Introduction

Niemann-Pick disease, type C1 (NPC1) is a fatal, genetic disorder that results in severe, progressive neurodegeneration. Mutations of the NPC1 gene result in a loss of function of the NPC1 protein (Carstea et al 1997), thereby, impairing cholesterol and glycosphingolipid trafficking. A deficiency of NPC1 protein function results in an accumulation of unesterified cholesterol and glycosphingolipids in the late endosomes/lysosomes (Pentchev et al 1987; Zervas et al 2001; Vanier and Millat 2003). Following the initial defect, a complex cascade of pathological events occurs in NPC1 including oxidative stress (Reddy et al 2006; Zampieri et al 2009; Fu et al 2010; Porter et al 2010; Klein et al 2011; Vazquez et al 2011), neurofibrillary tangle formation (Love et al 1995; Suzuki et al 1995), and neuronal apoptosis among others (Ong et al 2001; Sarna et al 2003; Wu et al 2005). Clinical symptoms in NPC1 patients are heterogeneous and include hepatosplenomegaly, ataxia, vertical gaze palsy, and dementia, in which progression in neurological severity occurs over time (Vanier 2010; Yanjanin et al 2010). To date, there is no FDA-approved therapy for NPC1. However, reports have indicated that treatment with miglustat, (Zavesca®), an imino sugar that blocks glycosphingolipid synthesis, slows the neurological progression of the disease in both animal models and NPC1 patients (Patterson et al 2007; Pineda et al 2009; Patterson et al 2010; Wraith et al 2010). Miglustat has been approved for the treatment of NPC1 by the European Medicines Agency. Currently, miglustat is only FDA-labeled for the treatment of Gaucher disease but has been used off-label for NPC1. Recently, 2-hydroxypropyl-β-cyclodextrin (HP-β-CD) has shown promise as a potential therapy for NPC1. HP-β-CD has been demonstrated to maintain neurological function and reduce the storage burden in both mouse and feline models of NPC1 disease (Davidson et al 2009; Liu et al 2009; Ramirez et al 2010; Ward et al 2010; Aqul et al 2011).

Neuroinflammation is a common feature found in many disorders particularly those affecting the central nervous system (CNS). Within the CNS, the innate immune system of microglia, astrocytes and perivascular macrophages serve as a first line of defense (Graeber et al 2011; Veerhuis et al 2011). The activation of microglia and astrocytes during the inflammation process results in a morphological change and is characterized by positive staining for CD68 and GFAP, respectively (Eng et al 2000; Kunisch et al 2004; Marin-Teva et al 2012). Secreted proteins from both microglia and astrocytes can be used as inflammation markers in cerebrospinal fluid-based analyses [as reviewed (Suk 2010)].

The first evidence of neuroinflammation in the Npc1 mouse is the activation of microglia at approximately 2 weeks post birth. This activation begins the neurodegenerative cascade with subsequent marked activation of astrocytes around 4 weeks post birth, beneath the Purkinje cell layer corresponding to the sites of early apoptosis in NPC1 disease (Baudry et al 2003). Smith et al tested the effectiveness of non-steroidal anti-inflammatory drugs in Npc1 mutant mice where a positive response with respect to survival was observed, as well as reduced microglial activation (Smith et al 2009). Pressey and co-workers investigated the effect of Npc1 deficiency on brain pathology at different stages of the disease process (Pressey et al 2012). The thalamus and cerebellum have been identified as particularly vulnerable to neurodegeneration, showing early activation of glia from 3 weeks post birth (Pressey et al 2012). The relationship between glia and neurons is variable within different brain regions, suggesting that the mechanism underlying the neuroinflammation in various brain regions may differ. Recently miglustat treatment in the feline model of NPC1 was shown to improve Purkinje cell survival, and reduced lipid storage and microglial activation suggesting that neuroinflammation may be affected with miglustat (Stein et al 2012).

