Niemann–Pick disease, type C (NPC) is an autosomal recessive, lysosomal storage disorder that affects ~1:120,000 births [1
]. Patients with NPC initially present with hepatosplenomegaly that is followed by progressive neurodegeneration leading to cerebellar ataxia and dementia [1
] with patients typically succumbing to the disease in the second decade of life [4
]. NPC is caused by mutations in NPC1 or NPC2, which encode for proteins that are involved in the transport of unesterified cholesterol out of the lysosome [5
]. Defects in either of these proteins result in the accumulation of unesterified cholesterol and other sphingolipids in the lysosome [9
]. A hallmark of the disease is the presence of neuroinflammation in the central nervous system [11
], leading to the anterior-to-posterior patterned cerebellar Purkinje neuron death within the cerebellum [15
]. The loss of Purkinje neurons is consistent with cerebellar ataxia, a major clinical finding in NPC. While much is known about the cause of NPC and the disease progression, there remains a need for a greater understanding of the molecular processes and pathways involved in the patterned neurodegeneration of the Purkinje neurons in order to develop new therapies.
Recently we performed RNA-sequencing analysis to examine the gene expression profiles of the cerebellar lobules of 4.5-week old Npc1-/-
]. This was done to identify genes and cellular pathways involved in the Purkinje neuron degeneration and potential protective expression patterns in lobule X of the cerebellum prior to disease progression. Our data identified several novel pathways of interest, including calcium, dopamine, and glutamate signaling, that may contribute to the lobule-specific susceptibility of Purkinje neurons to degeneration during disease progression [16
]. One gene of interest was Pcsk9
, that encodes proprotein convertase subtilisin/kexin type 9 (PCSK9) [17
]. This gene is expressed almost 8-fold higher in the anterior lobules of both Npc1+/+
cerebella compared to lobule X (Supplemental Figure S1
), suggesting a possible functional and lobule specific role for PCSK9 in the general physiology of the cerebellum and possibly of Purkinje neuron health during NPC disease progression.
PCSK9 is the ninth member of the proprotein convertase family of processing enzymes [18
]. Members of this family of enzymes are typically involved in the basic residue specific processing of proteins within the secretory pathway of cells to generate mature secreted biologically active peptides. Proprotein convertases are synthesized as pro-enzymes in the endoplasmic reticulum (ER) and require proteolytic activation by removal of their pro-domain in the trans Golgi network (TGN) or immature and mature secretory granules. However, PCSK9 appears to be unique in that its proteolytic function is inhibited through the tight binding of its cleaved pro-domain to its catalytic domain, rendering the enzyme inactive as a secreted proteinase [21
]. Hence, PCSK9 does not function as a protein processing enzyme.
The role of PCSK9 in the regulation of serum cholesterol has been extensively studied. PCSK9 associates with the extracellular EGF-like binding domain of the low-density lipoprotein receptor (LDLR) in the liver [22
]. After endocytosis of the LDLR, the bound PCSK9 triggers degradation of the internalized complex within the lysosome, thus reducing the levels of LDLR present on the plasma membrane [23
]. Such reduction of LDLR in the liver is sufficient to increase circulating levels of cholesterol. Hence, gain of function and loss of function mutations in PCSK9
have been found to modulate serum cholesterol levels in humans [25
] making PCSK9 inhibitors an attractive target to treat hypercholesterolemia in humans [30
]. During development in the mouse, Pcsk9
is expressed in the liver, kidney and intestine and in the brain it is expressed early in the telencephalon (embryonic day (E) 12) and later in the cerebellum (E17), with evidence it may be involved in neurogenesis [17
]. In the mouse, expression of Pcsk9
in the cerebellum persists during perinatal development, however, in the adult brain, overall levels are reduced with some signal in the external granule layer of the cerebellum [17
]. Such a function appears to be essential in zebrafish development, since reduced Pcsk9
expression in zebrafish results in significant abnormal neuronal development [31
], however, this is in contrast to the mouse model, since Pcsk9-/-
mice appear to develop normally [32
The function of PCSK9 in the central nervous system (CNS) is not fully understood despite implications for a role in Alzheimer disease (for review see [34
]). In the brain, two receptors closely related to LDLR; very low-density lipoprotein receptor (VLDLR) and apolipoprotein E receptor-2 (ApoER2), are expressed and have been studied as possible targets of PCSK9. PCSK9 and a gain of function mutant of PCSK9 enhanced cellular degradation of these receptors [35
], and PCSK9 binds to LDLR, VLDLR, and ApoER2 with similar sub-micromolar binding constants [36
]. Indeed, it has been proposed that PCSK9 modulates neuronal apoptotic signaling pathways via regulation of ApoER2 levels in the brain [37
]. VLDLR and ApoER2 are implicated in neuronal processes including cerebellar development and synaptic plasticity through Reelin signaling [38
]. In humans, loss of VLDLR
results in a non-progressive cerebellar ataxia phenotype [39
] and mutations in RELN
cause cerebellar hypoplasia [40
] similar to that of the Reeler
]. In addition, canonical and non-canonical signaling pathways of Reelin may contribute to its neuroprotective function, for example in the regulation of tau phosphorylation in the context of Alzheimer disease pathology [42
]. Hence, regulation of VLDLR and ApoER2 by PCSK9 may contribute to Reelin signaling and increased neuroprotection.
