Abstract
Niemann-Pick disease type C1 (NPC disease) is a neurodegenerative multi-lipid lysosomal storage disease caused by mutations in the NPC1 gene presenting with reduced lysosomal Ca2+ signalling and inhibited late endosome-lysosome transport. Elevating cytosolic Ca2+ levels in NPC cells has been shown to reduce lysosomal lipid storage. Treating Npc1-/- mice with the Ca2+ modulator curcumin led to reduced lipid storage, improved life expectancy and function. These studies led to reported utilisation of curcumin supplements by NPC disease families despite there being no clinical evidence of benefit and a report indicating no benefit of nanoformulated curcumin in Npc1-/- mice. The aim of this study was to determine whether various commercially available curcumin nanoformulations were capable of reproducing the findings obtained with unformulated pharmaceutical grade curcumin. We compared seven curcumin nanoformulations in Npc1-/- mouse astrocytes. All the nanoformulations elevate cytosolic Ca2+ levels but only two lowered lysosomal lipid storage. Importantly, some caused elevations in NPC lysosomal storage and/or decreased cellular viability. Although this is an in vitro study, our findings suggest that care should be taken when contemplating the use of curcumin supplements for NPC disease.
Background
NPC disease is a neurodegenerative lysosomal storage disease characterised by progressive ataxia, hallmarks of Alzheimer disease, dystonia and hepatosplenomegaly1. At the cellular level the disease is characterised by accumulation of multiple lipids including cholesterol, sphingosine, glycosphingolipids (GSLs), lyso-(bis)phosphatidic acid (LBPA) and sphingomyelin within late endosomes and lysosomes1-3. The disease is caused by mutations in either the NPC1 or NPC2 genes with the NPC1 gene encoding a 13 transmembrane domain protein (NPC1) residing within the limiting membrane of the lysosome1. Whilst the exact function of this protein remains unknown it has been shown to bind to cholesterol and to have greatest homology to the resistance nodulation division (RND) family of bacterial multi-substrate efflux pumps1,4. At present, the only approved therapy in Europe for NPC disease is miglustat5, which reduces GSL biosynthesis and storage6, and if given early enough can slow disease progression but does not reverse the disease course7.
A major contributing factor to the accumulation of lipids in NPC1 disease cells is the retardation in endosomal transport including defective delivery of material to lysosomes3, defective fusion between late endosomes and lysosomes2, and defective transport of material between late endosomes/lysosomes and other organelles including the Golgi or the endoplasmic reticulum (ER)1,8. We have previously shown that an important contributing factor to these defects is the reduced nicotinic acid adenine dinucleotide (NAADP) mediated Ca2+ release from lysosomes2, a necessary trigger for mediating endocytic transport9. This endocytic trafficking defect in NPC cells can be overcome by inducing transient global elevations in cytosolic Ca2+ levels2,10,11. Using curcumin, a weak inhibitor of the sarco/endoplasmic reticulum (ER) Ca2+ ATPase (SERCA)12, significant Ca2+ release from the ER can be triggered leading to a transient large elevation in cytosolic Ca2+ that in turn corrected the endocytic trafficking defect and led to a reduction in lipid storage in NPC1 disease cells2. This benefit was caused directly by the elevation in cytosolic Ca2+ as it could be blocked using the Ca2+ specific chelator BAPTA-AM and had no effect on the reduced lysosomal Ca2+ levels2.
