Background
Niemann–Pick disease type C is a fatal lysosomal storage disorder that causes progressive neurodegeneration along with visceral organ involvement. Symptom onset and disease severity are variable, but patients commonly develop hepatosplenomegaly, cognitive decline, and seizures, culminating in death in the second or third decades of life [
1,
2]. Niemann–Pick C patients have loss-of-function mutations in the NPC2 (~ 5%) or, more commonly, the NPC1 (~ 95%) protein. In the late endosome/lysosomal compartment (LE/Lys), LDL-derived unesterified cholesterol is bound by NPC2 and transferred to the transmembrane NPC1 protein [
3,
4]. Using a poorly defined mechanism, NPC1 exports unesterified cholesterol from LE/Lys. Unesterified cholesterol then moves to other sites within the cell where it alters membrane dynamics or is utilized for steroid production [
4]. In patients with Niemann–Pick C, mutations in NPC1/NPC2 prevent intracellular lipid trafficking and cause characteristic cholesterol accumulation [
5]. A biochemically similar lipid storage disease arises from mutations in the gene encoding the lysosomal enzyme acid sphingomyelinase. Deficiency of enzyme activity causes Niemann–Pick disease types A and B, in which the storage of sphingolipids and cholesterol in LE/Lys leads to hepatosplenomegaly and varying degrees of neurodegeneration [
6].
Endogenous mechanisms to maintain cellular cholesterol homeostasis include the removal of excess cholesterol by high-density lipoprotein (HDL) particles. Cholesterol is effluxed from peripheral cells by nascent HDL particles and esterified in plasma. Mature HDLs then travel to the liver where cholesterol is eliminated in the bile [
7]. Recent work has taken advantage of endogenous HDL function for the development of synthetic HDL (sHDL) nanoparticles as potential therapeutics for cardiovascular diseases [
8‐
11]. These nanoparticles are composed of the HDL protein apolipoprotein A-1 (ApoA1) or ApoA1 mimetic peptides surrounding a lipid bilayer to form 10–12-nm diameter discoidal lipoprotein particles [
12,
13]. Chemical synthesis of sHDL permits modifications that alter lipid and ApoA1 peptide composition and thereby impact potency, pharmacokinetics, and safety [
14‐
17]. sHDL nanoparticles were initially designed for removal of cholesterol from lipid-laden atherosclerotic plaques. In clinical trials involving ~ 2000 cardiovascular disease patients, sHDL was safe and well-tolerated [
10,
11,
18‐
21], and a large phase III clinical trial in 17,400 patients is currently ongoing (
https://clinicaltrials.gov/ct2/show/NCT03473223).
Here, we developed and optimized a sHDL nanoparticle which significantly reduces the accumulated cholesterol in Niemann–Pick type C cells. The sHDL contains a 37-amino acid ApoA1 mimetic peptide, termed 5A, and sphingomyelin (SM). 5A-SM sHDL at 1:1.15 (wt/wt) peptide to lipid ratio is safe in primates, and with established sterile manufacturing, this sHDL is well positioned for rapid clinical translation [
22,
23]. We show that 5A-SM sHDLs are non-toxic and effective at reducing cholesterol storage in Niemann–Pick C patient fibroblasts and brain slice cultures from
Npc1 mutant mice. We establish that 5A-SM requires the ATP-binding cassette transporter 1 (ABCA1) to efflux stored cholesterol. In vivo studies using
Npc1 mutant mice show evidence of target engagement and rescue of peripheral phenotypes and neuronal cholesterol storage. Furthermore, we show that 5A-SM also rescues sphingomyelin storage in Niemann–Pick type A fibroblasts. Together, these studies provide proof-of-concept data to support the therapeutic potential of sHDL for the Niemann–Pick diseases.
Methods
Mice
All
Npc1-I1061T mice [
24] were backcrossed to C57BL/6 (≥ 10 generations). Approximately equal numbers of males and females were used for all experiments, and littermates were used when available. Mice were randomly assigned to vehicle or experimental groups. All procedures involving mice were approved by the University of Michigan Committee on Use and Care of Animals (PRO00008133) and conducted in accordance with institutional and federal guidelines.
