Sigma Receptor Ligands Are Potent Antiprion Compounds that Act Independently of Sigma Receptor Binding

Abstract

Prion diseases are invariably fatal neurodegenerative diseases of humans and other animals for which there are no effective treatment options. Previous work from our laboratory identified phenethylpiperidines as a novel class of anti-prion compounds. While working to identify the molecular target(s) of these molecules, we unexpectedly discovered ten novel antiprion compounds based on their known ability to bind to the sigma receptors, σ₁R and σ₂R, which are currently being tested as therapeutic or diagnostic targets for cancer and neuropsychiatric disorders. Surprisingly, however, knockout of the respective genes encoding σ₁R and σ₂R (Sigmar1 and Tmem97) in prion-infected N2a cells did not alter the antiprion activity of these compounds, demonstrating that these receptors are not the direct targets responsible for the antiprion effects of their ligands. Further investigation of the most potent molecules established that they are efficacious against multiple prion strains and protect against downstream prion-mediated synaptotoxicity. While the precise details of the mechanism of action of these molecules remain to be determined, the present work forms the basis for further investigation of these compounds in preclinical studies. Given the therapeutic utility of several of the tested compounds, including rimcazole and haloperidol for neuropsychiatric conditions, (+)-pentazocine for neuropathic pain, and the ongoing clinical trials of SA 4503 and ANAVEX2–73 for ischemic stroke and Alzheimer’s disease, respectively, this work has immediate implications for the treatment of human prion disease.

Graphical Abstract

INTRODUCTION

Prion diseases are invariably fatal neurodegenerative diseases of humans and other animals.¹ The most common human prion disease, Creutzfeldt–Jakob disease (CJD), accounts for ∼95% of cases; it has a sporadic etiology with a worldwide incidence of 1–2 cases per million people per year. A smaller number of cases (5–10%) are due to germline mutations in PRNP (Prnp in mouse gene nomenclature), the gene encoding the prion protein (PrP^C).² Fewer still are due to infection as a result of exposure to contaminated tissues, by either ingestion or nosocomial means.³

The central event of prion disease is the structural rearrangement of PrP^C, a cellular glycoprotein with α-helical and unstructured domains, to PrP^Sc, a primarily β-sheet conformer that propagates by a self-templated seeding mechanism.^4,5 This three-dimensional change endows PrP^Sc with biochemical properties that differentiate it from PrP^C, including resistance to proteolytic digestion and insolubility in detergents. Following protracted and clinically silent incubation periods, accumulation of PrP^Sc in the central nervous system (CNS) results in the loss of neurons and eventual death of the host.¹ Divergent PrP^Sc atomic structures with identical amino acid sequences are the basis of prion strains, which cause diverse clinical and pathological outcomes by currently unknown mechanisms.^4,6

Extensive efforts have been devoted to identifying effective therapeutics for prion disease. Most antiprion compounds have been discovered by exposing prion-infected mouse neuroblastoma cells (ScN2a) to compound libraries and assessing changes in the levels of proteinase K (PK)-resistant PrP. Molecules discovered in this manner have been shown to prolong the disease course of mice infected with mouse prions but none have yet been effective in transgenic mice expressing human PrP infected with CJD prions or in patients.⁷ The efficacy of antiprion compounds has also been shown to be strain dependent, both in vitro and in vivo^8,9 complicating efforts to develop therapeutics. It has been demonstrated that some of these molecules prevent prion propagation through a direct interaction with PrP^C by stabilizing its structure or by sterically occluding interactions with PrP^Sc that are critical for conversion.^10,11 Others are known to inhibit the conversion process by direct interaction with PrP^Sc.¹² The majority of antiprion compounds, however, do not interact with either PrP conformer and, presumably, target other molecules.¹³ These compounds present a more attractive pharmaceutical target as they could potentially be used to combat multiple prion strains.^4,6

Our laboratory recently identified phenethylpiperidines as a novel class of antiprion compounds.¹⁴ This discovery was made through application of the drug-based cellular assay (DBCA) in a high-throughput screen for molecules that suppress the increased sensitivity to antibiotics induced by the expression of an internally deleted form of PrP (ΔCR).¹⁵ These compounds were refined through structure–activity studies, and the most potent derivative, JZ107, was capable of permanently curing ScN2a cells of infection by multiple prion strains. JZ107 also prevented prion-induced synaptotoxicity in cultured hippocampal neurons.^14,16,17 Phenethylpiperidine molecules do not appear to bind to PrP at concentrations used in these assays, making the identity of their interaction partners of significant interest.

