feasible during behavioral assessments due to observable motor differences between the 3-NP-administered animals and those receiving the treatment. To minimize observer bias, behavioral experiments were standardized in handling, timing, testing conditions, and scoring criteria. A third limitation is that we only measured caspase-1 protein levels, without assaying its activity due to technical constraints. Nevertheless, caspase-1 protein levels often correlate with caspase-1 activity and the consequent cell death events [63]. Although the modulation of pyroptosis and ferroptosis-related markers by roflumilast provides valuable information, it may not fully provide direct pathways’ involvement. However, our present findings were based on assessing the association of well-established markers of these pathways with the roflumilast effects, which can direct future research. In-depth mechanistic validation using PDE-4 isoforms, functional assays such as lactate dehydrogenase release, TUNEL co-localization with GSDMD, or caspase-1 activity, and ferroptosis-specific rescue experiments (using ferrostatin-1, liproxstatin-1) represents a promising direction for future work to establish direct causal involvement of pyroptosis and ferroptosis pathways. Additionally, results of the correlation analysis and PPI network were only exploratory and hypothesis-generating and have to be verified in future studies. Moreover, we focused on short-term outcomes, and future studies are necessary to examine long-term neurological and functional outcomes, with an emphasis on pyroptosis and ferroptosis. Although the roflumilast dose used is established in different CNS animal models, the drug dose and duration were optimized for rats and may diverge from clinical regimens. Thus, clinical trials are necessary to confirm this therapeutic approach, investigate its optimal dosing, and evaluate its efficacy, safety profile, and individual response variability.
Conclusion
This study uncovered the potential neurorestorative and neuroprotective effects of roflumilast in 3-NP-induced HD-like neurodegeneration in rats by mitigating pyroptotic and ferroptotic neuronal cell deaths, lessening astrogliosis, microglial activation, and neuroinflammation, and restoring synaptic plasticity. These findings may establish the impact of pyroptosis and ferroptosis alleviation in HD and pave the way for roflumilast as a novel therapeutic option in HD management.