Reagents
2-Hydroxypropyl-β-cyclodextrin (H-107) and amiloride (A7410) were from Sigma. EndoH (P0702) and PNGaseF (P0704) were from New England Biolabs. Dynasore (14061) was from Cayman Chemical; Human HDL (J64903) and acetylated LDL (J65029) were from Alfa Aesar. 5A peptide (DWLKAFYDKVAEKLKEAF-P-DWAKAAYDKAAEKAKEAA, 4078379) was from Bachem Americas (Torrance, CA). 22A peptide (PVLDLFRELLNELLEALKQKLK) was synthesized by Genscript (Piscataway, NJ). Lipids including egg-sphingomyelin (SM, Coastome NM-10), 1,2-dimyristoyl-sn-glycero-3-phosphocholine (DMPC, Coastome MC-4040), and 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine (POPC, Coastome MC-6081) were from NOF America Corporation. 1,1′-Dioctadecyl-3,3,3′,3′-tetramethylindodicarbocyanine, 4-chlorobenzenesulfonate salt (DiD, D7757) and 4-(4-(dihexadecylamino)styryl)-N-methylpyridinium iodide (DiA, D3883) were from Invitrogen. Cholesteryl [1,2,6,7-3H(N)] linoleate (ART 1203) and sphingomyelin [choline methyl-3H] were from American Radiolabeled Chemicals (Saint Louis, MO). N-[6-[(7-Nitro-2-1,3-benxoxadiazol-4-yl)amino]-sphingosine-1-phosphocholine (NBD-Sphingomyelin, 810218P) was from Avanti Polar Lipids (Alabaster, AL).
Antibodies
(Antigen, dilution, vendor, cat. no.): NPC1, 1:500, Abcam, ab134113; Actin, 1:4000, Sigma, A544; LAMP1, 1:100, Developmental Studies Hybridoma Bank University of Iowa, H4A3; Calbindin, 1:500–1:2000, Sigma, 02724; GFAP, 1:500, Dako, z0344; IBA1, 1:250, Abcam, ab5076; NeuN, 1:500, Millipore, abn78; EEA1, 1:400, Abcam, ab2900; F4/80, 1:400, abcam, ab6640.
sHDL synthesis
sHDL particles were prepared using a lyophilization method where peptide (5A or 22A) and lipid (SM, DMPC or POPC) were dissolved in acetic acid at 1:1.5 wt/wt ratio and then lyophilized together for 24 h. HDL was fluorescently labeled by adding 4 μg DiD or DiA per 1 mg peptide directly to the acetic acid mixture of peptide and SM. The resulting lyophilized dry pellet was re-hydrated in PBS, pH 7.4, to a final peptide concentration of 10 mg/mL, vortexed, and thermocycled 3× between 55 °C and room temperature to generate sHDL particles. pH was adjusted to 7.4 and sHDLs were sterile filtered using 0.22-μm Millipore filters. Labeling of 5A peptide in sHDL (5A-SM-DiA) with AlexaFluor 647 dye was performed using Invitrogen protein labeling kit (A20173). Purification of 5A-SM-DiA-Alexa647 post-labeling was done on the size-exclusion column supplied in the kit, and the final concentration of sHDL was determined according to the manufacturer’s instructions using plate reader measurements (SynergyTM NEO HTS Multi-Mode Microplate Reader, Bio-Tek).
sHDL characterization
Fluorescently labeled sHDL particles were analyzed by UPLC (Waters Aquity UPLC BEH125 sec 1.7 μm, 4.6 × 150 mm column) equipped with UV (220 nm) and fluorescence detectors (ex/em 644/665 nm DiD, 456/590 nm DiA, 650/665 nm AlexaFluor647). The hydrodynamic diameters of sHDL were determined by dynamic light scattering on Zetasizer Nano ZSP, Malvern Instruments (Westborough, MA). The volume intensity average values were reported. Transmission electron microscopy images were obtained on an FEI Morgagni electron microscope run at 100 kV at a magnification of 22,000× (2.1 Å/pixel) and then recorded on a Gatan Orius charge-coupled device camera. sHDL samples (3 μL of 2 μg/ml) were adsorbed for 1 min to a glow discharged 400-mesh copper grid covered with carbon-coated collodion film (Structure Probe). The grids were washed twice and then negatively stained in 0.07% uranyl formate. 22A and 5A peptides, SM, POPC, and DMPC lipids combined at 1:0.5, 1:1, and 1:2 wt/wt ratios were described previously [
14,
15].
