Psychosine Disrupts Multiple Cellular Functions in Oligodendrocyte Progenitor Cells

We began our studies with an analysis of psychosine (Psy, also referred to as galactosylsphingosine), a lipid that is thought to be of central, and potentially causal, pathogenic importance in KD [56–60]. Psy is one of the most extensively studied of all the lipids known to accumulate in LSDs and is known to exhibit toxicity for multiple cell types in vitro [61–75] and to cause extensive damage when injected intracranially in wild-type (WT) mice [59]. Psy accumulates in tissues of individuals with KD due to galactocerebrosidase ([GALC], EC 3.2.1.46) mutations that cause abnormal processing of lipids that are important components of myelin, the insulating material that enwraps axons in the CNS and peripheral nervous systems (PNS), a primary target of damage in KD. Psy also accumulates in tissues of the naturally occurring, severe murine model of KD, the twitcher mouse [76–80], which also harbors GALC mutations and recapitulates most human pathologies.

As a prelude to analyzing the ability of Psy to alter cellular function, we first determined which CNS cells were most vulnerable to this lipid and found that the most sensitive cells were primary oligodendrocyte (OL)/type-2 astrocyte progenitor cells ([O-2A/OPCs], also referred to as OL precursor cells). These progenitors, which give rise to the myelin-forming OLs of the CNS during development and in response to myelin damage, were killed by pathophysiologically relevant low-micromolar (3 μM) concentrations of Psy [77] that had no effect on hippocampal and cortical neuron survival (see also, e.g., [65, 81]) and were as toxic to OLs (Fig 1A). The vulnerability of primary O-2A/OPCs to Psy was also an order of magnitude greater than seen in immortalized CNS glial progenitor cell lines (e.g., [82]) and in Schwann cells of the PNS [83]. This level of sensitivity falls well within the reported Psy concentrations in the CNS of symptomatic (postnatal day [P]25) and moribund (P40) twitcher mice, which are between 15 μM and 34 μM, respectively [77]. Similar concentrations in the twitcher CNS have been reported by other investigators, ranging from 0.7 μM (P10), 4 μM (P16), and 4.5 μM (P20–P25) to as high as 27–50 μM (P30) [78–80]. Comparable concentrations have been reported in postmortem Krabbe patient CNS tissue, ranging from 2.7 μM to 45 μM in cortical grey and white matter, respectively [84, 85].

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Psy causes a diverse array of cellular and biochemical toxicities in cultured O-2A/OPCs.

(A) Survival of Psy-treated purified rat O-2A/OPCs, OL, cortical, and hippocampal neurons for 5 d relative to untreated controls. (B) Quantification of the number of rat O-2A/OPCs per clone in the presence or absence of Psy over 5 d. (C) Quantification of the relative amount of neutral lipids and phospholipids in rat O-2A/OPCs exposed to positive controls cyclosporin A (10 μM) or propranalol (10 μM), 100 nM bafilomycin A (BafA), or 3.3 μM Psy for 48 h. (D) Quantification of time to half-maximal intensity for rat O-2A/OPCs exposed to Psy (1 μM) or vehicle (DMSO) for 24 h before addition of fluorescent nanobeads. A plot of relative intensity is also shown; lines indicate time to half-maximal intensity. (E) Quantification of relative Cathepsin B and D activities in rat O-2A/OPCs exposed to the indicated drugs or 1 μM Psy for 24 h. (F) Quantification of lysosomal pH in rat O-2A/OPCs exposed to 500 nM BafA, 10 μM chloroquine (CQN), 10 mM ammonium chloride (NH₄Cl), or 1 μM Psy for 24 h or 48 h. Data for all graphs displayed as mean ± standard error of the mean (SEM); *p < 0.05, **p < 0.01, ^†p < 0.001 versus control. Data presented in this figure can be found in S1 Data.

https://doi.org/10.1371/journal.pbio.1002583.g001

Further studies revealed that O-2A/OPCs exposed to still lower (1 μM) levels of Psy exhibited multiple abnormalities of potential relevance to understanding the decreased myelination and apparent failure to repair myelin damage seen in KD. In the absence of cell death, 1 μM Psy suppressed both cell division (Fig 1B) and differentiation of O-2A/OPCs into OLs (S1A Fig). It also disrupted cytoskeletal integrity and caused decreased cell migration (S1B and S1C Fig). Such sensitivity places these cells among those most sensitive to the effects of Psy exposure.

Psy Alters Lysosomal Function in O-2A/OPCs

We next discovered that exposure to 1 μM Psy has the previously unrecognized ability to cause multiple alterations in lysosomal function, indicating that this lipid may provide a direct link between enzymatic mutation and lysosomal dysfunction in KD.

Exposure to Psy caused abnormalities in lipid homeostasis, endolysosomal transport, and cathepsin activity. Exposure to 1μM Psy disrupted lipid homeostasis, causing the intracellular accumulation of both neutral triglycerides and phospholipids (Fig 1C, S1D Fig). Endolysosomal transport was also compromised by exposure to 1 μM Psy, as shown by a decreased rate of endocytic import of fluorescently labeled polystyrene nanobeads (time to half-maximal staining intensity: 4.6 ± 1.0 min for vehicle-treated control versus 22.2 ± 5.7 min for Psy, p < 0.05; Fig 1D, S1E Fig). Psy exposure also increased the activity of resident lysosomal proteases cathepsin D and B, which can cause cellular damage or death upon export to the cytoplasm and the activities of which are known to be elevated in a number of LSDs (Fig 1E) [86–88]. Psy’s ability to disrupt lysosomal function was as great as that seen with bafilomycin A (BafA), which disrupts lysosomal function by antagonizing the lysosomal vacuolar-type H⁺-ATPase [89].

Exposure of O-2A/OPCs to 1 μM Psy significantly increased intralysosomal pH from 4.88 ± 0.04 to 5.62 ± 0.08 after 24 h, an increase maintained for at least 48 h after a single exposure (p < 0.001; Fig 1F). This elevation in lysosomal pH was observed in both fixed (Fig 1F) and live (S1F Fig) O-2A/OPCs. Psy exposure was as potent at increasing lysosomal pH as multiple compounds well known to exert such effects, including BafA, chloroquine, or the weak base NH₄Cl (Fig 1F) [90]. This increase was evident within minutes of exposure to Psy and was comparable to treatment with BafA (S1G Fig, S1–S3 Movies), and the effects on lysosomal pH were sustained over 24–48 h after Psy exposure.

Unbiased Screening Identifies Chemically Diverse Candidate Protective Agents That Prevent Psy-Induced Cellular and Lysosomal Dysfunctions

To identify potential means of preventing Psy-induced toxicities that might be suitable for eventual clinical utilization, we conducted unbiased analysis of multiple concentrations of 1,040 mostly United States Food and Drug Administration (FDA)-approved small molecules [91] and a custom panel of 12 growth factors with known neuroprotective activity. We examined prevention of Psy-induced suppression of O-2A/OPC division in these analyses (Fig 2A).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Unbiased screening identifies chemically diverse candidate protective agents that reduce Psy toxicities.