Given our recent results showing higher expression of Pcsk9 in the anterior lobules of the cerebellum where degeneration of the Purkinje neurons begins in NPC disease and the possible association with neuronal modulators ApoER2 and VLDLR, we hypothesized that decreased expression of Pcsk9 would increase surface expression of ApoER2 and VLDLR and thus confer neuroprotection of the Purkinje neurons in the cerebellum. To test this, we characterized the Npc1-/- mouse phenotype in the presence and absence of the Pcsk9 gene.
The role of PCSK9 in serum cholesterol regulation has been well established. Indeed, since its discovery in 2003, our current knowledge represents the complement of comprehensive basic and translational science, leading ultimately to an effective treatment of hypercholesterolemia [30
]. However, until more comprehensive studies on PCSK9 are undertaken to explore its role in the central nervous system (CNS), treatment of PCSK9-regulating drugs may affect the CNS in unseen and unwanted ways.
The exact role of PCSK9 in the CNS remains unknown. However, due to its well characterized cell biological mechanism in regulating the levels of the LDL receptor in the periphery, it is expected to function similarly in the brain. Studies have shown it can bind and regulate the degradation of the VLDL and ApoE2 receptors in a similar manner to that of the LDL receptor. In addition, since mature neurons receive their cholesterol from astrocytes through VLDLR and ApoER2, it is predicted that PCSK9 could contribute to brain cholesterol metabolism by regulating these receptors. In the adult mouse brain, it has been reported that PCSK9 is not involved in LDLR degradation in the cortex and hippocampus [36
]. However, Pcsk9
is increased in the dentate gyrus after transient cerebral ischemia and in a mouse model of hyperlipidemia where neuronal apoptosis occurs in the hippocampus [47
]. In both cases, reduced LDLR levels and increased neuronal apoptosis in these tissues appear to correlate with the increase in Pcsk9
. This suggests that in the context of disease or injury, levels of PCSK9 may regulate neuronal apoptotic signaling pathways. Based on these observations we therefore sought to test whether PCSK9 plays a role in NPC1, a disease of cholesterol metabolism.
Our studies however have shown that PCSK9 does not appear to play a role in NPC1 pathology even though we found expression differences in the lobules susceptible and resistant to Purkinje neuron death [16
]. In all parameters measured (disease severity score, weight, lifespan, selected motor functions and Purkinje neuron staining and serum liver enzymes), we did not discern an obvious difference in NPC1 disease progression in mice with or without PCSK9. Perhaps the levels of expression of PCSK9 are too low in the adult cerebellum to affect receptor levels especially in the Purkinje neurons layer, or perhaps specificity and sensitivity of binding to VLDLR or ApoER2 are different in vivo in the context of a cholesterol trafficking defect.
While further studies could be undertaken to characterize the molecular effects in the brain, we have decided that in the context of NPC1 disease outcome, this would not be informative. We present this work here to allow colleagues in the field to assess this aspect of NPC1 and conclude as we do that PCSK9 does not play a significant role in NPC1 disease.
4. Materials and Methods
4.1. Animal Maintenance
All animal work was reviewed and approved by the Eunice Kennedy Shriver National Institute of Child Health and Human Development (NICHD) Animal Care and Use Committee under protocol #18-002 (28 May 2018). Heterozygous Npc1+/- mice (BALB/cNctr-Npc1m1N/J) and B6;129S6-Pcsk9tm1Jdh/J (Pcsk9-/-) were purchased from The Jackson Laboratory (Bar Harbor, ME, USA). Npc1+/- mice were crossed with Pcsk9-/- mice to obtain double heterozygous mutants (Npc1+/-/Pcsk9+/-). These mice were then inter-crossed to generate double mutant mice (Npc1-/-/Pcsk9-/-) and the corresponding control mice. Pups were weaned 3 weeks post-birth and subsequently had free access to food and water. Genotypes were confirmed by PCR using DNA extracted from ear clips and the genotyping protocol reported by The Jackson Laboratory (protocol ID=27858). A humane end of life timepoint was determined in conjunction with the animal facility veterinarian, by assessing severity of motor impairment, at which point the mice were euthanized and tissues analyzed as necessary.