Cells
Cell lines were obtained from the NIGMS Human Cell Repository at the Coriell Institute for Medical Research. GM08399 was used as a control (CTRL) cell line. Niemann–Pick C cell lines with mutations in the
NPC1 gene: GM18453 (I1061T/I1061T), GM17912 (P1007A/T1036 M), and GM03123 (I1061T/P237S); Niemann–Pick A (NPA) cell line with mutation in the
SMPD1 gene GM00112 (L302P/L302P). Cells were cultured in MEM, PSG, and 20% FBS [
25].
Treatments
Endocytosis inhibitors
Cells were pretreated with dynasore (80 μM) or amiloride (1 mM) for 30 min. Cell culture media was replaced with fresh media containing vehicle (saline), dynasore, or amiloride along with 5A-SM-DiD for 2 h. ImageJ was used to quantify DiD label intensity inside cells.
sHDL in cells
Cells were plated 24 h before treatment. At the start of treatment, cell culture media was replaced with media containing vehicle or sHDL. Culture media containing vehicle or sHDL was refreshed after 24 h.
sHDL treatment of brain slices
Slices were treated with fresh particles/media daily at a concentration of 5 mg/ml for a period of 4 days.
siRNA transfection
Predesigned ON-TARGETplus SMARTpools containing 4 individual siRNAs per target sequence (Dharmacon Non-targeting SMARTpool D-001810-10-05, ABAC1 L-004128-00, SR-B1 L-010592-00) were transfected using TransIT-X2® (Mirus) reagent at
t = 0 and
t = 24 h. Imaging or RNA analysis occurred 48 h after the first transfection [
25].
Western blot
A bullet blender (Next Advance) was used to homogenize cell lysates. Protein concentrations were normalized by DC™-protein assay (Bio-Rad), and equal amounts of protein were loaded into 4–12% gradient SDS PAGE gels (Invitrogen). After electrophoresis and transfer onto a PVDF membrane, immunoreactivity was detected by ECL (Thermo Scientific) and imaged using an iBright (Thermo Fisher Scientific). ImageJ was used to quantify band intensity [
25]. For Endoglycosidase H assay, lysates were separated into three reactions containing: negative control (NT), EndoH (E) (NEB P0702L), or PNGaseF (P) (NEB P0704L) [
25]. After a 3-h incubation at 37 °C, samples were loaded on SDS PAGE gels as indicated above.
Filipin staining
After treatment, cell membranes were labeled with wheat germ agglutinin® (Thermo Fisher). Cells were fixed in 4% PFA for 20 min, washed 3× in PBS, and 1× in glycine. Unesterified cholesterol was labeled with filipin labeling solution for 2 h. Filipin labeling solution: 10% FBS + 0.4% DMSO+ 0.03 mg/ml (tissue) or 0.1 mg/ml (cells) filipin. Slides were washed 3× with PBS and mounted with ProLong® Gold (Thermo Fisher) [
25].
RT-qPCR
RNA was converted to cDNA using the High Capacity Reverse Transcription kit (Applied Biosystems 4368814). Quantitative real-time PCR (RT-qPCR) was conducted in technical triplicates using 15 ng cDNA, TaqMan™ probes (Thermo Fisher) for human HMGCR (Hs 00168302), HMGCS1 (Hs 00940429), ABCA1 (Hs 01059118), ABCG1 (Hs 00245154), LDLR (Hs 01092524), NPC1 (Hs 00264835), SCARB1 (SR-B1) (Hs 00969821), SREBF-2 (SREBP) Hs 01081778, GAPDH (loading control) (4325792), and mouse HMGCS (Mm 01304569). RT-qPCR was performed using an ABI 7900HT Sequence Detection System and relative expression calculated by the 2−ΔΔCt method using SDS software.
Immunofluorescence staining
Cells were washed 3× with HBSS and fixed with 4% PFA for 20 min at room temperature. Cells were washed with PBS and glycine before addition of blocking solution (0.02% saponin, 10% normal goat serum (NGS), 1% BSA) for 1 h. Slides were incubated with primary antibodies overnight at 4 °C, washed with PBS + 0.02% saponin, and incubated with secondary antibody for 1 h [
25]. Slides were mounted with Vectashield + DAPI (Vector Laboratories).