(A) Schematic depicting the work flow for unbiased screening with Celigo adherent cytometer (Nexcelom). (B) Representative plot of relative cell number over 5 d for rat O-2A/OPCs exposed to 1 μM Psy or vehicle (DMSO). Quantification of the relative proliferation rate for all 1,040 drugs at three concentrations over 5 d. Blue: “hit” drugs selected for further validation. A bar graph quantifying the relative proliferation rate of all vehicle and Psy controls is included. (C) Quantification of the relative proliferation rate of rat O-2A/OPCs exposed to 1.5 μM Psy, with and without verified pro-division “hits,” over 5 d. (D) Quantification as in (C) for cells exposed to Psy 48 h before exposure to the indicated drugs. (E) Summary of clinical metadata for lead “hits” as % of total. (* worldwide approval, including FDA. ** reported blood–brain barrier permeability; not all drugs have been examined/reported.) (F) Quantification of cell survival in cells exposed to 3.3 μM Psy for 5 d, with and without the indicated drugs, for 5 d. (G, H) Quantification of the (G) relative survival and (H) number of rat O-2A/OPCs per clone exposed to of Psy (1 μM in H), with and without 100 ng/mL insulin-like growth factor (IGF-1), for 5 d. (I–K) Quantification of (I) neutral lipid and phospholipid accumulation, (J) cathepsin B and D activities in rat O-2A/OPCs exposed to the indicated drugs, with and without 1 μM Psy, and (K) time to half-maximal intensity for endocytosis of fluorescent nanobeads for (I) 2 d or (J, K) 24 h. (L) Quantification of lysosomal pH for rat O-2A/OPCs exposed to 1 μM Psy, with and without the indicated “hits” (blue) or “non-hits” (gray), for 24 h. NT-3: 10 ng/mL; 1E04: 5 μM; 5F05, 9A06, 9H10: 1 μM. ^ap < 0.05, ^bp < 0.01 versus Psy-only. Data for all graphs displayed as mean ± SEM; ns = not significant; *p < 0.05, **p < 0.01, ^†p < 0.001 versus untreated; ^ap < 0.05, ^bp < 0.01, ^cp < 0.001 versus Psy-only treatment. See also S2 Fig, S1 and S2 Tables for drug names and concentrations. Data presented in this figure can be found in S1 Data.

https://doi.org/10.1371/journal.pbio.1002583.g002

We found 16 structurally and functionally diverse compounds (S2A Fig, S1 and S2 Tables), in addition to 4 growth factors, that had the unexpected properties of rescuing cell division (Fig 2B and 2C, S2B Fig). Eight of the 9 most protective agents were effective at rescuing cell division even when their administration was delayed 48 h after Psy exposure (Fig 2D). Importantly, all small molecules were optimally protective at physiologically achievable concentrations (i.e., nanomolar to low micromolar), and most are approved for use in humans (94%) and are blood–brain barrier permeable (>80%) (Fig 2E).

Five agents (chlorotrianisene [1G05], NKH-477 [(9C06), also known as colforsin], clofoctol [8D08], tulobuterol [9E07], and insulin-like growth factor [IGF-1]) revealed by our screens not only significantly rescued cell division but were also able to reduce cell death caused by exposure to higher concentrations of Psy (Fig 2F). With the exception of IGF-I, none of our compounds of interest were previously identified as being able to protect against toxicity of Psy (or of other lipids accumulating in LSDs). Even in the case of IGF-I, previous studies reported that supraphysiological (>10 μg/mL) concentrations decreased Psy-induced apoptosis in OLs [92] and in an O-2A/OPC cell line [82]. In our studies, by contrast, 100 ng/mL IGF-1 shifted Psy’s cytotoxicity curve by an order of magnitude and significantly deceased Psy-dependent suppression of self-renewal (Fig 2G and 2H).

Further examination of three of the most protective agents—clofoctol, NKH-477, and IGF-I—demonstrated that these compounds also prevented Psy-induced alterations in lysosomal function. All three suppressed Psy-induced increases in lipid accumulation and cathepsin activity and restored normal endocytosis (Fig 2I–2K). Moreover, all three compounds significantly decreased Psy-dependent increases in pH (Fig 2L). In contrast, five randomly selected molecules that did not significantly reduce any Psy toxicities in our screens (caffeine [1E04], acetarsol [5F05], mepartricin [9A06], avobenzone [9H10], and neurotrophin-3; S1 Table) had no effect on Psy-induced increases in lysosomal pH.

Protective Agents Converge on a Limited Number of Common Necessary Pathways for Their Activity

As the protective compounds we discovered are structurally and functionally diverse, we next attempted to define regulatory pathways on which these agents might converge to confer their protective activity. To do this, we focused on multiple signaling pathways previously described to be antagonized by exposure of cells to Psy (e.g., mitogen-activated protein kinase [MAPK], phosphoinositide-3-kinase [PI3k]/Akt, protein kinase C [PKC], cyclic Amp (cAMP)-dependent signaling [51, 74, 82, 93–95]), as well as other proteins that have been implicated in mediating stress responses in O-2A/OPCs, including Jun N-terminal kinase (Jnk) [74], mammalian target of rapamycin (mTOR) [96–98], and estrogen receptor [99–101]. We next generated a secondary screen consisting of pharmacological inhibitors to components of these various signaling pathways. In these experiments, O-2A/OPCs were exposed to Psy; a combination of Psy and a protective agent; or a combination of Psy, a protective agent, and one of 15 pharmacological inhibitors targeting important signaling pathways and proteins (S3 Table). This allowed for the generation of a compound-specific “fingerprint of protection” that revealed putative signaling pathways used by the candidate compound to overcome Psy-induced suppression of division (e.g., Fig 3A and 3B).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. Protective agents converge on a limited number of common necessary pathways for their activity.

(A, B) Representative “fingerprint of protection” for candidate drug 1G05. Quantification of the relative proliferation rate of rat O-2A/OPCs exposed to 1.5 μM Psy, with and without 100 nM 1G05, and with and without pharmacological inhibitors targeting the indicated protein, for 5 d. To account for differences in their ability to reduce Psy-induced suppression of division between candidate agents, relative changes in cell division were normalized as in (B). (C) Unsupervised hierarchical clustering of “fingerprints” for the indicated drugs. Data for all graphs displayed as mean ± SEM. See also S1 and S2 Tables for drug names and concentrations, and S3 Table for details on the “fingerprinting” screen. Data presented in this figure can be found in S1 Data.

https://doi.org/10.1371/journal.pbio.1002583.g003

We generated fingerprints for 12 of the most efficacious compounds in reducing Psy-induced suppression of division across 5 d using this approach. The results were then hierarchically clustered to identify similarities and dissimilarities between individual compound fingerprints and between the signaling pathways implicated in protection (Fig 3C). Despite structural and functional diversity among candidate protective agents, there was striking similarity in the signaling pathways needed for protection.

The activity of the diverse protective agents was antagonized by pharmacological inhibition of the Ras/rapidly accelerated fibrosarcoma gene (Raf)/MAPK pathway, Akt, estrogen receptor, protein kinase A (PKA), and geranylgeranyl transferase ([GGT], which is needed for activation of small GTPases that are involved in cell division and migration). Despite their structural diversity, there was a surprisingly high degree of correlation between groups of small molecules; the cluster of structurally and functionally unrelated drugs 2G08 (ethopropazine, an antiparkinsonian drug), 2F11 (estradiol valerate, a synthetic estrogen), and 8D08 (clofoctol, an antibiotic), for example, showed the highest degree of similarity (correl. = 0.97) (S3 Fig).

