Combination of Hydroxychloroquine and Indapamide Attenuates Neurodegeneration in Models Relevant to Multiple Sclerosis

Abstract

As the underlying pathophysiology of progressive forms of multiple sclerosis (MS) remains unclear, current treatment strategies are inadequate. Progressive MS is associated with increased oxidative stress and neuronal damage in lesions along with an extensive representation of activated microglia/macrophages. To target these disease mechanisms, we tested the novel combination of generic medications, hydroxychloroquine (HCQ), and indapamide, in tissue culture and in mice. HCQ is an anti-malarial medication found to inhibit microglial activation and to ameliorate disease activity in experimental autoimmune encephalomyelitis. We are currently completing a phase II trial of HCQ in primary progressive MS ( ClinicalTrials.gov Identifier: NCT02913157). Indapamide is an antihypertensive previously discovered in our laboratory drug screen to be an anti-oxidant. As these medications have a different spectrum of activities on disease mechanisms relevant to progressive MS, their use in combination may be more effective than either alone. We thus sought preclinical data for the effectiveness of this combination. In vitro, indapamide had robust hydroxyl scavenging activity, while HCQ and indapamide alone and in combination protected against iron-induced neuronal killing; TNF-α levels in activated microglia were reduced by either drug alone, without additional combination effects. In mice with a lysolecithin lesion that manifests demyelination and axonal loss in the spinal cord, the combination but not individual treatment of HCQ and indapamide reduced CD68+ microglia/macrophage representation in lesions, attenuated axonal injury, and lowered levels of lipid peroxidation. Our study supports the combination of indapamide and HCQ as a new treatment strategy targeting multiple facets of progressive MS.

Keywords: Multiple sclerosis; anti-oxidant; axonal loss; hydroxychloroquine; indapamide; microglia; neurodegeneration; neuroprotection; oxidative stress.

Figures

Fig. 1

Anti-oxidant activity of treatments against the hydroxyl radical. Representative relative fluorescence unit (RFU) graphs from HORAC assay of (A) indapamide, (B) HCQ, and (C) indapamide and HCQ alone or in combination. Gallic acid was used as a positive control anti-oxidant. Each data point is represented in RFU, taken every 5 min for 60 min. The greater magnitude of decay shown in the no anti-oxidant and 5 μM HCQ groups demonstrates an inability to counter the hydroxyl radical degradation of fluorescein. (D) AUC analysis from HORAC assay with various drug treatments. Bar graphs display the mean with error bars representing the SEM. Data were analyzed using a one-way ANOVA with Dunnett’s post hoc multiple comparison test. Each data point in (D) represents one experiment. ****p < 0.0001 compared to no anti-oxidant control. IND: indapamide; HCQ: hydroxychloroquine

Fig. 2

Drug treatments are neuroprotective against FeSO4 treatment in vitro on human fetal neurons. Representative images of human fetal neurons in vitro of (A) healthy DMSO control; (B) FeSO4 control; (C–E) 50 μM FeSO4 with drug treatments of (C) 1 μM indapamide, (D) 5 μM HCQ, and (E) combination. (F) Quantitation of surviving neurons with 1 μM indapamide, 5 μM HCQ, and their combination after 24-h incubation with 50 μM FeSO4. Scale bar represents 50 μm. Bar graphs represent the mean with error bars depicting SEM. Data were analyzed using a one-way ANOVA with Dunnett’s post hoc multiple comparison test to 50 μM FeSO4. Each data point represents one replicate well. The protective trend observed from HCQ and indapamide treatment was reproduced in two additional experiments. ****p < 0.0001 compared to iron. IND: indapamide, HCQ: hydroxychloroquine

Fig. 3

Drug treatments alone and in combination reduce TNF-α levels in vitro. Representative data from supernatant collected from cultured mouse microglia which were stimulated with 10 ng/ml LPS and subsequently subjected to drug treatments of 5 μM HCQ, 1 μM indapamide, and their combination. Supernatant was analyzed for TNF-α with an ELISA. Bar graphs display the mean with error bars representing the SEM. Data were analyzed using a one-way ANOVA with Dunnett’s post hoc multiple comparison test against the 10 ng/ml LPS group. Data has been reproduced with similar trends in two other experiments. Each data point represents one replicate well. **p < 0.01, ****p < 0.0001 compared to LPS. IND: indapamide, HCQ: hydroxychloroquine