For slice culture: slices were floated in HBSS+/+ in 6-well plates containing Netwell™ Inserts (Corning). Samples were fixed in 4% PFA and 0.1% Triton X-100 for 1 h, rinsed 3× in PBS, then treated for 10 min of 1.5 mg/ml glycine. After three washes in PBS, slices were blocked in PBS containing 5% NGS for 1 h at room temperature. Slices were labeled with primary antibody (diluted in blocking) overnight. The following day, slices were washed 3× in PBS and labeled with Alexa conjugated secondary (1:500) for 1 h. After 3 washes in PBS, slices were stained with filipin labeling solution for 2 h, washed 3× with PBS, and mounted in ProLong Gold (ThermoFisher) and imaged by confocal microscopy. Calbindin was used to outline Purkinje cells, and filipin intensity was calculated using ImageJ.
Preparation of cerebellar organotypic slice cultures
Cerebellar organotypic slice cultures were prepared using 30- μm thickness sagittal brain slices [
26]. Four slices per brain were used in each set of experiments, split evenly between control and experimental medium. Two slices were placed together on a cell culture insert (Millipore; 0.4-μm pore size, 30 mm diameter) which contained 1.2-ml slice culture medium (either control or experimental) and were pre-incubated at 37 °C in 95% O
2/5% CO
2 in a 6-well plate. Control medium contained 50% minimal essential medium with Earle’s salts, 25% horse serum, 25% Hank’s balanced salts solution, 25 mM HEPES, 2 mM
l-glutamine, and 6.5 mg/ml glucose. Experimental medium was prepared by adding nanoparticles at a concentration of 5 mg/ml to the aforementioned control medium. Every 24 h, cell culture inserts were transferred to a new 6-well plate which was pre-incubated at 37 °C in 95% O
2/5% CO
2 with control or experimental medium, as described above. Imaging and analysis of Purkinje neuron cholesterol content was performed after 96 total hours of incubation. In all cases, wild-type and NPC samples were matched so that slices were prepared on the same day and using the same reagents.
Stereotaxic mouse ICV bolus delivery
Stereotaxic administration of nanoparticles into the right lateral ventricle via an intracerebral ventricular (ICV) injection was performed on mice under vaporized isoflurane anesthesia according to IACUC guidelines. Six- to 7-week-old mice received a single ICV bolus injection of sHDL or vehicle using established protocols [
25,
27]. Each anesthetized mouse received a small scalp incision to expose the skull, and a small burr hole was drilled relative to Bregma suture: anterior-posterior + 0.3 mm, medio-lateral − 1.0 mm. A beveled needle (7758-04, Hamilton, Reno, NV) connected to a 10-μL syringe (7653-01, Hamilton, Reno, NV) was placed dorso-ventral − 3.0 mm at a rate of 1 mm/s. A 3-min wait was allotted for the brain to seal around the needle and prevent backflow of treatment around the injection site. A total of 10 μL vehicle or sHDL at a concentration of 100 mg/ml was delivered at an infusion rate of 0.5 μL/s using an injection pump (UMC4, World Precision Instruments, Inc., Sarasota, FL). Five minutes after the infusion was completed, the needle was retracted at a rate of 1 mm/s and the incision site was sutured with synthetic non-absorbable sutures (1011209, Henry Schein, Melville, NY). Mice were recovered in a temperature-controlled environment, and following surgery, the mouse weight, grooming activity, and home cage activity were recorded for up to 7 days according to IACUC guidelines.
Microscopy
Epifluorescence: Filipin was imaged on a Zeiss Axio Imager Z1 microscope with an automated stage. Cells were focused in the green channel (wheat germ agglutinin), and 16 tiled images were captured per experiment. Images with ≥ 90% cell confluence were quantified using NIH ImageJ software [
25].
Confocal imaging of cells: Fluorescently labeled sHDL particles were imaged on a Nikon A-1 confocal microscope. Co-localization coefficients were calculated using Nikon elements software (Pearson). Brightness and contrast were applied equally across the entire image to both control and experimental groups using Photoshop.
Macrophages were outlined in F4/80-stained sections of liver and area was quantified using ImageJ by an investigator blinded to genotype and treatment.
Confocal imaging of tissue: One-week post intraventricular injection, vehicle or 5A-SM-treated mice were perfused with saline and tissues were placed in 4% PFA overnight. The liver and right hemisphere of the brain were embedded in OCT, frozen, and cut into 10-μm-thick sections. Sections were permeabilized (0.1% triton/10% NGS/1% BSA in PBS) for 30 min and placed into blocking buffer (10% NGS/1% BSA in PBS) for 60 min. Sections were placed in primary antibody overnight at 4 °C, washed three times in PBS for 5 min, then incubated in secondary antibody for 1 h at room temperature. Sections were stained with filipin and imaged on a Nikon A-1 confocal microscope. Purkinje neuron soma were defined using a calbindin-DK28 antibody, and filipin was quantified using ImageJ.