We observed similar benefit in vivo in the Npc1-/- mouse model with a reduction in lipid storage, an increase in life expectancy and improved function2. More recently, it has also been shown that curcumin treatment of the Npc1-/- mouse leads to ∼16% increase in lifespan (not dissimilar to the 25% increase observed with miglustat), improved performance on the balance beam13 and greater survival of Purkinje neurons with limited effect on inflammation, suggesting that the benefit is mediated by modulating Ca2+ levels rather than the antioxidant properties of curcumin14. Despite these studies it has been suggested that the concentration of curcumin that enters the blood and tissues is not high enough to modulate Ca2+13. A study in rats found that an oral dose of 2g/kg reached a maximum serum concentration of ∼4μM after 50 minutes15 and prolonged treatment of mice with 83mg/kg/day can maintain similar levels (∼1μM) within the brain16, concentrations that have been shown to have impact on behavior17. However, to boost the bioavailability of curcumin, and to reduce dosing, it is often encapsulated in a lipid mixture13,18. A study in mice demonstrated that 0.5% of a 5mg/kg injection of nanocurcumin was capable of entering the brain within 1h and achieving serum concentrations of ∼10μM18. Furthermore, a recent study has highlighted that different sized curcumin nanosupensions/nanoformulations occupy tissues at varying concentrations dependent on their particle size. For example, with initial plasma concentrations of around 5μM following an intravenous injection the highest concentration in the liver were observed with 200nm curcumin particles at 10 min post injection whereas in the brain the highest concentration was observed with 70nm particles at 20-30mins post injection19. These studies indicate that higher therapeutic concentrations of curcumin can be achieved using nanosuspensions and that tissue penetrance may be dependent on particle size. Interestingly however, no benefit on Npc1-/- mouse function was observed when curcumin was combined with a lipid carrier to increase bioavailability13. This suggests that either the lipid carrier interferes with the mode of action of curcumin or that particle size was sub-optimal.
It has come to our attention that some NPC disease patients are taking curcumin supplements20. Several of these curcumin supplements contain lipid mixtures consisting of phospholipids and fatty acids15. NPC disease cells are known to have defects in the efflux of phosphatidylcholine to apolipoprotein A121, suggesting that some of these lipid mixtures could accumulate in NPC disease cells. Whilst some of the encapsulated curcumin would be delivered to cells of the intestine, which does not function normally in some NPC1 disease patients22, several nanoparticle vectors actually reduce clearance of the drugs they encapsulate resulting in accumulation in tissues such as the liver23, which is also affected in NPC1 disease22. As no information exists about the potential effects of these combined curcumin/lipid complexes on NPC1 disease cells, and as the previous study mentioned above did not observe any benefit when using curcumin complexed with a lipid vehicle13, we have undertaken a study to determine whether these curcumin nanoformulations are capable of reducing NPC lipid storage in vitro and whether their mode of action is in any way related to their nanoparticle size distribution.
Methods
Reagents
Unless otherwise stated, all reagents were obtained from Sigma-Aldrich, Dorset, UK.
Cells
Primary astrocytes from post-natal day 1 Npc1+/+ and Npc1-/- mouse were prepared as previously described2 and were maintained in culture as monolayers grown in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% foetal bovine serum (FBS) and 1% L-glutamine only at 37°C in a humidified incubator with 5% CO2. For all experiments, cells were seeded at equal densities of 5,000 or 20,000 cells on 8 well chamberslides (Ibidi, Thistle Scientific, UK) or Cellbind 96 well plates (Sigma-Aldrich, UK) respectively and left to adhere overnight prior to treatment with the curcumin formulations for the indicated times.
Preparation and solubilisation of curcumin neutraceuticals
Curcumin supplements, summarized in Table 1, were purchased in multiple batches from commercial retailers. Three randomly selected capsules from each batch were opened, the powder weighed and solubilised in dimethylsulfoxide (DMSO, VWR, UK). We determined the curcumin content from the manufacturers’ stated ratios of curcumin to lipid carrier and bulking agents (Table 1, Fig. 1B) and generated a 10mM stock solution in each case. Each curcumin supplement was used at a final concentration of 30μM ensuring that the DMSO content in the treated cultures always remained at 0.3% v/v, appropriate DMSO and, where possible, lipid or bulking agent controls were also used.