Protection Against Psy-Induced Lysosomal and Cellular Dysfunctions Can Be Provided by Promotion of Lysosomal Re-acidification

Lysosomal ion homeostasis, maintained through the activity of several channels and transporters, is critical to the normal function of lysosomes. For example, H⁺ import is necessary for the maintenance of an acidic pH [102] and is achieved through the activity of the V-ATPase, Ca²⁺ is important for vesicle trafficking [103] and fusion [104], Na⁺ and K⁺ are required for the regulation of membrane potential [105, 106], and Cl⁻ serves as a counterion to regulate lysosomal membrane potential and to facilitate the acidification of the lysosome lumen [107–109]. Although any of these may be potential therapeutic targets, we focused our attention on identifying those channels or transporters regulated by signaling pathways uncovered through our fingerprinting analysis.

Of the several pathways that are required for activity of our protective agents, the one for which there is a clearly established linkage to at least one aspect of lysosomal function is the requirement for PKA activity. Previously, it has been shown that cAMP can promote lysosomal re-acidification [110], as can PKA, which is activated by cAMP [111]. In addition, we found that increases in cAMP not only normalized lysosomal pH but also prevented Psy-induced decreases in O-2A/OPC division (Fig 4A and 4B), raising the theoretical possibility that intervention at this point would provide additional benefits beyond that of pH restoration. We therefore focused attention on the role of lysosomal re-acidification as a potential therapeutic target.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 4. Promotion of lysosomal re-acidification is critical in protecting from multiple aspects of Psy toxicity.

(A–D) Quantification of the lysosomal pH and proliferation rate in rat O-2A/OPCs exposed to (A, C) 1 μM Psy for 24 h, (B, D) 1.5 μM for 5 d, with and without (A, B) 1 mM cAMP (or 10 μM forskolin), or (C, D) 333 nM RP-107 and 1 μM cystic fibrosis transmembrane conductance regulator (CFTR) inhibitor 172 (CFTRi-172). (E) Quantification of neutral lipid and phospholipid accumulation in rat O-2A/OPCs exposed to the indicated drugs, with and without 1 μM Psy, for 2 d. (F, G) Quantification of the lysosomal pH and proliferation rate in rat O-2A/OPCs exposed to 1 μM Psy for (F) 24 h or (G) 5 d, with and without 100 nM KT-5720 or 3.3 μM H89 (cAMP: 1 mM). Data for all graphs displayed as mean ± SEM; *p < 0.05, **p < 0.01, ^†p < 0.001 versus untreated; ^ap < 0.05, ^bp < 0.01, ^cp < 0.001 versus Psy-only treatment. See also S1 and S2 Tables for drug names and concentrations. Data presented in this figure can be found in S1 Data.

https://doi.org/10.1371/journal.pbio.1002583.g004

The most attractive explanation for how cAMP/PKA activity could restore lysosomal pH would be through activation of the cystic fibrosis transmembrane conductance regulator (CFTR), a PKA-activated transmembrane chloride channel that promotes lysosomal re-acidification [112]. Unlike the CLC-7 Cl⁻/H⁺ antiporter, another chloride channel that is localized to the lysosomal membrane and thought to play a role in the basal maintenance of lysosomal pH [113], the CFTR channel appears only to be critical for re-acidification. Moreover, although the CFTR can be activated by PKA and cAMP, there is no evidence for such activation of CLC-7. In addition, specific agonists and inhibitors exist for the CFTR, enabling a direct test of whether promoting re-acidification can prevent Psy-induced toxicity.

As we predicted, treatment of cells with the cAMP-independent CFTR agonist RP-107 [114] restored lysosomal pH in cells exposed to Psy. Although control of lysosomal pH and/or lysosomal re-acidification has not been thought to have any upstream role in the multiple cellular dysfunctions caused by Psy exposure, we nonetheless found that RP-107 protected against Psy-induced suppression of division, as well as elevated storage of both neutral lipids and phospholipids (Fig 4C–4E). To test the hypothesis that these benefits were not due to off-target effects of RP-107, we also co-exposed cells to CFTR-inhibitor 172 (CFTRi-172) [115], which attenuated the protective effects of RP-107 treatment (Fig 4C–4E).

The effects of RP-107 were CFTR dependent, as determined by knockdown of CFTR in O-2A/OPCs using small interfering RNA (siRNA) pools targeting rat CFTR, as well as a pool of nontargeting (NT) siRNAs as a control for transfection; the reduction in CFTR protein levels was confirmed by western blot analysis. Knockdown of CFTR did not significantly affect lysosomal pH when compared to cells exposed to NT controls (4.96 ± 0.13 versus 4.81 ± 0.11, respectively). Moreover, in the presence of Psy, in both NT and CFTR siRNA pools, there was a significant increase in lysosomal pH (5.47 ± 0.08 versus 5.61 ± 0.09, respectively), with no significant difference between these two treatment groups. However, when we tested the effect of RP-107, a specific CFTR agonist, we found that lysosomal pH was significantly reduced in cells exposed to NT siRNA but that CFTR knockdown attenuated RP-107’s protective effect (5.18 ± 0.09 versus 5.62 ± 0.03, p < 0.01; S4A Fig). Thus, as with our pharmacological experiments, genetic loss of CFTR does not appear to significantly affect basal lysosomal pH in untreated cells. However, the protective capacity of RP-107 is CFTR dependent. These results are consistent with the original studies demonstrating the role of the CFTR in control of lysosomal re-acidification [112].

Moreover, we found that the most effective protective agents identified in our studies did not themselves reduce the basal acidic pH of lysosomes in the absence of Psy (S4B Fig) but instead seemed to work to promote re-acidification. Indeed, their ability to normalize lysosomal pH, as well as rescue cell division, in cells exposed to Psy was blocked by co-exposure to inhibitors of PKA (Fig 4F and 4G). As these protective agents are able to rescue cells even when applied 48 h after Psy exposure (Fig 2D), it appears that their protective activity is not mediated simply by blocking lysosomal alkalization.

Psy’s Free Amine Group Is Critical for Its Toxicity

The observations that multiple Psy-induced lysosomal and cellular dysfunctions can be prevented by lysosomal re-acidification with RP-107 (Fig 4C–4E) and that Psy exposure causes rapid increases in lysosomal pH (Fig 1F, S1D Fig), raise complementary questions about how Psy causes such changes. One possibility is that Psy disrupts the function of particular proteins involved in lysosomal re-acidification, but another possibility is that structural features of Psy itself are directly relevant to understanding effects on lysosomal pH.

Although multiple studies have attempted to understand the molecular mechanisms underlying Psy’s toxicity [61–75], we noted that Psy has unusual physicochemical features that might be of relevance to understanding its effects on lysosomes. Psy is unusual as a cationic, weakly basic lipid, carrying a net positive charge at physiological pH. With a pKa value of 7.18 [116], Psy is predicted to be 99.9% protonated in the acidic pH of the lysosome. If this aspect of Psy’s structure is important in altering lysosomal and cellular function, then the protonatable free amine group on Psy should be critical in mediating the changes in lysosomal pH that we observed. We therefore tested whether removing this free amine group altered effects on lysosomal pH and on other outcomes of Psy exposure.