Stimulated Raman scattering (SRS) microscopy: Cell monolayers were imaged at 2845 cm
−1 Raman shift wavenumber to generate a greyscale image channel. Images acquired at the 2845 cm
−1 are chemically selective for lipids, stimulating vibrational resonance of the CH
2 symmetric stretching mode [
28]. Individual fields of view (FOVs) for lipid quantification were generated and quantified using a two-layer automated thresholding method to avoid selection bias. Over a full 2 mm × 2 mm SRS image, a 250 pixel × 250 pixel sliding window with 100-pixel step size was used to detect FOVs with greater than 90% cellular confluence. Mean background pixel intensity values for each image were used to set the FOV threshold for background (i.e., media) and foreground (i.e., cells). Only FOVs with a foreground/background ratio greater than 90% were included for lipid quantification. After selection of FOVs, a second thresholding procedure was used to segment intracellular lipid droplets, which have high 2845 cm
−1 SRS signal compared to the remainder of the intracellular contents. For each FOV, a ratio between the area of intracellular lipid to total intracellular space was calculated and normalized to the number of cells within each image.
Amplex Red
The Amplex® Red Cholesterol Assay Kit A12216 (Invitrogen) was used to quantify total free cholesterol following the manufacturer’s instructions.
Cell death
Cell viability was assessed using Promega CellTiter 96 Aqueous One Solution cell proliferation colometric assay (G3580). Briefly, Niemann–Pick C cells were cultured in 96-well plates at 10,000 cells per well for 24 h, washed 3× with PBS, and treated as indicated with compounds diluted in media for 24 h. Cells were washed 3× with PBS and re-suspended in media supplemented with Promega CellTiter 96 reagent (20 μl reagent per 100 μl of media). After 45-min incubation at 37 °C, absorbance was read at 490 nm using a microplate reader. Each treatment was performed in triplicate, and the average absorbance reading of non-treated (Veh) cells was set to 100%. The percent viability was determined by dividing the average absorbance of treated over non-treated cells and multiplying by 100.
Sphingomyelin loading
C6-NBD sphingomyelin was dissolved in 100% ethanol to make a 10 mM stock solution. Cells were treated with 40 μM C6-NBD sphingomyelin in cell culture media overnight. The following day (t = 0), wells were briefly washed 2× with PBS and fresh media without C6-NBD sphingomyelin was added. At t = 0 and t = 24 h, cells were treated with fresh media containing vehicle (saline) or 5A-SM.
Radioactive cholesterol efflux assay
Preparation of [3H] cholesteryl linoleate-loaded acLDL
Cholesteryl [1,2,6,7-3H(N)] linoleate (60 Ci/mmol) was loaded into acetylated human LDL (acLDL) according to procedure adapted from Brown et al. [
29]. Briefly, 30 μCi (0.5 nmol) cholesteryl [1,2,6,7-3H(N)] linoleate in toluene was evaporated to dryness under a stream of nitrogen gas. Then, a thin film of cholesteryl [1,2,6,7-3H(N)] linoleate was dissolved in 10 μl DMSO followed by the addition of 100 μl acLDL (5 mg of protein/ml). The mixture was incubated for 2 h at 37 °C with gentle shaking to incorporate cholesteryl [1,2,6,7-3H(N)] linoleate into acLDL and then dialyzed at 4 °C against 20 mM Tris/HCl, 0.3 mM EDTA, 0.15 M NaCl, pH 7.4 using 3.5K MWCO slide-A-Lyzer mini device (ThermoFisher 88,400). Cholesteryl [1,2,6,7-3H(N)] linoleate-acLDL mixture routinely contained 90–95% of starting radioactivity as determined by scintillation counting before and after dialysis.