Microglial activation in chronic neurodegeneration can be beneficial, harmful or non-significant (Ransohoff and Brown 2012). The role of inflammation in the disease progression of NPC1 has yet to be determined. It is particularly unclear whether neuroinflammation is a primary pathological process, or a secondary event resulting from the initial genetic and trafficking defects. Recent data suggest that neuroinflammation in NPC1 is a secondary process (Lopez et al 2012a, b). While some studies have discussed the prevalence of inflammation in Npc1 mouse brain tissue and the potential benefit of anti-inflammatory treatment, human studies are lacking. Understanding the role of neuroinflammation and identification of biomarkers associated with this aspect of NPC1 pathology will be of utility in designing and translating potential therapies to NPC1 patients.

In an effort to further characterize and potentially target therapeutic interventions along the neuropathogenic cascade in NPC1, we sought to further understand the neuroinflammation processes occurring in both Npc1 mice and NPC1 patient brain tissue using differential expression analysis. Our work additionally aimed at establishing which inflammatory markers are altered in the cerebrospinal fluid from NPC1 patients relative to controls. Using discovery and targeted based approaches, several inflammation markers were found to be altered in the Npc1 mouse model as well as in cerebrospinal fluid from NPC1 patients. These data will provide a more comprehensive understanding of the array of biological processes that are involved in NPC1 pathology and provide a deeper understanding of the disease process while concurrently providing potential biomarkers for monitoring disease progression and evaluating potential therapeutic agents’ efficacy.

Materials and methods

Animal breeding and tissue isolation

Animal work was performed under an NICHD Animal Care and Use Committee-approved animal study protocol. Heterozygous Npc1 +/− mice (BALB/c Nctr-Npc1 m1N/J strain) were intercrossed to obtain control (Npc1 +/+) and mutant (Npc1 −/−) littermates. For tissue collection, female Npc1 −/− and Npc1 +/+ mice were sacrificed at 1, 3, 5, 7, 9 and 11 weeks of age using a rising concentration of carbon dioxide. The cerebral cortex was collected, flash frozen and stored at −80 °C until use. Mice homozygous for the Ccl3-targeted mutation (B6.129P2-Ccl3 tm1Unc/J strain, C57BL/6 genetic background) were used to test the effect of Ccl3 deficiency on the NPC1 pathological process. Pups were weaned 3 weeks after birth and subsequently had free access to water and standard mouse chow. PCR genotyping was performed using tail DNA. Primers and PCR conditions to genotype the Npc1 locus were previously described (Loftus et al 1997). Ccl3 locus genotype was determined using two pairs of primers: the wild-type allele was amplified using the sense 5′ATGAAGGTCTCCACCACTGC3′ and the antisense 5′AGTCAACGATGAATTGGCG3′ (yielding a 668 bp fragment); the mutant allele was amplified using the sense 5′CTTGGGTGGAGAGGCTATTC3′ and the antisense 5′AGGTGAGATGACAGGAGATC 3′ (yielding a 280 bp fragment). The PCR conditions were as follows: 94 °C x 3 min., then 35 cycles of: 94 °C × 30 s., 66 °C × 1 min., and 72 °C × 1 min. The final step was to heat at 72 °C × 7 min and hold at 4 °C until further analysis. Ccl3 −/− mice were intercrossed with Npc1 +/− mice to generate double-heterozygous animals, which were then backcrossed to Npc1 +/− for five generations. Double heterozygous animals from the N5 generation were then intercrossed to generate Npc1 +/+ Ccl3 +/+, Npc1 −/− Ccl3 +/+, Npc1 −/− Ccl3 +/− and Npc1 −/− Ccl3 −/− animals. Weight measurements were determined twice a week from 4 weeks of age. Animals were euthanized according to the ACUC protocol when they had lost 20 % of their maximal weight, and this was defined as the age of death for survival analysis.