We found that the free amine group on Psy is critical in its ability not only to disrupt lysosomal pH but also to cause other toxic effects. We compared Psy toxicity to that of N-acetyl-Psy (N-AcPsy), a structural derivative containing an amide-linked acetyl group, rendering it no longer protonatable (Fig 5A). Unlike Psy, N-AcPsy did not induce cell death or alter O-2A/OPC self-renewal at similar concentrations (Fig 5B and 5C). Moreover, N-AcPsy did not elevate neutral lipid and phospholipid storage, increase endocytic transport time, increase cathepsin activity, or elevate lysosomal pH (Fig 5D–5G). Thus, the positively charged free amine group present on Psy was critical to increasing lysosomal pH and also to the subsequent lysosomal and cellular impairments observed after exposure in O-2A/OPCs.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 5. Psy’s free amine group is critical for its toxicity.

(A) Structures of Psy and N-AcPsy. (B) Quantification of rat O-2A/OPC survival in response to Psy or N-AcPsy. (C–G) Quantification of rat O-2A/OPC self-renewal, lipid accumulation, endocytosis, cathepsin activity, and lysosomal pH in cells exposed to 1 μM Psy or N-AcPsy for (C) 5 d, (D) 2 d, or (E–G) 24 h. Data for all graphs displayed as mean ± SEM; ns = not significant; *p < 0.05, ^†p < 0.001 versus control. Data presented in this figure can be found in S1 Data.

https://doi.org/10.1371/journal.pbio.1002583.g005

Structurally Related Lysosphingolipids from Multiple LSDs Alter Lysosomal Function

To further test the hypothesis that the presence of free amine group on a cationic lipid is critical to lipid-induced toxicities, and that such lipids provide a direct link between enzymatic mutation and lysosomal disruption, we examined a series of lipids known to accumulate in other LSDs. Other lipids of potential interest include lyso-sulfatide (lyso-SF) (which accumulates in MLD [49]), glucosylsphingosine (GlcSph) and glucosylceramide (GlcCer) (which accumulate in Gaucher disease [47]), and lactosylsphingosine (LacSph) and lactosylceramide (LacCer), which accumulate in several LSDs (Fig 6A) [78, 117, 118]. Some of these lipids appear to have been only rarely studied for their effects on cell function in vitro (lyso-SF, GluSph, LacSph, LacCer) [119]. In the case of Gaucher disease, the majority of previous in vitro studies appears to have focused on GlcCer, and studies on both GlcCer and GlcSph often have required lipid concentrations severalfold greater than those at which Psy’s effects were observed (e.g., [22, 50, 51, 53, 55, 120–122]).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 6. Lysosphingolipids accumulating in other LSDs suppress critical O-2A/OPC behaviors and lysosomal function.

(A) Chemical structures of the indicate lipids. Gal: galactose; Glc: glucose; R: variable hydrocarbon chain. (B–G) Quantification of (B) the relative survival, (C) the relative number of O-2A/OPCs per clone, (D) lipid accumulation, (E) endocytic import time, (F) cathepsin activity, and (G) lysosomal pH in rat O-2A/OPCs exposed to the indicated lipids for (B and C) 5 d, (D, E, G) 24 h, or (F) 2 d. Data for all graphs displayed as mean ± SEM; ns = not significant; *p < 0.05, **p < 0.01, ^†p < 0.001 versus untreated, unless otherwise indicated. S4 Table for lipid concentrations used in (C–G). Data presented in this figure can be found in S1 Data.

https://doi.org/10.1371/journal.pbio.1002583.g006

In order to eliminate differences in cell types as potential contributors to different outcomes, we examined the survival and self-renewal of O-2A/OPCs exposed to lyso-SF, GlcSph, GlcCer, LacSph, and LacCer. Use of these cells also provided a test of the hypothesis that the structure of a lipid is of primary importance in determining toxicity. We also examined the effects of N-acetyl-sulfatide (N-AcSF) as a direct comparison with N-AcPsy. We found that sphingosine-derived lipids that accumulate in different LSDs and that contain a free amine group (and thus are structurally similar to Psy) caused significant cell death and suppression of self-renewal (Fig 6B and 6C) at similarly low lipid concentrations as we observed with Psy. In contrast, exposure to their ceramide-based counterparts GlcCer and LacCer, or to N-AcSF, did not cause cellular toxicities at comparable or 10-fold higher concentrations (Fig 6B and 6C).

We also found that lysosphingolipids accumulating in other LSDs [22, 31, 47–55, 123–127] had similar effects as Psy on lysosomal function. Exposure to sublethal concentrations of GluSph, lyso-SF, or LacSph caused increases in neutral lipid and phospholipid accumulation, endocytic transport time, cathepsin activity, and lysosomal pH. In contrast, exposure to their non-lyso counterparts did not have such effects (Fig 6D–6G).

If the hypotheses are correct that other toxic lysosphingolipids that accumulate in LSDs work through similar mechanisms as Psy, and that such mechanisms are relevant to understanding the efficacy of our protective agents, then our protective agents also should rescue cells from the toxic effects of lipids from other LSDs. If correct, such findings would provide both the first structural predictors of toxicity and the first example of protective agents of potential relevance in different LSDs.

We found that our candidate protective agents also reduced the toxic effects of GlcSph, lyso-SF, and LacSph (Fig 7A). Three of our most effective agents—IGF-1, clofoctol (8D08), and NKH-477 (9C06)—prevented lipid-induced suppression of division and also attenuated increases in neutral lipid and phospholipid accumulation, endocytic import time, cathepsin activities, and lysosomal pH in rat O-2A/OPCs exposed to sublethal concentrations of GlcSph, lyso-SF, or LacSph (Fig 7B–7G, S5A Fig). These agents also rescued cell division in cells exposed to GlcSph or Lyso-SF for 48 h before addition of protective agents (S5B Fig).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 7. Candidate protective agents reduce multiple lysosphingolipid-induced lysosomal and cellular toxicities.

(A) Venn diagram summarizing the number of protective drugs, including IGF-1, that reduce suppression of division in rat O-2A/OPCs exposed to the indicated lyso-lipids for 5 d. (B–E) Quantification of (B) lipid accumulation, (C) endocytic import time, (D) cathepsin activity, and (E) lysosomal pH in rat O-2A/OPCs exposed to the indicated lyso-lipids, with and without 100 ng/mL IGF-1, 100 nM 8D08, or 333 nM 9C06, for (B) 2 d or (C–E) 24 h. Data for all graphs displayed as mean ± SEM; ns = not significant; *p < 0.05, **p < 0.01, ^†p < 0.001 versus untreated; ^ap < 0.05, ^bp < 0.01 versus Psy-only treatment, unless otherwise indicated. See also S4 Fig and S1 and S2 Tables for drug names and concentrations and S4 Table for lipid concentrations. Data presented in this figure can be found in S1 Data.

https://doi.org/10.1371/journal.pbio.1002583.g007

Lysosphingolipids Disrupt Human OL Progenitor Cell Behaviors, and Protective Compounds Rescue Human Cells

We next examined the question of whether the principles revealed in our studies on cells derived from the CNS were applicable to human cells. In these experiments, we used an anti-CD140a (PDGFRα) antibody to enrich for a population of human O-2A/OPCs from the corpus callosal field of mid-gestation fetal tissue (S6 Fig) [128] and exposed cells to Psy and potential protective agents as for rat-derived cells.