Cytosolic Ca2+ assays
Changes in cytosolic Ca2+ in response to curcumin formulations were measured using the membrane permeant acetoxymethyl (AM) derivative of the intracellular ratiometric Ca2+ indicator Fura-2,AM (Stratech Scientific, UK). Cells were loaded in DMEM as previously described2 with 5μM Fura-2AM and 0.025% Pluronic F127 for 1h at room temperature followed by 10min de-esterification of the AM ester group to release free cytosolic Fura 2. For all experiments, cells were bathed in an extracellular solution of Hank’s balanced salt solution (HBSS, Invitrogen, Paisley, UK) containing 1mM MgCl2, 1mM CaCl2 and 10mM HEPES (pH 7.2). Changes in cytosolic Ca2+ in response to the curcumin formulations were measured every second using 360nm and 380nm LED excitation wavelengths with emission measured at 520nm on a Zeiss Colibri LED fluorescence microscope equipped with a high speed Mrm CCD camera. Videos were recorded using the Axiovision Physiology Ca2+ imaging package which was also used for preparation and analysis of the ratiometric Fura 2 Ca2+ traces. Regions of interest (ROIs) were drawn over the whole cell and a minimum of 20 cells were analysed per experimental repeat.
Lipid cytochemistry
Intracellular cholesterol and ganglioside GM1 localisation and accumulation were measured in cells fixed for 10 minutes with 4% paraformaldehyde (Fisher Scientific, Loughborough, UK) using filipin and fluorescein isothiocyanate (FITC) labelled cholera toxin B subunit (FITC-CtxB) respectively, as previously described2,6. Briefly, for cholesterol staining, fixed cells were incubated with 175μg/ml filipin in DMEM with 10% FBS at room temperature for 30mins followed by one wash in DMEM with 10% FBS and two washes in Dulbecco’s Phosphate Buffered Saline (DPBS) prior to mounting in a solution of 4.8% MOWIOL 4-88 (Merck, Darmstadt, Germany) and 12% glycerol in PBS. For ganglioside GM1 staining, fixed cells were incubated with 1μg/ml FITC-CtxB in DPBS containing 1% bovine serum albumin (BSA) and 0.1% saponin at 4°C for 16h followed by three 5 min washes in DPBS and mounting in MOWIOL 4-88. Lysotracker Red (Invitrogen) staining for live visualisation of lysosomes in cells was performed as described6, cells grown on chamberslides were incubated with 200nM Lysotracker Red in DPBS for 15min at room temperature followed by three washes in DPBS and immediate imaging. Nuclei were counterstained by addition of 2μg/ml Hoechst 33358 or Sytox Green (Invitrogen) either in PBS or added to the lysotracker staining solution. In all cases, cells were imaged using a Zeiss Colibri LED microscope with Axiovision 4.8 software, post-analysis was performed using ImageJ (NIH) and Adobe Photoshop CS6.
Texas Red dextran endocytosis
Fluid phase endocytosis of Texas Red dextran was measured in live cells incubated for 4h with the curcumin nanoformulations in conjunction with 0.25mg/ml 10kDa Texas Red dextran in DMEM medium supplemented with 10% FBS and 1% L-glutamine. Cells were then washed 3 x 5min with cold DMEM medium with 10% FBS supplemented with 1% BSA and 0.5mg/ml unlabeled 10kDa dextran to remove non-internalised Texas Red dextran not internalized into the cells that is associated with the plasma membrane. Cells were then washed three times with DPBS and were immediately imaged using a Zeiss Colibri LED microscope.
Thin layer chromatography (TLC)
Lipids were extracted from 100mg of the powdered curcumin nanoformulations in chloroform/methanol (VWR, West Sussex, UK) at 1:2 (v:v) as previously described3. After centrifugation (5 min, 1000 rpm) to remove sediment, solvent extracted lipids were washed three times by phase separation via the addition of 1mL of CHCl3 and 1mL of PBS, centrifugation and repeated removal of the aqueous phase, the final organic phase was dried under N2. Purified lipids were resuspended in ethanol, separated on high performance thin layer silica gel 60 chromatography (HPTLC) plates (Merck) and visualised with p-anisaldehyde as previously described24. The developing solvent systems used were: chloroform:methanol:H2O 65:25:4 for improved separation of phospholipids and 80:10:1 for improved separation of cholesterol from ceramides. Lipid standards were included on all TLC plates at the indicated concentrations and were obtained from Avanti Polar Lipids (Alabama, USA).