Fig. 4

Treatment paradigm and effect of drug treatments on day 3 lysolecithin lesion size. (A) Timeline of pretreatment paradigm. (B) Representative image of the lysolecithin lesion, visualized with eriochrome cyanine staining for lipids (blue) and neutral red for nuclei (red). Closed arrowhead depicts lesion area. (C) Lesion epicenter area quantitation of each treatment group. (D) Quantitation of the lesion volume of each treatment group. Pooled data from the three separate experiments (n ≥ 5 each), normalized and represented as fold change of the vehicle group within a particular experiment. Graphs represented are the mean with error bars depicting SEM. Data were analyzed using a one-way ANOVA with Dunnett’s post hoc multiple comparison test, compared to vehicle treatment. Each data point represents one animal combined from 3 separate experiments. Horizontal dashed line represents fold change of 1 (the mean of vehicle group). IND: indapamide, HCQ: hydroxychloroquine

Fig. 5

Drug treatments have no effect on myelin degradation or oligodendrocytes in the lesion. (A–D) Representative images of treatment groups from the day 3 lysolecithin lesion, visualizing Olig2+ cells (white), NFH+ axons (green; see Fig. 7), MBP+ myelin (red), and the overlay image. (E) Number of oligodendrocyte lineage cells in the lesion. (F) Comparisons of MBP+ staining within the lesion, indicative of myelin debris. Pooled data from the three separate experiments (n ≥ 5 each), normalized and represented as fold change of the vehicle group. Graphs represented are the mean with error bars depicting SEM. Data were analyzed using a one-way ANOVA with Dunnett’s post hoc multiple comparison test, compared to vehicle treatment. Each data point represents one animal. Horizontal dashed line represents fold change of 1 (the mean of vehicle group). IND: indapamide, HCQ: hydroxychloroquine

Fig. 6

Effect of drug treatments on CD68+ microglia/macrophages in vivo. (A) Representative images of the day 3 lysolecithin lesion, visualizing DAPI for nuclei (blue) and CD68 for microglia/macrophages within the lesion (red). (B) Number of CD68+ cells within the lesion, expressed as percent coverage of the lesion area. (C) Mean fluorescence intensity of CD68+ microglia/macrophages within the lesion. Lower immunoreactivity represents reduced cell activation. Pooled data from the three separate experiments (n ≥ 5 each), normalized and represented as fold change of the vehicle group. White dashed line depicts lesion area. Graphs represented are the mean with SEM. Data were analyzed using a one-way ANOVA with Dunnett’s post hoc multiple comparison test, compared to vehicle treatment. Each data point represents one animal. Horizontal dashed line represents fold change of 1 (the mean of vehicle group). ***p < 0.001 compared to vehicle control. IND: indapamide, HCQ: hydroxychloroquine

Fig. 7

Combination treatment reduces the magnitude of axonal damage in vivo. (A–E) Representative images of normal-appearing white matter (NAWM), or of lysolecithin lesions in mice treated with vehicle, IND, HCQ, and combination. Each image includes NFH for axons (green), APP for compromised axons (magenta), and an overlay image. The insets of the overlay image depict APP accumulation on NFH+ axons (white). (F) Total number of NFH+ axons in the lesion. (G) Quantitation of number of NFH+ axons with accumulated levels of APP, represented as fold change of vehicle group. Graphs represent the mean with error bars depicting SEM. Data were analyzed using a one-way ANOVA with Dunnett’s post hoc multiple comparison test; *p < 0.05 compared to vehicle treatment. Each data point represents one animal. Pooled data from the three separate experiments (n ≥ 5 each), normalized and represented as fold change of the vehicle group. Horizontal dashed line represents fold change of 1 (the mean of vehicle group). IND: indapamide, HCQ: hydroxychloroquine

Fig. 8

HCQ and combination treatment reduce lipid peroxidation in vivo. (A) Representative images of the day 3 lysolecithin lesion for all treatment groups. Images are overlays of 4-HNE (green) for lipid peroxidation and DAPI (blue) for nuclei; the DAPI signal has been reduced to illuminate the 4-HNE better. White dashed line depicts lesion area. (B) Percent of the lesion containing 4-HNE+ immunofluorescence. Graph represents the mean with error bars depicting SEM. Data were analyzed using a one-way ANOVA with Dunnett’s post hoc multiple comparison test, compared to vehicle treatment; **p < 0.01; ***p < 0.001. Each data point represents one animal, normalized and represented as fold change of the vehicle group. Horizontal dashed line represents fold change of 1 (the mean of vehicle group). IND: indapamide, HCQ: hydroxychloroquine