Here, we describe our efforts to identify the molecular target of JZ107. We discovered that ligands of the sigma receptors (σ₁R and σ₂R) potently reduced the levels of PK-resistant PrP, a biochemical indicator of infection, in ScN2a cells and prevented prion-induced retraction of dendritic spines on hippocampal neurons. Surprisingly, however, we find that these effects are independent of the direct interaction of the compounds with the sigma receptors. Because these molecules are known to cross the blood–brain barrier, and some are already in clinical use for other diseases, they make excellent candidates for preclinical therapeutic studies.

RESULTS

Discovery of Novel Antiprion Compounds.

We previously described the ability of a phenethylpiperidine, JZ107, to reduce the levels of PrP^Sc in ScN2a cells infected with two prion strains (RML and 22L) at low micromolar concentrations.¹⁴ In an effort to identify potential non-PrP molecular targets of JZ107, we utilized the National Institute of Mental Health Psychoactive Drug Screening Program (NIMH PDSP), which assays the binding affinity of small molecules to a panel of central nervous system channels, receptors, and transporters.¹⁸ By performing radioligand binding competition assays, a list of inhibition constants (K_i) was obtained for JZ107 and an inactive analogue, JZ103, against 37 potential target proteins (Table 1). We found that the two targets in the PDSP panel with the highest affinity for JZ107 were the sigma-1 and sigma-2 receptors (σ₁R and σ₂R), which bound JZ107 with K_i values of 7.9 and 5.1 nM, respectively. In contrast, the K_i values for JZ103 binding were 339 nM (σ₁R) and 292 nM (σ₂R), 43× and 57× higher, respectively, than JZ107. Transcriptomic data obtained through RNA-seq was used as an additional filter of this data set, with 0.5 FPKM (Fragments Per Kilobase of transcript per Million mapped reads) chosen as the lower limit for the expression of each gene. Surprisingly, expression of the Sigmar1 (σ₁R) and Tmem97 (σ₂R) genes in N2a cells was found to be higher than that of any other proteins assayed in the PDSP (Table 1).

Table 1.

Binding Affinities Obtained through the PDSP and Expression Values in N2a Cells Ranked by Increasing JZ107 K_i

gene	protein	ki JZ107 (nM)	ki JZ103 (nM)	mRNA expression (FPKMa)
Tmem97	σ2	5.1	292	48.78
Sigmar1	σ1	7.9	339	12.85
Drd3	D3	13.7	4396.3	0.02
Hrh2	H2	30	522	0.00
Oprm1	MOR	40.3	>10,000	0.00
Htr2b	5-HT2B	52.3	>10,000	0.00
Drd2	D2	101	>10,000	0.02
Htr1a	5-HT1A	151.3	36.5	0.00
Htr2c	5-HT2C	181.8	1848.7	0.00
Htr6	5-HT6	289	>10,000	1.37
Slc6a4	SERT	375	>10,000	0.01
Htr2a	5-HT2A	444	442.7	0.00
Drd4	D4	550.5	>10,000	0.61
Drd1a	D1	909.3	>10,000	0.08
Slc6a2	NET	927	>10,000	0.00
Adra1b	α1B	1125	>10,000	0.45
Slc6a3	DAT	1401	2720	0.00
Adra1a	α1A	1486	859.5	0.00
Chrm3	M3-muscarinic	1495	>10,000	0.19
Chrm2	M2-muscarinic	1567.7	>10,000	0.02
Oprk1	KOR	1909.7	>10,000	0.00
Adra1d	α1D	2123	>10,000	0.00
Hrh3	H3	2336.7	1262	3.04
Chrm4	M4-muscarinic	2613	>10,000	7.27
Hrh4	H4	2797.7	>10,000	0.00
Chrm1	M1-muscarinic	3107.5	>10,000	0.04
Htr5a	5-HT1A	7596.3	>10,000	0.00
Htr1d	5-HT1D	7746	2682.7	0.02
Adra2b	α2B	>10,000	1525	1.08
Hrh1	H1	>10,000	>10,000	2.24
Htr1b	5-HT1B	>10,000	>10,000	0.01
Drd5	D5	>10,000	>10,000	0.00
Adra2c	α2C	>10,000	374	0.00
Adrb3	β 3	>10,000	>10,000	0.01
Adra2a	α2A	>10,000	1615.5	0.02
Chrm5	M5-muscarinic	>10,000	>10,000	0.00
Htr7	5-HT7	>10,000	2697.7	0.01

^aFragments per kilobase of transcript per million mapped reads (FPKM).