Cell viability assays
Cellular viability following either 16h or 48h treatment with the curcumin formulations was determined on live cells by fluorescence microscopy using the early apoptotic marker Annexin A5. Following curcumin treatment, cells were incubated for 30mins on ice with 5μg/ml Alexa Fluor 488-Annexin A5 (Invitrogen) in HBSS supplemented with 10mM HEPES (pH 7.2), 1mM MgCl2 and 1.2mM CaCl2. Cells were then washed three times and imaged in the same buffer at 4°C (to prevent internalization of the Annexin A5) or by MTS assay (Promega) in 96 well plates respectively.
Nanoparticle size analysis
The size distribution and concentration of nanoparticles in the curcumin formulations were analysed using a NanoSight. Following solubilisation in DMSO, the curcumin formulations were diluted 1 in 50,000 in mqH2O prior to loading onto the NanoSight and pump assisted flow over a 488 laser at a speed of 50. Movies were captured at 25FPS using a sCMOS camera and analysis was performed using the NanoSight software at a detection threshold of 2.
Results
For this study we compared a range of curcumin formulations all of which, apart from turmeric root extract (TRE), are complexed with a range of lipid carriers in order to improve bioavailability (Table 1). Almost all, apart from BCM95s, which is solubilised in the essential oils of turmeric, contain one or a mixture of the following; lecithin, fatty acids, triglycerides, phospholipids and stearic acid. One, CGM, contains fenugreek galactomannans. We have performed the first comparison of the size properties of these curcumin nanoformulations and have confirmed that all are nanoparticles (Fig. 1a). Four, including TRE, CGM, SLNM and SLNA are in the range of 75-85nm whilst three have peak sizes at 100-110nm including SLNL, BCM95s and BCM95g. A second broader peak ranging from 110-175nm also exists for SLNL, SLNM and BCM95s representing aggregation of these particles. A third peak at 220nm can also be seen for BCM95s indicating this formulation made from the essential oils of turmeric has the most diversity in terms of particle size which is tempered by the addition of lecithin and beeswax in the case of BCM95g (Table 1).
In addition to confirming that these curcumin formulations are nanoparticles we also determined the curcuminoid content in the nanoformulations via solvent extraction and separation by HPTLC as described in materials and methods. Using this method we were able to separate the three major curcuminoids, namely curcumin (largest band), desmethoxycurcumin and bis-desmethoxycurcumin in all of the nanoformulations (Fig. 1b). Our results are largely in keeping with the total curcumin content as reported (Table 1) with TRE having the highest overall curcuminoid content followed by BCM95g and BCM95s whereas SLNL, SLNA and SLNM have the lowest curcumin content.
In order to determine whether the curcumin supplements were likely to have any beneficial effect on NPC cells we first determined whether they could induce an elevation in cytosolic Ca2+ levels. All of the supplements were able to induce rapid elevation in cytosolic Ca2+ in both Npc1+/+ and Npc1-/- astrocytes at 30μM (Fig. 2a). Interestingly, whilst the TRE, BCM95s, and BCM95g supplements all released similar levels of Ca2+ in Npc1+/+ and Npc1-/- astrocytes, the CGM, SLNL, SLNM, and SLNA supplements all induced greater release in the NPC1 disease cells compared to controls (Fig. 2b). The SLNA formulation elevated intracellular Ca2+ by 2-2.5 times more in the Npc1-/- cells compared to the Npc1+/+ controls. Presumably this is a combined result of the ability of the SLN lipid vehicle in particular to enhance delivery across the plasma membrane coupled to altered NPC1 disease membrane fluidity25 leading to greater release of curcumin within the cell.