Human studies

Human post-mortem tissue from age-matched control and NPC1 patients was obtained from the NICHD Blood and Tissue Bank (http://medschool.umaryland.edu/btbank/). Cerebellar tissue was obtained from controls UMB#754 (asthma attack), UMB#914 (motor vehicle accident), UMB#1841 (motor vehicle accident), UMB#5282 (asphyxia) and from NPC1 patients (confirmed mutations annotated from NM_000271.4) UMB#4237 (c.1628C > T/not detected), UMB#4770 (c.3107C > T/c.3573_3574insACTT), UMB#5372 (c.2842G > A/c.3182 T > C), UMB#M4002M (c.973_974dup/not detected). Frontal cortex tissue was received from NPC1 patients UMB#4237, UMB#4770, UMB#5372, UMB#M4002M, UMB#4214 (apparent c.3182 T > C homozygosity), UMB#M4003M (c.3134_3135insG/c.3566A > G), UMB#M4004M (c.2819C > T/c.3182 T > C) and controls UMB#754, UMB#914, UMB#1841, UMB#5282, UMB#1573 (motor vehicle accident), UMB#4670 (accident), UMB#5387 (drowned). No other data are available on these individuals.

NPC1 patients included in this study were enrolled between August 2006 and January 2011 in an Institutional Review Board-approved, longitudinal, Natural History/Observational trial at the National Institutes of Health (06-CH-0186, NCT00344331). Written, informed consent was obtained for all subjects. Assent was obtained when appropriate. Clinical diagnosis was confirmed by filipin staining of fibroblasts and NPC1 mutation analysis. Cerebrospinal fluid was collected via lumbar puncture. Control cerebrospinal fluid was obtained from 30 gender and age-matched patients who were undergoing cerebrospinal collection for another clinical indication. Four control patients were febrile (>38.5 °C) at the time of cerebrospinal fluid collection, but none had elevated white blood cell count or positive cultures. Age of control subjects ranged from 2 weeks to 20 years at the time of cerebrospinal fluid collection.

Cerebrospinal fluid biomarker measurements were made by Rules Based Medicine (Austin, TX) utilizing Multi-Analyte Profiling Technology. Statistical calculations were performed using GraphPad Prism. Log10 transformed concentration values were used to perform statistical analyses on normally distributed datasets. The D’Agostino and Pearson omnibus normality test was used. For cases in which less than 40 % of the total number of measurements was missing due to being below the limit of detection (LOD), the LOD/2 substitution method was used (Ganser and Hewett 2010). C3 concentrations in CSF of a subset of NPC1 patients covering the full age and severity range and controls (N = 14 and 11 respectively) were measured using the GenWay Biotech kit, according to the manufacturer’s recommendations. A 1:20 dilution of the CSF samples was used to bring the sample concentration into the assay concentration range.

RNA isolation

Frozen tissue samples, from a single mouse cortex, or approximately 100 mg from post-mortem human tissue, were homogenized in TRIzol (Life Technologies, Carlsbad, California) using Omni-TipsTM (Omni International, Marietta, GA). After chloroform addition and phase separation, total RNA was purified using the RNeasy Mini Kit (Qiagen, Valencia, CA) with an on-column DNase digestion step using the RNase-Free DNase Set (Qiagen, Valencia, CA).

PCR arrays

Following RNA extraction, total 7-week-old mouse or human RNA (1 μg) was reverse-transcribed using the RT2 First strand kit (SA Bioscience), as recommended by the manufacturer. The Mouse Inflammatory Cytokines and Receptors PCR Array (SA Bioscience) was performed using cerebral cortex tissue cDNA (N = 3) according to the manufacturer’s protocol. The Human Inflammatory Cytokines and Receptors PCR Array (SA Bioscience) was performed on both the human frontal cortex and cerebellum according to the manufacturer’s protocol (N = 3; NPC1 patients UMB#4237, 5372, M4002M, and controls UMB#754, 914, 1841). Genes for which an undetermined Ct value was reported in at least one sample were excluded from analysis. All raw data are available in Supplemental Table 1.