Cholesterol efflux assay
Niemann–Pick C fibroblast cells were grown in culture media until confluency. On day 1, 75,000 cells were plated in 24-well plates and grown for 24 h in 0.5-ml culture media. On day 2, cells were washed with PBS, pH 7.4, 1× at room temperature and grown overnight in media containing lipoprotein-deficient serum (10% v/v) in DMEM to upregulate LDL receptors. On day 3, cells were washed with PBS, pH 7.4, 2×, and labeled with cholesteryl [1,2,6,7-3H(N)] linoleate-acLDL for 24 h in DMEM (no phenol red)/BSA (1 mg/ml)/P-S media (0.5 ml) containing 1 μCi of [3H] cholesteryl linoleate per 1-ml media. On day 4, labeled cells were washed with PBS, pH 7.4, 3×, to remove cholesteryl [1,2,6,7-3H(N)] linoleate not taken up by cells. Radioactive cholesterol was effluxed from cells for 24 h using vehicle (media), 5A peptide (0.75 mg/ml), 5A-SM HDL, 5A-DMPC, 5A-POPC (0.75 mg/ml), or cyclodextrin (1 mM) diluted in DMEM/BSA/P-S. On day 5, media from each well was transferred into separate Eppendorf tubes and centrifuged at 3000 rpm for 10 min to remove any detached cells. The remaining cells on the plate were lysed with 0.1% SDS/0.1 M NaOH solution for 2 h at room temperature. Radioactive counts of media and cell fractions were measured separately using a Perkin Elmer liquid scintillation counter. Percent cholesterol effluxed from cells was calculated by dividing media counts by the total sum of media and cell counts and then multiplying by 100%. Non-specific cholesterol efflux by vehicle was subtracted from all data.
In vivo cholesterol mobilization
Total serum cholesterol concentrations from 7-week-old Niemann–Pick C mice pre- and 2 h post-treatment with 100 mg/kg 5A-SM i.p. were analyzed enzymatically by a colorimetric cholesterol oxidase assay (Wako Chemicals, Richmond, VA) using microplate reader.
Distribution of mobilized cholesterol in lipoproteins
Serum samples from Niemann–Pick C mice collected at baseline and 2 h post-treatment with 100 mg/kg 5A-SM i.p. were analyzed to assess the cholesterol distribution between VLDL, LDL, and HDL lipoprotein fractions. Separation of lipoproteins from serum was performed on a Waters HPLC system equipped with a Superose 6, 10/300 GL column (GE Healthcare, Piscataway, NJ) and a fraction collector. Serum samples were injected onto the HPLC and eluted with saline solution pH 7.4 at 1 ml/min. Eluent fractions containing different lipoproteins were post-column reacted in the HPLC with an enzymatic solution for total cholesterol detection [
30].
Radioactive sphingomyelin efflux assay
Cells (40,000 cells/well) were cultured for 24 h in 24-well plates and then incubated with 1 μCi (80 Ci/mmol) sphingomyelin [choline methyl-3H] per 1-ml media. After 24 h, cells were washed with PBS, pH 7.4, 3×, followed by the treatment with vehicle or 0.75 mg/ml 5A-SM in culture media. Radioactivity in media and cells was counted using a PerkinElmer scintillation counter. Percent sphingomyelin effluxed from cells was calculated by dividing media counts by the total sum of media and cell counts and then multiplying this number by 100%. Non-specific sphingomyelin efflux by vehicle was subtracted from all data.
Serum analysis
Whole blood was collected and allowed to clot for 5 min in BD microtainer® SST gold cap tubes (365967). Tubes were centrifuged for 5 min at 3000×g to remove the clot. Liver enzymes were blindly analyzed by the University of Michigan In-Vivo Animal Core.
Statistics
Significance (p < 0.05) was determined by Graphpad Prism 7.0. Figure legends indicate when unpaired Student’s two-tailed t test, one-way or two-way ANOVA with Tukey, or Bonferroni post hoc analysis were used. All error bars are s.e.m. Graphpad outlier analysis was used to remove one outlier per group for Purkinje neuron filipin quantification.
Discussion
We describe an innovative approach to ameliorating lipid storage in the Niemann–Pick family of diseases by harnessing the activity of the body’s endogenous cholesterol scavenging particle, HDL. The sHDL particles characterized here potently remove stored cholesterol from Niemann–Pick C fibroblasts (Fig.
2) and neurons (Fig.
5f). The particles show evidence of cholesterol target engagement and rescue disease phenotypes when administered to Niemann–Pick C mice (Fig.
5). The 10–12 nm sHDL nanodiscs are generated at high purity by assembling peptide-lipid nanoparticles by a co-lyophilization and thermocycling process (Fig.
1). Notably, the degree of cholesterol removal was affected by altering the constituent ApoA1 mimetic peptide, lipid, and peptide to lipid ratio (Fig.