These results led us to test several known sigma receptor ligands for their ability to reduce the levels of PK-resistant PrP in cell culture. RML-infected ScN2a cells were cultured in the presence of increasing concentrations of each compound or dimethyl sulfoxide (DMSO) vehicle for a total of 7 days, with passaging on the third day. Cell lysates were then exposed to 10 μg/mL of PK before western blotting, and MTT assays were performed in parallel as a measure of cell viability. Using this approach, we identified ten novel antiprion compounds (Figure 1A–J; Supplementary Figure 1A–J): PD 144418¹⁹ (EC₅₀ = 5.2 μM); BD1047²⁰ (EC₅₀ = 19 μM); BD1063²⁰ (EC₅₀ = 12.3 μM); PB-28²¹ (EC₅₀ = 5.1 μM); rimcazole (BW234U)²² (EC₅₀ = 3.5 μM); haloperidol²³ (EC₅₀ = 13.2 μM); SA4503 (Cutamesine)²⁴ (EC₅₀ = 27.2 μM); ANAVEX2–73 (Blarcamesine)²⁵ (EC₅₀ = 39.5 μM); (+)-pentazocine²⁶ (EC₅₀ = 35.8 μM); ditolylguanidine (DTG)²⁷ (EC₅₀ = 68.1 μM). BMY-14802²⁸ was found to be ineffective in lowering the levels of PK-resistant PrP at concentrations up to 50 μM, and PrP^Sc reduction at higher concentrations paralleled a loss of cell viability (Figure 1K; Supplementary Figure 1K). For ease of visual comparison, dose–response curves are plotted together (Figure 1L), and all data, including compound structures, are compiled in Table 2, along with corresponding data for JZ107 and JZ013.¹⁴ Based on MTT assays of cellular toxicity (Figure 1, dashed lines), molecules with antiprion activity exhibited LC₅₀ values ranging from 7.8 to >50 μM, with corresponding therapeutic indices (TI = LC₅₀/EC₅₀) of 1.5 to >4.2 (Table 2).

Figure 1.

Sigma receptor ligands reduce the levels of PK-resistant PrP in ScN2a-RML cells. N2a cells chronically infected with RML prions were incubated with the indicated concentrations of compound for a total of 7 days, before being lysed for PK digestion and analysis by western blot. Cultures treated in parallel were subjected to MTT assay. All data points represent three independent replicates. (A) PD 144418: EC₅₀ = 5.2 μM, R² = 0.9367; LC₅₀ = >22 μM. (B) BD1047: EC₅₀ = 19 μM, R² = 0.8873; LC₅₀ = >25 μM. (C) BD1063: EC₅₀ = 12.3 μM, R² = 0.9591; LC₅₀ = >50 μM. (D) PB-28: EC₅₀ = 5.1 μM, R² = 0.9023; LC₅₀ = 7.8 μM, R² = 0.9398. (E) Rimcazole: EC₅₀ = 3.5 μM, R² = 0.9567; LC₅₀ = 10 μM, R² = 0.9305. (F) Haloperidol: EC₅₀ = 13.2 μM, R² = 0.9649; LC₅₀ = >25 μM. (G) SA 4503: EC₅₀ = 27.2 μM, R² = 0.9662; LC₅₀ = >50 μM. (H) ANAVEX2–73: EC₅₀ = 39.5 μM, R² = 0.9705; LC₅₀ = >50 μM. (I) (+)-pentazocine: EC₅₀ = 35.8 μM, R² = 0.9773; LC₅₀ = >50 μM. (J) DTG: EC₅₀ = 68.1 μM, R² = 0.9696; LC₅₀ = >50 μM. (K) BMY-14802: EC₅₀ = not active <50 μM; LD₅₀ = >50 μM. (L) EC₅₀ curves of A–K plotted together. All curves were fit by least-squares regression using GraphPad software, n = 3 replicates.