Validation of PCR array data

Total RNA (10 μg) was reverse-transcribed into cDNA according to the manufacturer’s instructions using the High-Capacity cDNA archive kit (Applied Biosystems). The following TaqMan assays (Applied Biosystems) were used to assess the expression level of several genes with altered expression including: Ccl8 (Mm01297183), Il1b (Mm01336189), Cxcl9 (Mm00434946), C3 (Mm00437858) and Cxcl10 (Mm00445235); C3 (Hs01100879), C4A (Hs00167147), CXCL3 (Hs00171061), CXCL5 (Hs01099660), IL10RA (Hs00155485), CEBP (Hs00270923), TOLLIP (Hs01553188) SPP1 (Hs00959010), CCR5 (Hs00152917), and CCL5 (Hs00174575). All gene assays were first validated using serial dilutions of control cDNA and compared to Gapdh/GAPDH (mouse Ccl8, Cxcl9, Il1b, and Cxcl10 assays; human CXCL3, CXCL5, CEBP, CCR5, TOLLIP, and SPP1 assays), or Actb/ACTB assays (mouse C3 assay; human C3, C4A, CCL5 and IL10RA assays). We were unable to confirm the probes for the following genes due to low expression levels in the tissues we tested: LTB, IL1F9, IL5 and IL9R.

Quantitative real-time PCR (qPCR) was performed using an Applied Biosystems 7900 real-time PCR system. Each sample was analyzed in triplicate, using 50 ng of total cDNA for each reaction. Human gene expression was validated on the full set of NPC1 patients and controls available (seven samples for each group for frontal cortex; four samples for each group for cerebellum). Mouse gene expression was validated on the 7-week-old samples used for the PCR array, as well as on an independent set of 1, 3, 5, 7, 9, and 11 week-old samples. The relative quantification of gene expression was performed with the comparative cycle number measured with the threshold method (CT) (Livak and Schmittgen 2001), using the control samples as a reference for quantification, and was plotted with mean and standard deviation (SD). Log10 transformed quantitation values were used to perform statistical analyses on normally distributed datasets. An unpaired t-test with Welch’s correction, when appropriate, was performed to assess the significance of the difference of means between control and mutant or patient samples. A list of all genes or proteins analyzed in this study is provided in Supplemental Table 1.

Pathology study on human brain post-mortem sample

Fixed tissue slices embedded in paraffin were obtained from the NICHD Brain and Tissue Bank for three patients and controls. Tissues were stained using a mouse anti-Human CD68 antibody (EBM11, Dako) and developed using the Vectastain ABC kit (Vector) using standard protocols.

Results

Expression of genes associated with inflammation in Npc1 mutant mouse cerebral cortex tissue

7-Week Npc1 mouse analysis

We conducted an inflammation-targeted gene expression analysis from cerebral cortex tissue isolated from 7-week-old Npc1 +/+ and Npc1 −/− mice. Twelve genes (Table 1) met our screening criteria of an absolute fold-change ≥2 and p-value ≤0.1. These 12 genes include members of the complement pathway, cytokine/chemokine family and the interleukin family. Il1b, Ccl8, C3, Cxcl9 and Cxcl10 genes were further validated via qPCR (Supplemental Fig. 1) with significantly increased expression of Il1b, Ccl8 and Cxcl10 genes. C3 and Cxcl9 expression appeared to be increased; however, not significantly (p-values 0.06 and 0.17 respectively), likely due to the high variability in expression levels for C3 and Cxcl9 in the mutant and control groups respectively.

Table 1 Genes significantly altered in cerebral cortex tissue from 7-week Npc1 −/− mice relative to control (Npc1 +/+) mouse tissue. A total of 12 transcripts were altered with a fold-change of ≥2 (p ≤ 0.1)
Full size table

Npc1 mouse model inflammation time course study

We additionally investigated the mRNA expression of the same five genes listed above as well as Ccl3 in an independent set of Npc1 mouse cerebral cortex tissue collected at 1, 3, 5, 7, 9 and 11 weeks of age (Fig. 1). The expression of all genes analyzed with the exception of Ccl8 was significantly increased in the Npc1 −/− mouse tissue compared to control littermates from 3 weeks of age across the time course study. Only Cxcl10 expression was found to be significantly increased at the 1 week time point.