2), demonstrating that sHDLs provide a flexible platform that can be tuned to adjust therapeutic potency. Moreover, our observation that the sHDL that rescues cholesterol storage in type C disease also rescues sphingolipid storage in type A disease (Fig.
6) raises the possibility that alternative sHDL compositions may be beneficial for additional lipid storage disorders. The initial in vivo analyses presented here provide a proof of concept of activity for a single sHDL formulation, 5A-SM, at limited points. 5A-SM treatment of Niemann–Pick C mice induces cholesterol mobilization from the liver (increased HMGCS expression, Fig.
5c), increases serum cholesterol (Fig.
5a), and reduces liver inflammation (Fig.
5f). These data set the stage for additional analyses in Niemann–Pick animal models, including comparisons with other therapies currently administered to patients or in clinical trial. Future analyses are also needed to determine the extent to which optimized sHDL treatment regimens impact lysosomal cholesterol and sphingolipid storage in liver and normalize oxysterol biomarkers.
Prior studies have established that Niemann–Pick type C cells have normal ApoA1 receptor binding, endocytosis, and re-secretion [
56,
57], yet Niemann–Pick patients have reduced serum HDL levels [
56,
58‐
62] that likely worsen lipid storage. Previous reports have also demonstrated that Niemann–Pick type C cells are defective in loading cholesterol into ApoA1 [
56,
57,
63]. Similarly, we noted that ApoA1 mimetic peptides are not sufficient for reducing Niemann–Pick C cholesterol storage (Fig.
2a, b). Our strategy bypassed the HDL formation deficiencies in disease by utilizing an investigational new drug, 5A-SM sHDL, that exhibits no significant cellular toxicity. The activity of alternative sHDL formulations will be the subject of future research. Additionally, the extent to which incorporation of ApoE rather than ApoA1 mimetic peptides and the addition of brain targeting peptides enhance therapeutic efficacy in the CNS remain to be defined.
While the rescue of cholesterol storage from mutant fibroblasts required expression of ABCA1 (Fig.
2e), fluorescently labeled sHDLs readily entered cells by macropinocytosis (Fig.
4a). The lipid and peptide constituents of the nanoparticles remained tightly associated within cells (Fig.
4b, c), with some trafficking to LAMP1 and filipin staining vesicles (Fig.
4c). Other intracellular sHDLs remained outside these lipid storage vesicles, and the precise location of cholesterol loading remains to be determined. In contrast to LDL, there is still incomplete understanding of HDL endocytosis and subcellular trafficking [
64]. Interestingly, multiple intracellular pools of HDL have been described. After endocytosis, HDL can be re-secreted, sent to the lysosome, or trafficked to the Golgi before re-secretion [
64‐
66]. The location of the subcellular pools of HDL is still being described and is likely cell type dependent. Although we were not able to define where all of the 5A-SM localized within cells, treatment with sHDLs did trigger cholesterol efflux (Fig.
4d), a finding that may reflect release of cholesterol-laden nanoparticles and/or enhanced lysosomal exocytosis.
The identification of therapeutic rescue of lipid storage in Niemann–Pick type A cells by sHDL was greatly facilitated by SRS microscopy (Fig.
6c). This technique was used to circumvent challenges associated with labeling endogenous sphingolipids and shortcomings of applying exogenous sphingolipids to study intracellular trafficking. We were not able to identify the accumulated lipid species by current technology and cannot exclude the possibility that sphingomyelin correction is a consequence of cholesterol removal. However, this seems unlikely as treatment with the cholesterol-removing agent cyclodextrin did not alter sphingomyelin accumulation or lipid content in Niemann–Pick type A cells (Fig.
6b, c). Notably, limited techniques allow live cell imaging of endogenous lipids. While SRS cannot currently delineate lipid subspecies, we anticipate that continued development of this technology will enable this process. Moreover, the robust rescue of lipid storage in Niemann–Pick type A cells, along with the amelioration of peripheral phenotypes in Niemann–Pick C mice following i.p. administration, is particularly encouraging in the context of Niemann–Pick type B. Niemann–Pick B is characterized by peripheral organ system phenotypes, but not CNS involvement. Even with the successful removal of sphingomyelin in Niemann–Pick A cells by 5A-SM, future research is needed to establish the optimal sHDL formulation for removing stored lipids in this disorder.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.