Table 2.

Structures and Pharmacological Properties of Compounds Used in This Study^a

compound	structure	EC50** (μM)	LC50¶ (μM)	TI*	[SRA]min# (μM)	Kiσ1R (μM)	Kiσ2R (μM)	Selectivity (σ1R/ σ2R or σ2R/ σ1R)	activity
PD 144418		5.2	>22	>4.2	10	0.0011	0.351	σ1R (319.1 x)	antagonist4
BD1047		19	>25	>1.3	0.5	0.0028	0.013	σ1R (4.6 x)	antagonist5
BD1063		12.3	>50	>4.1	1	0.0027	0.1235	σ1R (45.7 x)	antagonist5
PD-28		5.1	7.8	1.5	0.1	0.00256	0.004	σ1R (1.6 x)	σ1R antagonist σ2R agonist3
rimcazole		3.5	10	2.9	0.5	0.1577	0.0496	σ2R (3.2 x)	antagonist2
haloperidol		13.2	>25	>1.9	2.5	0.00328	0.02921	σ1R (8.9 x)	antagonist4, 5
SA 4503 (Cutamesine)		27.2	>50	>1.8	ND	0.0051	0.0175	σ1R (3.4 x)	agonist7
ANAVEX2–73 (Blarcamesine)		39.5	>50	>1.3	ND	0.442	2.026	σ1R (4.5 x)	agonist8
(+)-pentazocine		35.8	>50	>1.4	ND	.00549	2.46711	σ1R (457 x)	agonist12
DTG		68.1	>50	ND	ND	.0719	.0219	σ2R (3.4x)	agonist10
Other compunds:
amiodarone		4.4	6.3	1.4	ND	.0852	0.171	σ2R (2 x)	unknown
JZ107		3.11	11.41	3.71	5.01	0.0079	0.0051	σ2R (1.5 x)	unknown
JZ103		not active up to 7.5 1	ND1	NA1	Not active	0.339	0.292	σ2R (1.2 x)	unknown
BMY-14802		not active up to 50	>50	NA	ND	0.085	0.0825	σ2R ≈ σ2R	antagonist6

^a**EC₅₀ = Effective concentration for 50% reduction of PrP^Sc levels in ScN2a-RML cells.

^¶LC₅₀ = Lethal concentration for 50% reduction of cell viability based on MTT assay.

^*Therapeutic Index,

^#Minimum concentration capable of preventing hippocampal spine retraction. Numbered footnotes indicate information reported in the following references: 1¹⁴, 2²⁹, 3³⁰, 4³¹, 5³², 6³³, 7²⁴, 8²⁵, 9³⁴, 10³⁵, 11³⁶, 12³⁷. ND = not determined, NA = not applicable.

Sigma Receptor Binding Characteristics of Antiprion Compounds.

K_i values reported in the literature for sigma receptor ligands are often difficult to compare, as they have been obtained using different assay methods and/or source tissues. We therefore submitted our set of antiprion compounds to the PDSP to experimentally verify their sigma receptor binding characteristics. The PDSP assays the ability of test compounds to compete with the binding of [³H]-(+)-pentazocine and [³H]-DTG to membrane preparations of receptor-expressing HEK293T cells to determine K_i values for σ₁R and σ₂R, respectively. We observed that, irrespective of antiprion activity, all compounds bound to both σ₁R and σ₂R, most with submicromolar affinity (Table 2). Of the ligands with antiprion activity, PB-28 and JZ107 had similar affinities for σ₁R and σ₂R (1.6X and 1.5X selectivity for σ₁R and σ₂R, respectively). Showing slight selectivity for σ₁R were BD1047 (4.6×), haloperidol (8.9×), SA 4503 (3.4×), and ANAVEX2–73 (4.5×), while rimcazole and DTG displayed the opposite preference (3.2X and 3.4X for σ₂R, respectively). Finally, three ligands were highly selective for σ₁R: PD 144418 (319.1×), BD1063 (45.7×), and (+)-pentazocine (457×).