Fig. 1
figure 1

Validation of mouse inflammation PCR array results. qPCR validation of six genes in Npc1 +/+ (white columns) and Npc1 −/− (black columns) cerebral cortex. One-week-old control mice were used as reference group for the relative quantitation, except for Ccl3 for which the 3-week-old control group was used as 1-week-old mice had undetectable transcript levels (BLD: below limit of detection). Mean and standard deviation are shown for each group (N = 4). An unpaired t-test with Welch’s correction, when necessary, was performed to determine the significance of the difference in means between the log10 values of control and mutant mice relative quantitation at each age: *p-value ≤ 0.05; **p-value ≤ 0.005; ***p-value ≤ 0.0001

Full size image

Effect of Ccl3 up-regulation in Npc1 pathological process

Ccl3 gene expression has previously been described as increased in the Npc1 mouse model (Liu et al 2010; Aqul et al 2011; Lopez et al 2012a). Our time course study indicated an early alteration of Ccl3 expression (Fig. 1), progressive dysregulation in the Npc1 mouse model and increased levels of CCL3 protein in cerebrospinal fluid from NPC1 patients (see below). These data, combined with the prior observation (Wu and Proia 2004) that deletion of the Ccl3 gene delayed the onset of neurological symptoms and increased survival in a mouse model of GM2-gangliosidosis (Sandhoff disease, HexB mutations), prompted us to further characterize the role of Ccl3 in NPC1 pathology. We hypothesized that, similarly to observations in the HexB mutant mouse, deletion of Ccl3 would improve the phenotype in Npc1 mutant mice. To test this hypothesis, we intercrossed Ccl3-deficient mice with Npc1 +/− mice to produce Npc1 +/− Ccl3 +/− mice. These mice were then intercrossed to obtain Npc1 +/+ Ccl3 +/+, Npc1 −/− Ccl3 +/+, Npc1 −/− Ccl3 +/− and Npc1 −/− Ccl3 −/− animals on a mixed C57BL/6 and BALB/c background. Most of the Npc1 −/− Ccl3 +/− and Npc1 −/− Ccl3 −/− animals showed weight loss and age of death similar to Npc1 −/− Ccl3 +/+ animals, with the notable exception of two double mutant mice that survived to ∼13 weeks of age. Npc1 deficiency in a C57BL/6 genetic background is more severe than in the BALB/c background (Parra et al 2011), thus the mixed genetic background could mask a beneficial effect associated with Ccl3 deletion. We therefore backcrossed Npc1 +/− Ccl3 +/− mice to the original BALB/c Npc1 +/− animals for five generations (∼97 % BALB/c) and compared weight loss and survival of these animals. No significant difference between Npc1 −/− Ccl3 −/− and Npc1 −/− Ccl3 +/+ mice was observed (Supplemental Fig. 2). Female Npc1 −/− Ccl3 +/− and Npc1 −/− Ccl3 −/− mice gained and maintained a closer-to-normal weight than Npc1 −/− Ccl3 +/+ mice between P40 and P55 days approximately, but all animals with a deletion of the Npc1 gene had lost 20 % of their maximal weight by approximately 75 days of age. Males with the Npc1 deletion showed comparable weights at all ages independent of the Ccl3 genotype.

Expression of genes associated with inflammation in human NPC1 frontal cortex and cerebellar tissue

We tested the presence of neuroinflammation in NPC1 human brains by CD68 immunostaining which is a common microglial activation marker. Frontal cortex and cerebellar tissue from control (Fig. 2a, b) and NPC1 patients (Fig. 2c, d) were analyzed where increased CD68 staining was observed for the NPC1 tissues.

Fig. 2
figure 2

Representative human tissue stained with the macrophage/microglia marker CD68. The top row shows cortex (400x) from control (a) and NPC1 (c); The bottom shows cerebellum (200X) from control (b) and NPC1 (d). Staining was performed on sections from the three patients (UMB#4770, 5372, and M4003M) and respective controls for which brain sections were available. Images are from patient UMB#5372, and control UMB#914

Full size image

A qPCR array analysis, analogous to that described above for the Npc1 mouse model, was performed on human frontal cortex and cerebellar brain tissue from control and NPC1 subjects. Nine genes (Table 2a) and ten genes (Table 2b) were differentially expressed with an absolute fold-change ≥2 and p-value ≤ 0.1 for frontal cortex and cerebellar tissue, respectively. Unexpectedly, only one identified gene, complement 3 (C3), was concordantly altered in mouse brain, human frontal cortex, and human cerebellar tissue. Interestingly, CCL5 was increased in Npc1 mutant mouse cerebral cortex and in cerebellar tissue from post-mortem NPC1 patients.