Exposure to lysosphingolipids caused death in human cells at concentrations comparable to those used in rat progenitor cells (S4 Table), as well as suppression of cell division and elevation of lysosomal pH at sublethal concentrations, whereas their non-lyso counterparts did not cause similar toxicities (Fig 8A–8C). Notably, cell division and normalization of lysosomal pH were restored in cells exposed to Psy with clofoctol, NKH-477, and IGF-I, as we observed for rat O-2A/OPCs (Fig 8D and 8E).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 8. Lysosphingolipids disrupt human O-2A/OPC behaviors.

(A–C) Quantification of (A) the relative survival, (B) the relative proliferation rate, and (C) lysosomal pH of human O-2A/OPCs exposed to the indicated lipids for (A, B) 5 d or (C) 24 h. BafA: 100 nM. (D, E) Quantification of the relative proliferation rate and lysosomal pH in human O-2A/OPCs exposed to 1 μM Psy for (D) 5 d or (E) for 24 h, with and without the indicated drugs. ^ap < 0.05 versus Psy-only. Data for all graphs displayed as mean ± SEM; ns = not significant; *p < 0.05, **p < 0.01, ^†p < 0.001 versus untreated control; ^ap < 0.05, ^bp < 0.01 versus Psy-only treatment. See also S5 Fig and S1 and S2 Tables for drug names and concentrations, and S4 Table for lipid concentrations. Data presented in this figure can be found in S1 Data.

https://doi.org/10.1371/journal.pbio.1002583.g008

Alterations in O-2A/OPC Biology in Twitcher Mice Are Like Those Induced By Psy Exposure In Vitro

In the final section of our studies, we asked whether discoveries made on WT cells exposed exogenously to Psy in vitro revealed principles applicable to cells with an LSD-relevant mutation, both in respect to cellular pathologies and to rescue of lysosomal function. These studies were carried out using twitcher mice, a naturally occurring model of KD that recapitulates most human symptoms. Multiple studies have demonstrated that this mouse is a reliable model of KD in respect to enzymatic dysfunction and tissue pathology [77–80, 129, 130] and is also one of the most useful models for studying LSDs in general. In particular, twitcher mice progress from a lack of apparent pathology to severe disease over a relatively rapid time course, with function appearing to be normal at birth, followed by disease symptoms manifesting about 20 d after birth and with death ensuing at about 42 d after birth. This time course allows pathology and the effects of treatment to be studied at different stages of disease progression.

In our studies on twitcher mice, we first determined that changes in O-2A/OPC function were like those induced by exposure to low doses of Psy in vitro. We found significant reductions in both myelin content and OL cell number (OLs; Olig2⁺/GST⁺) in the corpus callosum—the major myelinated tract of the CNS—at P40 when compared to age-matched WT littermates (Fig 9A and 9B), consistent with previous analyses of human and twitcher tissue [76, 131]. We also observed a significant reduction in the percentage of dividing (Ki67⁺) O-2A/OPCs (54.0% ± 1.9% of WT, p < 0.01; Fig 9C) at this late time point, during which time O-2A/OPCs should be undergoing rapid expansion through cell division to replace damaged OLs and myelin.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 9. NKH-477, a protective compound identified in vitro, protects against multiple toxicities in treated twitcher mice.

(A–C) Quantification of (A) fluoromyelin intensity, (B) number of GSTpi+/Olig2+ OLs, and (C) the relative number of dividing Ki67+/Olig2+ O-2A/OPCs in the corpus callosa of P40 twitcher mice and age-matched WT littermates (n = 3 from different litters). (D) Analysis of clonal composition of P15 twi O-2A/OPCs and WT littermates across 5d. (E) Quantification of the relative number of dividing Ki67+/Olig2+ O-2A/OPCs in the corpus callosa of P15 twitcher mice and age-matched WT littermates (n = 3 from different litters). (F) Quantification of lysosomal pH of O-2A/OPCs acutely isolated from P17 twitcher and WT mice. (G) Overview of treatment paradigm and clinical course for twitcher mice. (H) Quantification of lysosomal pH of O-2A/OPCs isolated from P35 treated mice. (I) Quantification of the number of dividing callosal O-2A/OPCs in P35 treated mice. (J–L) Representative confocal images of fluoromyelin-stained corpus callosa of the indicated treatment groups at P40, in addition to quantification of staining intensity and the number of OLs. (M) Kaplan–Meyer survival curve for treated and untreated twi mice. Median survival of twi mice, with dotted lines indicating reported median survival of single-therapy treatments [15]. (N–Q) Quantification of travel speed, stance time, beam traverse time, and relative weights for P25 saline-treated WT and twi, as well as NKH-477–treated twi, mice. (R) Quantification of brain Psy levels at P35. Data for all graphs displayed as mean ± SEM; ns = not significant; *p < 0.05, ^†p < 0.001 versus WT; ^bp < 0.01 versus vehicle-treated twi. Data presented in this figure can be found in S1 Data.

https://doi.org/10.1371/journal.pbio.1002583.g009

We additionally found that O-2A/OPC function was compromised in presymptomatic twitcher mice in ways similar to those induced by Psy exposure. We isolated O-2A/OPCs from presymptomatic P15 twitcher mice and examined their self-renewal capacity in vitro. These cells showed impaired self-renewal in comparison with cells of age-matched WT cells when maintained in vitro for 5 d (Fig 9D). Such findings were mirrored by significant reductions in the pool of dividing O-2A/OPCs in vivo at P15 (Fig 9E).

To determine whether cells harboring a mutant GALC gene exhibit changes in lysosomal pH, we examined the endolysosomal pH of corpus callosal O-2A/OPCs acutely isolated from presymptomatic twitcher mice (P17). We found that the lysosomal pH was significantly less acidic than that of cells isolated from age-matched WT littermates (Fig 9F), similar to what was observed in vitro with exogenous Psy treatment (Fig 1F). Thus, O-2A/OPCs isolated at developmental time points in which symptoms are not obvious (prior to P18–20) show altered lysosomal pH and alterations in critical cellular behaviors like those induced by exposing WT cells to Psy in vitro.

NKH-477 Protects against Multiple Toxicities in Twitcher Mice

We next investigated whether the analytical approach employed in our in vitro studies could identify compounds able to provide clinically relevant benefits in vivo. We focused our studies on NKH-477 (9CO6), a water-soluble derivative of forskolin that is approved for treatment of acute heart failure in Japan [132], as this agent is known to be CNS penetrant and elevates cAMP levels (through direct activation of adenylyl cyclase) in brains of rats after systemic administration [133]. Moreover, unlike the other identified protective agents, the linkage of NKH-477 to PKA regulation (and thus to lysosomal re-acidification) is both defined and mediated through widely expressed proteins, consequentially not requiring cells to express specialized drug-targeted receptors in order to be responsive.

We initiated treatment at P10, a time when CNS concentrations of Psy are already approaching the range at which we see effects on O-2A/OPCs [78–80], using once-daily intraperitoneal (IP) injections (1 mg/kg; Fig 9G). This is a point in time when disruptions in neuronal function can already be observed in twitcher mice [134], raising the possibility of initiating treatment only after subtle clinical changes are first observable. This delayed initiation of treatment is in marked contrast with the well-studied need to initiate the application of bone marrow transplantation and/or gene therapy in the first few days after birth in order to obtain benefit [1, 10, 16, 40, 130].