Having confirmed that all the curcumin formulations were capable of elevating intracellular Ca2+ levels we next determined whether this could induce a reduction in NPC1 lysosomal lipid storage as previously reported with pure unformulated curcumin2,10. Surprisingly, although SLNM had the greatest effect on elevating cytosolic Ca2+ in Npc1-/- disease astrocytes it had no beneficial effect on lysosomal storage. Indeed, we observed an increase in lysosomal accumulation of cholesterol (Fig. 3a), a smaller increase in ganglioside GM1 (Fig. 3b), and further expansion of the lysosomal system visualized by lysotracker staining (Fig. 3c) in Npc1-/- cells treated with SLNM. Despite their ability to elevate cytosolic Ca2+ to a greater degree in the Npc1-/- astrocytes, CGM had no effect on lipid storage or lysosomal expansion whereas SLNL, in a manner similar to SLNM, consistently led to an increase in lipid storage of cholesterol (Fig. 3a), gangliosides (Fig. 3b), and an expansion of lysosomes (Fig. 3c), no changes were observed in the Npc1+/+ cells (not shown). In contrast to the other two SLN nanoformulations, SLNA had no effect on cholesterol storage (Fig. 3a), ganglioside storage (Fig. 3b) or lysosomal expansion (Fig. 3c). Two curcumin supplements consistently emerged as having the greatest impact on lowering lysosomal lipid storage in the Npc1-/- cells, namely BCM95s and TRE with reduction in cholesterol (Fig. 3a), ganglioside GM1 (Fig. 3b), and lysosomal expansion (Fig. 3c) observed with BCM95s and a reduction in cholesterol (Fig. 3a) and lysosomal expansion (Fig. 3c) with TRE. No detrimental effect of any of the supplements on inducing lysosomal storage of these lipids in Npc1+/+ cells was observed (not shown).
To determine the cause of the elevated lipid storage levels in the Npc1-/- cells treated with SLNL and SLNM we investigated whether incubation with these curcumin nanoformulations had any effect on endocytosis, which is known to be altered in NPC disease and is the main route for bulk lipid entry into the cell. Following a joint incubation of the cells with both the curcumin nanoformulation and 10kDa texas red dextran for 4h we observed some key differences between the different formulations. In parallel with the reduced lipid storage observed in Npc1-/- cells treated with TRE and BCM95s (Fig. 3), and in keeping with previous data on curcumin2 we also observed a partial correction in the endocytic transport defect (Fig. 4) with these two curcumin formulations. NPC disease cells have been shown to have a delay in transport between early and late endosomes2,6,26, following 4h treatment with Texas Red dextran, this probe can be seen to cluster around the nucleus in late endosomes and lysosomes in the Npc1+/+ cells (Fig. 4), whereas in the Npc1-/- cells it has a broader distribution representative of early endosomes as well as some late endosomes. Both TRE and BCM95s, as well as SLNA, appear to have partially rescued this transport defect with Texas Red dextran staining now clustered in a peri-nuclear region indicative of late endosomes and lysosomes, very little staining in proximity to the plasma membrane, indicative of early endosomes, can be seen. Interestingly, both SLNM and SLNL appear to have either reduced the entry of texas red dextran into the Npc1-/- astrocytes or enhanced it’s recycling out of the cell as the total level of cellular fluorescence is lower by ∼65% and ∼85% respectively compared to the untreated cells (Fig. 4). This would appear to suggest a connection between the elevated lipid storage levels and a further defect in endocytosis in the Npc1-/- cells, however, we also observed reduced fluorescence indicating reduced internalization of the Texas Red dextran probe in the Npc1-/- cells treated with BCM95g (∼88%) and the Npc1-/- cells treated with CGM (∼65%). As no lipid storage was observed in these cells (Fig. 3) it must be concluded that the defect in endocytosis of texas red dextran is not the cause of the elevated lipid storage observed with SLNM and SLNL.