4.2. Disease Severity Scoring
After weaning, mice were assessed weekly for weight and disease severity using a standardized scoring system modified for NPC1 [46
]. The disease severity score consisted of 5 components: Grooming, gait, kyphosis, hindlimb clasp, and ledge test, with each observation scored from 0-3; 0 being the least severe and 3 indicating a severely affected phenotype. Upon entry to the procedure room, mice were allowed to acclimate to the new environment for 15 minutes. For the assessment, a mouse was randomly selected and weighed. The mouse was then placed into a new cage and assessed for disease phenotype. A composite disease severity score was calculated by adding the scores for each assessment. Evaluations were performed by two observers and mouse identification was revealed post-assessment. For the ledge test, mice were placed on the ledge of the cage and assessed for their ability to walk around the ledge with coordinated use of hindlimbs, proper balance, and ability to dismount effortlessly.
4.3. AccuScan Activity Assessment
One of the measures of motor coordination is the ability of the mice to rear up on their hind legs. We therefore measured the rearing activity of the mice in a 400 mm X 400 mm AccuScan Activity box (AccuScan Instruments Inc., Columbus, OH, USA), weekly from 5-10 weeks of age. This box contains 16 infra-red laser beams that cross the area of the box at 1-inch intervals and 1 inch above the floor to measure horizontal activity. An additional set of 16 lasers are active at 3.5 inches above the floor to measure vertical activity. Mice were initially brought to the behavioral room and allowed to acclimate for 30 minutes. In the room, ambient light and sound were constant and the researcher remained silent during data acquisition. Individual mice were placed in the chambers and activity was measured automatically by beam break counts collected by computer. Horizontal and vertical counts were recorded for three contiguous 5-minute periods. Total movements were calculated and reported for the combined second and third 5-minute periods (10 minutes).
4.4. Immunohistochemical Analysis of Cerebellar Tissue from Mutant Mice
Mice were euthanized at 7 weeks of age and transcardially perfused with room temperature phosphate buffered saline (PBS). The whole brain was dissected immediately and fixed in 4% paraformaldehyde (PFA) in PBS, pH 7.4, 4 °C, for 24 h followed by cryoprotection in 30% sucrose until sectioning. Brains were cryostat-sectioned parasagitally (20 μm) and sections were collected in PBS containing 0.25% Triton X-100. The floating sections were blocked for 30 min in 10% normal goat serum and then incubated overnight at 4 °C with rabbit anti-Calbindin (1:400; Cell Signaling, Danvers, MA, USA) in 0.25% Triton X-100/PBS. After washing, the sections were incubated with Alexa Fluor 594 goat anti-rabbit IgG, (1:1000, Thermo Fisher Scientific, Waltham, MA, USA). Nuclei were stained with Hoechst 33342, Trihydrochloride, Trihydrate I (1:1000, Invitrogen, Carlsbad, CA, USA). Images were taken on a Zeiss Axio Observer Z1 microscope (Zeiss, Germany).
4.5. Serum Liver Enzymes and Cholesterol Analyses
To assess liver function, mice were euthanized by CO2 and blood obtained by cardiac puncture. The samples were allowed to coagulate at room temperature for 45 min and then centrifuged for 10 min at 2000 rpm, 4 °C. The serum was saved at –20 °C until sent for analysis. Sera from control and affected mice (7–8 weeks of age) were sent to the Department of Laboratory Medicine, Clinical Center, NIH for hepatic and lipid panel screens. The liver enzymes assayed were alkaline phosphatase (AP), alanine transaminase (ALT), and aspartate transaminase (AST) using the following Roche Diagnostics reagent kits; AP #03333752 190; ALT #20764957 322; AST #20764949 322 and the lipids analyzed were total cholesterol using reagent kit #03039773 190. All assays were analyzed by Roche/Hitachi cobas c 311/501 analyzers (Roche Diagnostics GmbH, Mannheim, Germany).
4.6. Statistical Analysis
All statistical analyses were performed using the GraphPad PRISM software, version 8.3.0 (San Diego, CA, USA, www.graphpad.com
). Data were analyzed by one-way ANOVA followed by Tukey’s multiple comparisons test. The log rank Mantle–Cox test for significance was used for the Kaplan–Meyer survival curve. Statistical significance was set to p