The primary endpoints of interest in our in vivo studies were whether we could rescue lysosomal and cellular function in O-2A/OPCs and whether once-daily treatment with NKH-477 is sufficient to provide benefit on both parameters. O-2A/OPCs were isolated at P35 to examine the effects of NKH-477 treatment on lysosomal pH, and we found a normalization of pH in cells isolated from treated twitcher mice when compared to vehicle-treated mice (Fig 9H). NKH-477–treated twitcher mice also showed an increase in the numbers of dividing O-2A/OPCs at P35 to near-normal levels, as well as increases in myelin content and increased OL cell numbers at moribund ages, when compared to vehicle-treated twitcher mice, again to levels not significantly different from WT mice (Fig 9I–9L).

Remarkably, we also found that NKH-477 treatment provided significant lifespan extension that was comparable to published single-therapy treatments aimed at restoring GALC activity, including bone marrow transplantation (the current standard of care in patients) or viral-mediated gene therapy (Fig 9M) [1, 2, 10, 13, 15, 16]. Moreover, twitcher mice that received daily injections of NKH-477 also showed significantly improved locomotor and gait function (Fig 9N–9P, S7 Fig) and significantly improved weight gain throughout their lifespan when compared to vehicle-treated twitcher littermates (Fig 9Q). These benefits were observed despite the fact that we were not correcting the genetic defect; indeed, we did not find that NKH-477 treatment reduced the overall tissue burden of Psy in the CNS (Fig 9R).

Ethics Statement

The University of Rochester RSRB has reviewed this study and determined that based on federal (45 CFR 46.102) and University criteria the study does not qualify as human subjects research and has waived the need for consent (RSRB#00024759). All animal procedures were performed under guidelines of the National Institutes of Health and approved by the Institutional Animal Care and Utilization Committee (IACUC) of the University of Rochester Medical Center, Rochester, NY (UCAR#2001–140).

Lipids

Lipids used in this study were purchased from Matreya and were of highest purity. All lipids were resuspended in anhydrous dimethyl sulfoxide (DMSO) to 10 mM, stored at -20°C or -80°C, and resuspended in media before use. Comparable results for OPC division and survival were obtained with Psy purchased from Sigma-Aldrich and Santa Cruz Biotechnology.

O-2A/OPC Isolation, Purification, and Culture

Corpus callosa of P7 Sprague-Dawley rats (Charles River) were micro-dissected, finely minced with a sterile blade, and digested for 20 min in 2.2 mg/mL collagenase (Worthington #4189), 20 Kunitz/mL DNase (Sigma D4263) in HBSS (Gibco #114170–161) supplemented with Sato medium (see below). Collagenase-containing medium was then replaced with 20 Kunits/mL DNase and papain solution (1:40, activated per manufacturer’s directions; Worthington #LS003127) in HBSS/Sato for 20 min. Tissue was then sequentially triturated with 21-, 25-, and 26-guage needles in 35 Kunits/mL DNase in DMEM:F12 complete media (see below) before dissociated cells were plated on tissue culture plastic for 10 min (37°C, 7% CO₂). Nonadherent cells were pelleted by centrifugation (5 min, 500 xg), resuspended in degassed HBSS/Sato supplemented with 1% BSA Fraction V and anti-A2B5 MACS beads (1:50; Miltenyi Biotec #130-093-388), and incubated on ice for 20 min. Cells were pelleted and sorted with MACS columns as per manufacturer’s directions (Miltenyi).

A2B5⁺ cells were plated on tissue culture plastic coated with poly-L-lysine (1 μg/cm² for 20 min; Sigma #P1274) in DMEM:F12 (Gibco #11330–057) supplemented with 10 μg/mL insulin (Sigma #I5500), 100 μg/mL holotransferrin (Sigma #T2252), Sato media (final concentration: 0.03% BSA Fraction V [Sigma #A7979-50ML], 10 μM putrescine [Sigma #P7505], 200 nM progesterone [Sigma #P0130], 235 nM sodium selenite [Sigma #S1382]), 50 μg/mL gentamycin (Gibco #15750–060), 10 ng/mL PDGF-AA (R&D #221-AA), and 5 ng/mL basic FGF (Miltenyi #130–093) and maintained at 37°C (7% CO₂). Cells were passaged with 0.05% trypsin-EDTA (Gibco #2300), neutralized with 80 Kunitz/mL soybean trypsin inhibitor (Sigma #T9003), and replated in DMEM:F12 complete media supplemented with 10 ng/mL PDGF-AA (and without basic FGF). Purified O-2A/OPCs were passaged no more than once and were maintained in culture for at most 7–9 d in vitro for all experiments. To generate OLs, purified O-2A/OPCs were maintained in DMEM:F12 containing Sato components, transferrin, insulin, 100 pg/mL PDGF-AA, and 45 nM T3/T4 (Sigma #T6397/#0397) for 5 d before initiation of experiments. Human cells were isolated from corpus callosal fields of fetal week 18–21 de-identified tissue, purified with anti-CD140a-coupled magnetic beads (1:100; BD Biosciences #558774; Miltenyi), and maintained as above.

Cortical and Hippocampal Neuron Isolation and Culture

Embryonic neurons were isolated from E18 Sprague-Dawley rats (Charles River) and maintained as previously described [213]. Briefly, isolated tissue was digested in papain solution (1:50, activated per manufacturer’s directions; Worthington #LS003127) in HBSS/Sato for 20 min at 37°C (7% CO₂). Pelleted tissue was then triturated with a pulled glass Pasteur pipet in 80 Kunitz/mL DNase (HBSS; Sigma #D4263), and dissociated cells were pelleted through a 0.5 M sucrose cushion (10 min, 500 xg). Immature neurons were plated on poly-L-lysine-coated tissue culture plastic (1 μg/cm² for 20 min; Sigma #P1274) in NeuroBasal media (Gibco #21103–049) with 50 μg/mL gentamycin (Gibco #15750–060), 2 mM L-glutamine (Gibco #25030–081), and 1 X B27 serum-free supplement (Gibco #17504044). Neurons were allowed to mature for 7 d at 37°C (7% CO₂) before initiation of experiments, with a 50% media change every third day. Analysis of cell survival was determined with calcein-AM/propidium iodide, as described below.

Clonal Analysis

Purified (A2B5⁺) O-2A/OPCs were isolated from P7 rat corpus callosa as above and plated at a density of 25 cells/cm² in poly-L-lysine-coated 24-well plates in DMEM:F12 complete media supplemented with 10 ng/mL PDGF-AA immediately after purification. Experiments were initiated after 24 h. Cells were fixed after 5 d in 4% paraformaldehyde, stained with antibodies against A2B5 (1:4; in-house hybridoma, ATCC) and GalC (1:4; in-house hybridoma, ATCC), and counterstained with DAPI (1 μg/mL; Invitrogen #D1306), and the size and composition of each clone was then scored. Spontaneous generation of GalC⁺ OLs (i.e., in the absence of differentiation conditions) or Type 2 astrocytes was not detected in this paradigm, and so only the number of progenitor cells (A2B5⁺GalC⁻) per clone is reported.