We attempted to correlate the observed antiprion effects with the categorization of these compounds as agonists or antagonists of the sigma receptors. This categorization must be regarded cautiously since there is no agreed upon biochemical or physiological activity of either receptor to serve as a criterion for assigning the pharmacological effects of its ligands (Table 2). Although most compounds with antiprion activity have been classified as antagonists in published literature, some (SA 4503 and ANAVEX2–73) have been classified as agonists and one (PB-28) as an antagonist of σ₁R and an agonist of σ₂R. We also assessed the relationship between the observed antiprion activities of these molecules with their binding affinity for the two sigma receptors. When we plotted the EC₅₀ values determined using ScN2a cells against K_i values for σ₁R and σ₂R determined using the PDSP, we observed no correlation between these two parameters for either receptor (Figure 2). Taken together, these results demonstrate no significant correlation between the antiprion potency of the ten ligands and their affinity for or purported pharmacological action on σ₁R and σ₂R.

Figure 2.

Correlation of sigma receptor binding affinity and antiprion activity. The corresponding K_i for each compound for σ₁R or σ₂R as determined by the PDSP is plotted against the EC₅₀ determined against RML prions. PDSP-derived K_i values of (+)-pentazocine and DTG were determined by another group.^34,36 (A) σ₁R K_i vs RML EC₅₀: r = 0.3149, p = 0.3455. (B) σ2R K_i vs RML EC₅₀: r = 0.3701, p = 0.2626. Correlation analysis was performed using GraphPad software.

Sigma Receptor Knockout Does Not Abrogate Antiprion Effects.

To investigate the role of σ₁R and σ₂R in prion propagation and in the antiprion effects of their ligands, we edited both the Sigmar1 and Tmem97 genes in ScN2a cells using CRISPR/Cas9 to disrupt the respective open reading frames. Since N2a cells are well-known to display clonal variations in prion infection susceptibility and propagation,³⁸ we utilized a CRISPR/Cas9 platform capable of high-efficiency gene editing without the need for cloning (see Experimental Procedures). As a control, we used a sgRNA targeting the Rosa26 safe harbor locus. Editing efficiency was determined by Sanger sequencing of PCR amplicons and deconvolution using the Inference of CRISPR Edits (ICE) tool.³⁹ The editing efficiency (percentage of open reading frames with indels) and the knockout score (percentage of edits producing a frameshift) were >94% for both Sigmar1 and Tmem97 (Figure 3A). Quantitative reverse transcriptase polymerase chain reaction (RT-PCR) analysis demonstrated that Sigmar1 and Tmem97 transcripts underwent incomplete nonsense-mediated decay, with transcript levels of Sigmar1 and Tmem97 decreasing relative to Rosa26-targeted cells (Figure 3B). Confirming DNA sequence analysis, σ₁R and σ₂R could not be detected in Sigmar1 + Tmem97 knockout cells by western blot (Figure 3C; Supplementary Figure 2). These genetic manipulations had no effect on either the transcript level of Prnp (Figure 3B) or upon the basal levels of total or PK-resistant PrP (Figure 3C,D; Supplementary Figure 2).

Figure 3.

Combined CRISPR/Cas9-mediated editing of the σ₁R and σ₂R genes has no effect on basal levels of PrP^C or PrP^Sc. (A) Inference of CRISPR edits (ICE) analysis of Sigmar1 and Tmem97 gene disruption. Following PCR amplification of the targeted loci, ICE analysis allows for the deconvolution of Sanger sequencing data to provide indel % (editing efficiency) and a knockout score (KO: proportion of edits that result in a frameshift). The Rosa26 safe harbor locus is targeted with a single guide RNA as a control. (B) Quantitative RT-PCR analysis using the ΔΔC_t method. Fold expression in Sigmar1 + Tmem97 cells relative to Rosa26-targeted cells ± SEM: Sigmar1, 0.4 ± 0.012, p = .002; Tmem97, 0.07 ± 0.003, p < 0.0001; Prnp, 1.06 ± 0.018, p = 0.282. (C) High knockout efficiency is demonstrated by undetectable levels of σ₁R and σ₂R protein in double targeted cells by western blot. Blotting for PrP with and without PK demonstrates that levels are unaffected by the ablation of Sigmar1 and Tmem97. (D) Quantification of PrP blots in (C). For all experiments n = 3 replicates. p < 0.0001 = ****; p < 0.001 = ***; p < 0.01 = **; p < 0.05 = *; ns = not significant using a two-tailed students t test.