Table 2 a Array data displaying altered transcripts in NPC1 human post-mortem frontal cortex tissue versus control tissue. Nine transcripts were altered with a fold-change of ≥2 (p ≤ 0.1). b Transcripts found to be altered in NPC1 post-mortem cerebellar tissue versus control tissue. A total of ten transcripts were altered with a fold-change of ≥2 (p ≤ 0.1). The two genes listed in the sub-table for each tissue type were found to display drastic fold changes however with p-values > 0.1, but were chosen to validate via qPCR
Full size table

qPCR validation was performed for eight of these genes in both frontal cortex and cerebellar tissue from controls and NPC1 patients. We found six genes (C3, CEBP, CCR5, CXCL5, IL10RA, and SPP1) in which significant changes were observed in at least one of the tissue types tested (Supplemental Fig. 3). C3 expression was higher in NPC1 patients than the age-matched controls, except for the 5-year-old patient (Supplemental Fig. 3a). We additionally evaluated C4A (Supplemental Fig. 3b) and CXCL3 (Supplemental Fig. 3f) gene expression by qPCR, which displayed strong fold-change but not significant p-values on the PCR array data. In general, the expression levels of the ten genes tested were variable within the patient group, which may be related to the degree of variability in disease severity. The 5-year-old male patient showed normal levels of expression of all tested genes in both brain tissue types. The available clinical information indicates that he died before manifesting neurological symptoms, which might explain why none of these inflammatory-related genes showed altered expression.

Characterization of neuroinflammatory markers in cerebrospinal fluid from NPC1 patients

To expand and translate our findings from the mouse and human arrays, we evaluated neuroinflammatory markers in the cerebrospinal fluid from a cohort of NPC1 patients enrolled in the NIH Natural History/Observational study. This was accomplished by measuring the concentrations of 31 interleukins, cytokines and chemokines in the cerebrospinal fluid of NPC1 patients (n = 42) and pediatric controls (n = 30), using multi-analyte ELISA-based profiling (MAP). CSF levels of IL-3 (p = 0.0005), IL-16 (p = 0.04), CXCL5 (p = 0.03) and CCL3 (p < 0.0001) were increased in NPC1 subjects compared to controls, and CSF levels of IL-4 (p < 0.0001), IL-10 (p = 0.02), IL-13 (p = 0.006) and IL-12p40 (p < 0.0001) were decreased (Figs. 3 and 4a). Trends were observed for IL-1α (p = 0.06) and IL-7 (p = 0.07), while no statistical differences were detected for the interleukins IL-5, IL-6, IL-15 and IL-18 (Supplemental Fig. 4) and MCP-1, MMP3, ICAM-1, or stem cell factor (Supplemental Fig. 5).

Fig. 3
figure 3

Cerebrospinal fluid inflammation markers. Log10 transformed concentrations of seven cytokines in cerebrospinal fluid from controls (circles) and NPC1 patients (squares). Mean and standard deviation are shown for each group. An unpaired t-test with Welch’s correction when necessary was performed to determine significance

Full size image
Fig. 4
figure 4

CCL3 marker. a Log10 transformed concentrations of CCL3 in cerebrospinal fluid from controls (circles) and NPC1 patients (squares). The patient group was additionally separated between patients receiving miglustat treatment (triangles), and patients not receiving miglustat treatment (inverted triangles). Mean and standard deviation are shown for each group. An unpaired t-test with Welch’s correction when necessary was performed to determine the significance of the difference in means between controls and patients. b Correlation of CCL3 concentrations in cerebrospinal fluid with patient annual severity increment score (severity score divided by their age) followed by numerical ranking (i.e., 1 to 59 representing least to most severe phenotype)

Full size image

The classical pro-inflammatory cytokine TNFα has been shown to be elevated in astrocytes from 7-week-old Npc1 −/− mice (Wu et al 2005). Therefore we hypothesized that TNFα would be elevated in the CSF in NPC1 patients as a reflection of the neuroinflammation. Most NPC1 patients had undetectable levels of TNFα in CSF as measured using MAP (limit of detection 1.4 pg/mL). Using a high-sensitivity ELISA kit specific for TNFα, we still found that the majority of tested samples were below the limit of detection (0.1 pg/mL). The TNFα transcript was also measured via the array analysis where no changes were detected in the mouse or human cerebellum but elevation was observed in the human frontal cortex (Table 2a).