Cell Migration

Cell migration with agarose drops was performed as previously reported [214, 215]. Briefly, purified (A2B5⁺) O-2A/OPCs were isolated from P7 rat corpus callosa as above, resuspended in 0.3% low-melt agarose (at 37°C; Sigma #A0701), and diluted in DMEM:F12 complete media supplemented with 10 ng/mL PDGF-AA at a density of 4 x 10⁴ cells/μL, and 1.5 μL of the cell-agarose mixture was plated in the center of a poly-L-lysine-coated 24-well plate. The agarose was allowed to gel at 4°C for 10 min before DMEM:F12 complete media supplemented with 10 ng/mL PDGF-AA, with and without Psy. (PDGF-AA was omitted in some wells as controls for migration, as O-2A/OPC motility is stimulated in vitro by PDGF). Half of the media, with and without Psy, was replaced daily for 3 d, after which point the cells were loaded with calcein-AM (200 nM) and imaged. The distance migrated from the agarose drop to the leading edge was quantified as reported [215].

Immunocytochemistry

Cells were fixed with 4% paraformaldehyde for 20 min before permeabilization for 10 min with 0.5% Triton X-100 (Sigma #X100) in blocking media (Earl’s Balanced Salt Solution [Gibco] with 5% calf serum and 1% BSA Fraction V). Permeabilized cells were then blocked for 1 h at 25°C before overnight incubation with primary antibodies (A2B5 hybridoma [IgM, 1:4], GalC hybridoma [IgG3, 1:10], Olig2 [1:500; Millipore #MABN50], Ki67 [1:1000, BD Pharmingen #550609], GFAP [1:2000; DAKO #Z0334], Tuj1 [1:2000; Abcam #14545]) diluted in blocking media at 4°C. After washing, cells were incubated with species- and isotype-matched Alexa Fluor-conjugated secondary antibodies (1:2000; Invitrogen) and counterstained with DAPI (1 μg/mL; Invitrogen #D1306) for 30 min at 25°C before final washes with PBS and ddH₂O.

Immunohistochemistry

Mice were transcardially perfused with 4% paraformaldehyde/PBS. Isolated tissue was post-fixed for 24 h in 4% paraformaldehyde and normalized for 48 h in 20% sucrose. Brains were sectioned at 15-μm thickness in OCT (Tissue Tek) using a cryotome and immunostained with Ki67 (1:250; BD Pharmingen #550609), Olig2 (1:500; Millipore #MABN50), GST-pi (1:500; BD Biosciences #610718), Fluoromyelin (Invitrogen), and DAPI (Invitrogen). Mosaic images were acquired using a Leica TCS SP5 laser confocal microscope with a 40x oil immersion lens. Data represent analyses of the corpus callosa of at least three WT and three twitcher brains from separate litters.

Analysis of Cell Survival

Cells were incubated with 200 nM calcein-AM and 1 μg/mL prodium iodide for 30 min at 37°C to determine the proportion of live and dead cells, respectively, in an experimental condition. Single-cell analysis was performed using a Celigo cytometer (Nexcelom) and the %Live corrected values are reported (calcein⁺PI⁻).

Analysis of Cell Proliferation Rate

The total number of cells per well was determined using Brightfield analysis with a Celigo cytometer (Nexcelom) daily across 5 d, and cell numbers were normalized to the number of cells at the beginning of the experiment (Day 0). The proliferation rate was calculated as the fold change in cell number per unit time (days) using linear regression (0.95 < R² < 1.0 over 5 d) for each well.

Differentiation

To induce differentiation, purified (A2B5⁺) O-2A/OPCs were exposed to DMEM:F12 complete media supplemented with 1 ng/mL PDGF-AA and 45 nM T3/T4 mixture (Sigma #T6397/#0397) and allowed to differentiate for 5 d. Cells were fixed in 4% paraformaldehyde, stained with antibodies against A2B5 (1:4; in-house hybridoma ATCC) and GalC (1:4; in-house hybridoma, ATCC), and counterstained with DAPI (1 μg/mL; Invitrogen #D1306), and the numbers of GalC⁺ OLs and A2B5⁺GalC⁻ progenitor cells per condition were quantified.

Small-Molecule Screen

For analyses of cell division, purified (A2B5⁺) O-2A/OPCs were exposed to 1 μM Psy with each of the 1,040 compounds in the NINDS II Custom Collection library (Microsource) diluted to a final concentration of 0.2, 1, and 5 μM, each in duplicate, 15 compounds per plate. Each plate had control wells (in triplicate) of cells exposed to either vehicle (0.01% DMSO) or Psy (1 μM) alone. The proliferation rate was determined by daily cell counting using a Celigo cytometer (Nexcelom) across 5 d, as outlined above, and the increases in cell number within each well were internally normalized to the number of cells in that well at Day 0 (before the addition of Psy/compounds). The calculated proliferation rates were then normalized to the mean proliferation rate of in-plate vehicle-treated controls.

Selected “hits” in both screens (15 for survival and 36 for proliferation) were rescreened at 0.2, 1, and 5 μM for their ability to significantly reduce Psy-induced suppression of division across 5 d. Those compounds that significantly reduced Psy toxicities with at least one of the three selected concentrations (22 in total) were rescreened using nine-point dose-response curves, ranging from 1 nM to 10 μM, to identify optimally protective concentrations to significantly reduce Psy-induced cell death at 5 d and/or suppression of division across 5 d. The list of 22 compounds that significantly reduced Psy-induced suppression of division was shortened to 15 by elimination of those minimally protective compounds for which commercial sources were cost prohibitive or unavailable.

Growth Factor Screen

For analyses of cell division, cells were exposed to 1 μM Psy with each growth factor (including BSA, with which all growth factors were diluted) diluted to a final concentration of 10, 33, and 100 ng/mL, each in triplicate, six growth factors per plate. Each plate had control wells (in triplicate) of cells exposed to either vehicle (0.01% DMSO) or Psy (1 μM) alone. The proliferation rate was determined by daily cell counting using a Celigo cytometer (Nexcelom) across 5 d, as outlined above, and the increases in cell number within each well were internally normalized to the number of cells in that well at Day 0 (before the addition of Psy/compounds). The calculated proliferation rates were then normalized to the mean proliferation rate of in-plate vehicle-treated controls.

Fingerprint Analysis

Concentrations of small-molecule inhibitors of signaling proteins were selected based on their ability to reduce phosphorylation of target proteins when analyzed by immunoblot when possible; the selected concentrations minimally enhanced—or did not alter—Psy’s effects on cell division when examined in the absence of protective agents. Purified (A2B5⁺) O-2A/OPCs were pretreated with the small-molecule inhibitors for 1 h before the addition of protective agent and Psy (each in triplicate). All plates had vehicle- and Psy-treated wells (each in triplicate), as well as Psy combined with the protective agent of interest, wells as controls. The proliferation rates were analyzed and normalized against vehicle-treated controls, as outlined above. Data were hierarchically clustered with an unweighted Euclidean distance similarity metric (complete linkage clustering) using Cluster 3.0 and visualized using TreeView.