Having observed that some of the curcumin nanoformulations increased lipid storage in NPC disease cells and that this may not be due to defects in endocytosis induced by the curcumin formulations we decided to determine the nature of the lipid species in each formulation by solvent extraction and separation by HPTLC. Perhaps unsurprisingly, TRE had the lowest lipid content with very few bands present which correlate with those observed at a higher level in BCM95s and BCM95g (Fig. 5a and 5b), both of which contain essential oils of curcumin that are presumably present in lower concentrations in a total turmeric root extract (TRE). CGM, which has few lipids, also contains one of these bands, the identity of which is currently unknown but could possibly be related to the galactomannan present in CGM or is a component of the curcumin used in manufacturing CGM. Interestingly, the three curcumin species are visible in all lanes (compare with Fig. 1b) between the glucosylceramide and cholesterol bands (Fig. 5a). As well as TRE and CGM, two other curcumin nanoformulations, BCM95s and SLNA, contained very few lipids with only the reported triglycerides and fatty acids present for SLNA (the band above cholesterol). Otherwise, the remaining curcumin supplements (SLNL, BCM95g, and SLNM) had significant levels of a variety of lipids. SLNM has the highest lipid content (Fig. 5a and 5b) and is reported as using phospholipid to solubilize curcumin, by similarity to the standard this could represent lecithin (Fig. 5c) as well as to BCM95g (Fig. 5a, 5b and 5c) which incorporates lecithin (Table 1). SLNL is also reported to contain lecithin (Table 1), this appears to be the case although there are fewer bands when compared to the standard (Fig. 5c) with one or two fainter additional bands also present which are also seen in SLNM (Fig. 5c). For the purposes of this study, our TLC analysis largely confirms the stated lipid content of these formulations whilst also indicating the presence of a few other lipids. Perhaps of most importance is that none of the supplements contained lipids that are known to accumulate in NPC disease. To confirm this we ran the HPTLC plates in two different solvent systems (Fig. 5a and 5b). First we used a solvent system comprising CHCl3:MeOH:H2O 65:25:4 to separate LBPA from sphingomyelin and which also allows visualization of neutral glycosphingolipids such as glucosylceramide (three lipids that are stored in NPC disease). We observed a small amount of LBPA and sphingomyelin in SLNM but did not observe any of these lipids in the other formulations (Fig. 5a). In order to separate cholesterol and ceramide we used a solvent system comprising CHCl3:MeOH:H2O 80:10:1. We did not observe any appreciable amount of cholesterol in any of the formulations (Fig. 5b). We can therefore rule out that the lipid formulations are themselves the source of the additional lipid storage that we observe in the Npc1-/- astrocytes treated with SLNL and SLNM. However, both SLNL and SLNM contain high amounts of phospholipid (Fig. 5a-c), so a change in Npc1-/- lysosomal metabolism as a result of phospholipid accumulation cannot be ruled out.
Finally, having shown that some of the supplements could alter Ca2+ levels differentially between the Npc1+/+ and the Npc1-/- astrocytes, with greater release in Npc1-/- (Fig. 2), that some caused an increase in lysosomal storage (Fig. 3), and that some induced defects in endocytosis (Fig. 4) potentially caused by membrane damage triggered by the nanoformulation itself27, we decided to test whether any of the curcumin formulations had any effect on cellular viability. First we utilized an early marker of apoptosis, extracellular live binding of FITC-Annexin A5 to plasma membrane phosphatidylserine (PS). PS is externalized to the outer leaflet of the plasma membrane as one of the first events in apoptosis. Following 16h treatment with the supplements, no staining of extracellular PS by FITC-Annexin A5 is detected in any of the conditions apart from the Npc1-/- cells treated with CGM and SLNL (Fig. 6a). As a positive control to confirm that staining is indicative of apoptosis we treated cells with nigericin, a molecular poison, and observed plasma membrane FITC-Annexin A5 staining (Fig. 6a). Interestingly, some intracellular staining of FITC-Annexin A5 indicative of endosomes is observed in the Npc1-/- disease cells treated with SLNL (Fig. 6a), suggesting either the possibility of necrosis or that the curcumin nanoformulation has disrupted the plasma membrane sufficiently to allow Annexin A5 to enter but not a 10kDa dextran (Fig. 4a). To confirm our viability findings with FITC-Annexin A5 we used a metabolic marker of cellular viability, namely mitochondrial activity measured using MTS. Of the curcumin formulations tested, all bar SLNA had some effect on Npc1-/- mitochondrial activity and cellular viability following a 30h incubation with the supplements (Fig. 6b). TRE, BCM95g and BCM95s had a minimal ∼7-8% reduction in mitochondrial function whereas SLNM, SLNL and CGM substantially, and significantly, reduced cellular viability by ∼40-50% respectively. No detrimental effect on cell viability of any of these formulations was observed on Npc1+/+ cells (not shown).