Based on the elevated levels of C3 in both mouse and human tissue samples, we were interested in determining if CSF levels of C3 were altered. As C3 was not evaluated in the MAP experiment, we measured C3 concentrations in CSF from a subset of NPC1 patients and controls from our cohort. No significant differences were observed. Mean CSF C3 levels were 1096 ± 333 ng/mL and 726.6 ± 58.83 ng/mL for control (n = 11) and NPC1 (n = 14) samples respectively (p = 0.8). Another similarity between the array data and CSF data was for CXCL5. This gene was found to be elevated in NPC1 patients’ tissues (3.9-fold, p = 0.003 in cerebellum; 4.0-fold, p = 0.03 in cortex) compared to controls. Furthermore, in the CSF the mean CXCL5 level was found to be mildly elevated in the NPC1 patient cohort (Fig. 3g).

Approximately half of the patients in the NIH Natural History cohort are taking miglustat off-label; therefore, we investigated whether miglustat therapy had any effect on the concentrations of the eight significantly modified markers. IL-3 and IL-13 concentrations were decreased slightly when comparing miglustat treated and untreated patients (p-value < 0.05; Supplemental Fig. 6a, e). Since individual responses could be masked by the large degree of inter-patient variation, we compared CSF concentrations of these same eight inflammation markers in five NPC1 patients for whom we had serial cerebrospinal fluid samples prior to and after initiation of off-label miglustat use. In this subset of five patients only IL-10 had a coordinate response (Supplemental Fig. 6d). More extensive, long-term analysis is needed to establish if IL-10 CSF levels provide a useful indication of neuroinflamation in NPC1 based on these results.

Given the phenotypic heterogeneity observed in NPC1, we were interested in determining if cerebrospinal fluid concentrations of the markers correlated with NPC1 disease status. Using the NIH NPC1 neurological severity score (Yanjanin et al 2010), we observed a significant correlation between CCL3 levels and annual severity increment score (neurological severity score/age; r2 = 0.17, p-value < 0.002; Fig. 4b).

Discussion

This discovery based analysis revealed several inflammatory markers which are altered in the tissue of Npc1 mutant mice or NPC1 human tissue specimens relative to controls. The majority of candidate markers from the PCR array in NPC1 post-mortem tissue displayed increased expression in the NPC1 samples relative to controls with only one transcript displaying decreased expression. Inflammation in human NPC1 post-mortem tissue was also confirmed via CD68 staining for activated microglia in both the cerebellum and frontal cortex. Furthermore, measurements of inflammatory markers in cerebrospinal fluid from NPC1 patients and pediatric controls show marked differences for a small set of inflammatory proteins.

Complement C3 was the only gene to be found altered in all samples (mouse cerebral cortex, human cerebellum and human frontal cortex). The expression of some components of the complement pathway had previously been reported as increased in Npc1 mouse model cerebellum, specifically C1q, C2, C3, C4b and the two receptors C3ar1 and C5r1 (Liao et al 2010; Vazquez et al 2011; Lopez et al 2012a, b). All cell types in the brain are able to produce complement factors (Veerhuis et al 2011), and the complement pathway contributes to the inflammatory process of several brain diseases, including Alzheimer disease (Rubio-Perez and Morillas-Ruiz 2012). A recent study investigating the role of the complement pathway in neuron survival in NPC1 revealed that complement gene deletion did not rescue neuronal death, and therefore innate inflammation, while present in NPC1, is not the major causative factor of neurodegeneration (Lopez et al 2012b).

Activated microglia was observed in brain tissue of the Npc1 mouse model in several studies via positive CD68 staining (Liu et al 2009; Smith et al 2009; Liao et al 2010; Ramirez et al 2010; Aqul et al 2011; Pressey et al 2012). We also observed this marker of activated microglia in post-mortem tissue from NPC1 patients while it was absent in control tissue. The common classification of “activated microglia” stems from the classical morphological change and the production of neurotoxin or neurotrophic molecules as well as microglia migration to the site of injury.