Delayed Administration of Protective Agents

For experiments in which the administration of candidate protective agents was delayed, purified (A2B5⁺) O-2A/OPCs were exposed to the indicated concentrations of Psy for 2 d before each small molecule or IGF-1, prepared at a 10X concentration in DMEM:F12 complete media supplemented with 10 ng/mL PDGF-AA, was diluted to 1X so as to minimize dilution of Psy and perturbation of cells. An equal volume of DMEM:F12 complete media supplemented with 10 ng/mL PDGF-AA was added to untreated and Psy-only control wells. The proliferation rate was calculated from the daily change in cell number from time of administration (Day 2) for three days (Day 5), as outlined above.

Lysosomal pH Measurements

Purified (A2B5⁺) O-2A/OPCs were plated on poly-L-lysine-coated glass-bottom microwell dishes (MatTek Co., Ashland, MA; #P35G-1.5–14) in DMEM:F12 complete media supplemented with 10 ng/mL PDGF-AA after passaging. Cells were loaded with 500 μg/mL LysoSensor Yellow/Blue Dextran (Invitrogen) in complete media with PDGF for 24 h prior to treatment. After 24 h, the cells were fixed in 4% paraformaldehyde and were imaged using a Leica TCS SP5 laser confocal microscope with a 63X oil immersion lens. Using an excitation wavelength of 335 nm (405 diode), emission spectra at 450 nm (acidic) and 521 nm (alkaline) were quantified, and the ratio of these emissions was calculated using the Leica Advanced Fluorescence software. Live-cell imaging was performed as above, with the exception that cells were not fixed in PFA prior to imaging and analysis. To generate the lysosomal pH calibration curve, the pH of pre-loaded O2A/OPCs was measured as previously described [112]. Briefly, the cells were incubated in calibration buffers (20 mM MES, 110 mM KCl, and 20 mM NaCl containing 10 μM monensin and 20 μM nigericin; Sigma) adjusted to known pH values between 4.0 and 6.0 at 0.5 increments using HCl/NaOH for 1 h prior to imaging, and ratiometric quantification, as above. Calibration curves were generated using both fixed and live cells.

Endocytosis Measurements

Purified (A2B5⁺) O-2A/OPCs were exposed to indicated conditions for 24 h. Cells were trypsinized and resuspended at a density of 10⁶/mL in conditioned (treated) medium and 1:1000 FluoSpheres polystyrene beads (Invitrogen). The cell:bead suspension was incubated at room temperature and gently inverted every 5 min. At indicated time points, 10 μL (10,000 cells) were transferred to 1 mL of ice-cold PBS and pelleted at 13,000 rpm at 4°C for 5 min. Cell pellets were resuspended in ice-cold 2% paraformaldehyde/PBS and transferred to PLL-coated 96-well dishes to adhere during fixation. The integrated fluorescence intensity per cell was measured using a Celigo cytometer (Nexcelom) and plotted over time; the time-to-half-maximal intensity was calculated using curve-fitting software (Prism).

In Situ Cathepsin Activity Measurements

Cathepsin B activity was measured as per manufacturer’s instructions (MagicRed Cathepsin B substrate, ICT). Briefly, purified (A2B5⁺) O-2A/OPCs were treated as indicated for 24 h before exposure to cell-permeant CathB substrate (1 μM); substrate cleavage occurred at 37°C for 1 h before the integrated fluorescent intensity per cell was quantified with a Celigo cytometer (Nexcelom). For cathepsin D measurements, fluorescently labeled, cell-permeant CathD active-site inhibitor (BODIPY-FL Pepstatin A; 10 μM) was added to O-2A/OPCs that had been treated as indicated for 24 h; active CathD labeling occurred at 37°C for 1 h before the integrated fluorescent intensity per cell was quantified with a Celigo cytometer (Nexcelom).

In Situ Lipid Accumulation

Neutral lipid and phospholipid accumulation were quantified with the HCS LipidTox Phospholipidosis and Steatosis Detection Kit (Invitrogen) as per the manufacturer’s directions. Positive controls cyclosporin A (10 μM) and propranolol (10 μM) were used, respectively.

CFTR Knockdown

Rat O-2A/OPCs were exposed to either 50-nM pools of four siRNA constructs targeting rat CFTR or 50-nM pools of four control siRNA constructs that do not target the rat genome with DharmaFECT-1 transfection reagent (1:1000), as per manufacturer’s directions (Dharmacon), for 24 h. Four d post transfection, cells were passaged and either lysed for western blot analysis or plated on glass-bottom dishes for analysis of lysosomal pH, as outlined above. Western blot analysis was performed as previously reported [216] using an anti-rabbit CFTR antibody (Cell Signaling) and HRP-conjugated beta-actin (Santa Cruz).

Psy Quantification

Mice were killed at P35 and transcardially perfused with ice-cold PBS. Isolated tissue was flash-frozen in liquid nitrogen and stored at -80°C until analysis. Analysis of sphingolipids was performed by Dr. Jacek Bielawski from the Lipidomics Core at the Medical University of South Carolina (MUSC) using liquid chromatography-mass spectrometry (LC-MS/MS) and supercritical fluid chromatography-mass spectrometry (SFC-MS/MS) methodologies, as described previously [217].

Animal Treatment

Adult heterozygote (Galc^twi/+) C57Bl/6J (B6.CE-Galctwi/J) mice were originally obtained from Dr. Ernesto Bongarzone (University of Illinois at Chicago, Chicago, IL) and used as breeder pairs to generate homozygous (twi; Galc^twi/twi) twitcher mice and WT (Galc^+/+) C57Bl/6J mice. All animal procedures were approved by the Institutional Animal Care and Use Committee (IACUC) at the University of Rochester School of Medicine and Dentistry and conformed to the requirements of the Animal Welfare Act. In total, three cohorts of aged-matched mice from different litters received daily IP injections beginning at P10: WT mice receiving saline (n = 6); twi receiving saline (n = 6); twi receiving 1 mg/kg NKH-477 (Tocris) in saline (n = 8). Mice were euthanized at P35 and tissue was isolated after transcardial perfusion with 4% paraformaldehyde. Immunohistochemical analyses were completed as outlined above. For survival analysis, animals were provided moistened chow and hydragel water packs and monitored daily for weight gain. Animals were euthanized when moribund, as assessed by when the animals could no longer ambulate to maintain food and water intake or exhibited clinical signs of pain such as hunched posture and ruffled fur, as determined on a daily basis. Animals were euthanized using CO₂ exposure and cervical dislocation.

Motor Behavior Testing

Aged-matched mice from different litters received daily IP injections beginning at P10: WT mice receiving saline (n = 4); twi receiving saline (n = 3); twi receiving 1 mg/kg NKH-477 (Tocris) in saline (n = 4) were analyzed for locomotive ability and gait using the Phenoscan suite and Runwayscan software (CleverSys, Inc) at P25. Multiple parameters, including stance, stride, swing, brake, and propulsion time (milliseconds), stride length (millimeters), and average speed (millimeters/s) were collected for each animal over three compliant trials and averaged for both front and rear paws.

Statistical Analyses

Bar graphs are plotted as mean ± SEM and represent at minimum three independent biological replicates performed in triplicate, except where noted. Two-group comparisons were analyzed using a Student’s t test, and multiple-group comparisons were analyzed using an ANOVA with Bonferroni post-hoc test. Prism (v5.0; GraphPad) was used for data analysis and presentation.