Discussion
Unformulated curcumin has been shown to be able to ameliorate NPC disease in five separate studies2,13,14,28,29, whilst curcumin formulated into a lipidated vector has been shown to have little to no benefit13. The aim of the present study was therefore to determine whether formulated lipidated curcumin could mediate any benefit in NPC1 disease cells. Our project focused on a range of lipidated curcumin nanoformulations (Table 1) which are used for increasing the bioavailability of curcumin which is otherwise poorly absorbed through the intestines15. Formulation of curcumin into a lipid vector allows for improved delivery across the blood-gut barrier, the achievement of higher concentrations of curcumin in the blood and tissues, and reduced renal clearance, all of which is essential for treating disease19. Recent evidence has shown that formulation into nanoparticles allows for even greater penetration into the blood and higher steady state levels, dependent on particle size19. Our study is the first to demonstrate that all of these formulated lipidated curcumin particles are nanoparticles, ranging in size from 50-250nm.
With respect to the ability of these nanoformulations to modulate intracellular Ca2+, all were capable of inducing Ca2+ release from the ER and elevating cytosolic Ca2+ levels. It was interesting to note that four, CGM, SLNA, SLNM and SLNL, induced greater Ca2+ elevation in the Npc1-/- cells but this did not correspond with a reduction in lysosomal storage as would have been expected2. However, this enhanced intracellular Ca2+ release elicited by the majority of these nanoformulations (with the exception of SLNA) did correspond with increased cellular toxicity, namely reduced mitochondrial activity and apoptosis. However, another possibility is that CGM, SLNA, SLNM and SLNL may be the least stable of the nanoformulations and release their curcumin cargo in a more rapid manner upon entry into the cell compared to the other nanoformulations15, thus having a more significant effect on intracellular Ca2+.
A surprising result of our work is that although all of the nanoformulations of curcumin could modulate intracellular Ca2+, very few actually had an impact on NPC disease lysosomal lipid storage. As unformulated curcumin has been shown in several studies to be effective2,14,28,29, we hypothesised that this was due in some way to the properties of the lipid carrier. Although SLNA appears to have a small benefit on some components of NPC1 lysosomal storage, no effect is seen with CGM, whilst treatment with SLNM and SLNL led to an elevation in lysosomal storage. This worsening of the NPC1 lysosomal storage phenotype was greatest with SLNL, but the exact reasons underlying this unexpected phenotype are unclear. SLNL and SLNM both appeared to significantly reduce endocytosis of Texas Red dextran, which might explain the elevated lysosomal storage. However, we also observed this endocytosis defect with CGM and BCM95g, neither of which had any effect on lysosomal storage, which rules this out as a possibility. Based on the similar properties of the SLNL and SLNM particles one possibility is that their lipid content is related to the enhanced lysosomal storage we observed in the Npc1-/- cells treated with these nanoformulations. This is supported by the lack of effect of SLNA, which has a similar formulation but is substantially different in that it contains sodium alginate, which can restrict the diffusion of phospholipids30 and may therefore ameliorate its effects on the Npc1-/- cell. However, we did not observe the presence of any NPC disease storage lipids in these nanoformulations, ruling this out as a possible cause of the elevated storage. One further potential cause of the elevated lipid storage is that the curcumin itself is trapping cholesterol within the NPC disease lysosome. Curcumin has been suggested to be capable of interacting with cholesterol31 and as such, delivery of curcumin into the endocytic system as a nanoparticle could potentially result in the entrapment of cholesterol within these compartments that would reduce any benefit of the elevated intracellular Ca2+. This mechanism could explain why only a small number of