Several classical inflammatory markers have been reported as being dysregulated in the Npc1 mouse model, including IL1β (Baudry et al 2003; Repa et al 2007) and TNFα (Li et al 2005; Wu et al 2005; Langmade et al 2006; Aqul et al 2011). In our expression study of mouse brain tissue, we identified Il1b as being increased. However, it was not modified in human tissues. Similarly, TNFα was only detected as upregulated in NPC1 human cortex tissue. Additionally, both markers were undetectable in cerebrospinal fluid of most NPC1 patients and controls. Other classical markers of inflammation, such as IL-6, IL-8, and MCP1, had comparable levels in patients and controls cerebrospinal fluid, which suggests that the neuroinflammatory mechanism present in NPC1 might be different from a classical neuroinflammation process reported in other disorders.

Eight of the 31 inflammatory markers tested showed significantly different CSF concentrations in NPC1 patients compared to pediatric controls including decreased levels of IL-4 and 10. It has been shown that activated microglia result in the production of IL-4 and IL-10 to reduce neuron degeneration (Park et al 2005; Henry et al 2009); however, the mechanism for IL-10 production within the CNS is still a topic of ongoing research (Chabot et al 1999). Interestingly, little to no change was noted in any of the analyte levels measured from NPC1 patients on miglustat versus those not on miglustat treatments, and only IL-10 displayed a slight decrease after initiation of miglustat therapy. The fact that miglustat treatment did not change the CSF marker levels in NPC1 patients would suggest that the potential anti-inflammatory properties of miglustat may affect classical markers, but not those found to be significantly altered in this study.

From our gene expression experimental results Ccl3 was found to be altered in the cerebral cortex of 7-week-old Npc1 mutant mice. In an effort to understand if deletion of this gene would improve the survival and therefore the phenotype of NPC1 disease, we generated a double mutant mouse model. In contrast to the Sandhoff disease model, the genetic deletion of Ccl3 in the Npc1 mouse model did not show any improvement with regards to survival or weight loss in the Npc1 mutant animals. This suggests that the origins of the brain inflammation observed in the two lysosomal storage diseases are likely different. In particular, macrophage infiltration from the periphery is observed in Sandhoff disease (Wu and Proia 2004), and does not appear to occur in NPC1 disease (Lopez et al 2012a). The effect of Ccl3 deletion in Npc1 mice has been reported in the F1 generation (Lopez et al 2012a). In agreement with our finding, the authors did not observe an improvement in the double Npc1 and Ccl3 mutant phenotype compared to the Npc1 mutants. However, they reported a lower weight in double-mutant males compared to Npc1 −/− Ccl3 +/+, and a slightly shorter survival rate, but no differences between females. On the contrary, we observed a small temporary benefit in weight gain for Npc1 −/− females with deletion of one or two Ccl3 alleles, and no differences for males. Genetic background might have an impact on this specific trait, as Npc1 −/− mice have a more severe phenotype on a C57BL/6 background than on a BALB/c background (Parra et al 2011). Considering that we backcrossed our mice for five generations, the effect of the C57BL/6 background was probably negligible in our cohort of animals.

Despite the absence of a positive effect of Ccl3 deletion on NPC1 disease process, we identified increased concentrations of CCL3 in cerebrospinal fluid of NPC1 patients compared to age-matched controls. Correlation of CCL3 levels with disease severity and the clear separation of the patient group compared to controls make CCL3 an interesting biomarker for evaluation of NPC1 disease status. Future studies aimed at developing an appropriate pharmacological therapy for NPC1 disease will be strengthened by the ability to monitor CCL3 CSF concentration in these patients.

To summarize, extensive work conducted within the last decade has documented a progressive inflammatory process occurring in NPC1 disease. The vast majority of this work has focused on the Npc1 mouse model. We therefore investigated the neuroinflammatory process in NPC1 patients. Based on our studies, we have identified several potential inflammatory markers that may prove useful in correlating with neurological progression or be utilized to facilitate the design and implementation of clinical trials to demonstrate potential efficacy of candidate therapies.