the curcumin nanoformulations led to any observable benefit. Another possibility underlying the increased lipid storage is that three of these four nanoformulations, CGM, SLNM and SLNL, have elicited some toxicity in the cells, as observed in Fig. 6. The associated cellular stress would lead to reduced cellular turnover and a greater degree of lipid accumulation in the non-dividing cells, supported by the reduced mitochondrial activity (Fig. 6b). However this lipid accumulation occurs, it is clearly not beneficial for lysosomal lipid levels to be elevated in cells from a lysosomal storage disease. One additional outcome of our findings is that they argue against curcumin working to rescue NPC1 lysosomal storage by exocytosis, as has been suggested32, as the supplement that elevates cytosolic Ca2+ the most (and would therefore elicit the greatest degree of exocytosis), SLNA, only has a minimal effect on reducing lysosomal storage levels. Of the curcumin formulations we have tested it is those least modified with lipids that overall have had the greatest beneficial effect on reducing lysosomal lipid storage in NPC1 disease cells (BCM95s and TRE). These nanoformulations were able to elevate cytosolic Ca2+ without inducing toxicity and were able to reduce lipid storage in Npc1-/- astrocytes as previously reported with pure curcumin2. Our findings are in keeping with the published data from the Npc1-/- mouse model where unmodified curcumin had the greatest effect on survival and function2 whereas lipidated curcumin (SLNL) had no benefit on Npc1-/- mouse function13. Although this is an in vitro study it is important to note that the benefit of curcumin to NPC disease comes from modulation of Ca2+ at the ER and not as an anti-oxidant14. It is therefore the effect of curcumin on the individual cells of the body that needs to be considered and as such our study provides useful insight into the potential effects of curcumin formulations on NPC disease cellular function. For example, the function of the NPC1 intestine and liver is already abnormal22, and it is these tissues that will be primarily affected by short term treatment with these formulations. Some may transcytose, enter the blood stream and be carried to various other organs before releasing their cargo of curcumin and lipids23. What impact this may have on the disease course is unknown and as the only lipidated nanoformulation of curcumin to be tested in NPC disease has been SLNL in the mouse model13 it is clear that more work is needed to determine the safety and efficacy of these nanoformulations on NPC disease.
In conclusion, as NPC1 is a lipid storage disease the use of a lipidated vehicle may not be the best approach due to the possibility that the additional lipid load could alter metabolism or endocytosis and lead to further lipid storage as we have observed. Based on our evidence2, from this report and from others13,14,28,29 it is perhaps the least modified forms of curcumin that appear to have the greatest benefit for NPC1 disease both in vitro and in vivo. Ultimately, the utilisation of curcumin itself may not be ideal for treating NPC1 disease, owing to it’s low bioavailability15, other more bioavailable Ca2+ modulators10,11 may yet prove to be the most effective therapeutic approach for targeting the lysosomal Ca2+ dysfunction in NPC disease.
Acknowledgements
The authors would like to acknowledge the significant support and funding for this project from the Niemann-Pick Research Foundation (NPRF) as well as the support from professional services staff within the School of Biosciences. HRK and RS were supported by a summer student bursary from the NPRF. EM was supported by a PhD studentship award from the NPRF. LJH was supported by Sport Aiding Medical Research for Kids (SPARKS) and a BBSRC DTP PhD studentship alongside an MRC in vivo skills award. EHC was supported by a MRC DTP PhD studentship. HWE was supported by Action Medical Research and the Henry Smith charity. KW and JG were supported by a March of Dimes Basil O’Connor Starter Scholar award given to ELE who was supported by an RCUK Fellowship, a March of Dimes Basil O’Connor Starter Scholar Award and a research grant from the Royal Society.