We next sought to determine if the knockout of σ₁R and σ₂R influenced the ability of sigma receptor ligands to reduce PrP^Sc levels. For these experiments, we subjected these cells to our standard drug treatment regimen using the measured EC₅₀ of each compound against RML prions in unedited cells (Figure 1; Table 2). Surprisingly, we found that dual knockout of Sigmar1 and Tmem97 did not affect the ability of any of the compounds to reduce the levels of PK-resistant PrP (Figure 4; Supplementary Figure 3). These data demonstrate that, despite being sigma receptor ligands, the antiprion activity of these compounds does not require expression of σ₁R or σ₂R.

Figure 4.

Elimination of σ₁R and σ₂R expression does not alter the antiprion effects of sigma receptor ligands. Compound efficacy against RML prions in chronically infected Sigmar1 + Tmem97 KO cells. For each compound, the EC₅₀ measured against RML prions in N2a cells was used. Levels of PK-resistant PrP relative to DMSO-treated controls in Rosa26-targeted cells vs Sigmar1 + Tmem97 KO cells ± SEM: (A) JZ107 (3.1 μM): 21.9 ± 3.6% (p = 0.0004) vs 10.3 ± 4.7% (p = 0.0001); (B) PD 144418 (5.2 μM): 81.1 ± 5.0% (p = 0.0398) vs 56.6 ± 5.8% (p = 0.0046); (C) BD1047 (19 μM): 67.98/± 3.0% (p = 0.0156) vs 58.0 ± 4.6% (p = 0.0023); (D) BD1063 (12.3 μM): 66.0 ± 3.5% (p = 0.0073) vs 60.3 ± 4.0% (p = 0.0016); (E) PB-28 (5.1 μM): 51.4 ± 3.5% (p = 0.0033) vs 37.4 ± 5.8% (p = 0.0011); (F) Rimcazole (3.5 μM): 32.1 ± 4.5% (p = 0.0179) vs 18.6 ± 4.4% (p = 0.0001); (G) haloperidol (13.2 μM): 65.1 ± 2.0% (p = 0.0017) vs 52.3 ± 3.9% (p = 0.0006); (H) SA 4503 (27.2 μM): 54.2 ± 2.5% (p = 0.0042) vs 35.4 ± 3.8% (p = 0.0006); (I) ANAVEX2–73 (39.5 μM): 70.4 ± 1.1% (p = 0.0015) vs 57.9 ± 9.0% (p = 0.0240); (J) (+)-pentazocine (35.8 μM): 64.4 ± 1.9% (p = 0.0152) vs 38.1 ± 5.1% (p = 0.0013); (K) DTG (68.1 μM): 80.5 ± 1.6% (p = 0.0029) vs 66.3 ± 5.0% (p = 0.0068). For all experiments, n = 3 replicates, degrees of freedom = 4. p < 0.0001 = ****; p < 0.001 = ***; p < 0.01 = **; p < 0.05 = *; ns = not significant using the Holm–Sidak test for multiple comparisons.

Other Binding Partners Are Not Responsible for Antiprion Effects.

While σ₁R or σ₂R were the targets identified by the PDSP that were most highly expressed in N2a cells and that had the highest affinity for the test compound JZ107 (Table 1), other proteins that interact with one or more of sigma receptor ligands are expressed at lower levels (Figure 5A; Table 1; Supplementary Table 1). Using the same strategy employed for Sigmar1 and Tmem97, we disrupted the open reading frames of Htr6, Drd4, Hrh1, Hrh3, and Chrm4. Attempts to detect the corresponding proteins by western blot were unsuccessful, likely due to their low levels of expression in N2a cells (Table 1). However, disruption of the open reading frames observed through ICE analysis gives us confidence that these genes are no longer functional (Figure 5B). Again using each compound at its previously determined EC₅₀, we were unable to demonstrate a requirement for any of these proteins for the observed antiprion effects of the tested compounds (Figure 5C–N; Supplementary Figure 4).

Figure 5.

Other receptors identified in the PDSP do not mediate the observed antiprion effects. (A) Graphical representation of active compound K_i values for other receptors known to be expressed in N2a cells (FPKM > 0.5). (B) Inference of CRISPR Edits (ICE) analysis of gene disruption. The Rosa26 safe harbor locus is targeted with a single guide RNA as a control. (C–N) Compound efficacy against RML prions in chronically infected KO cells. For each compound, the EC₅₀ determined against RML prions in N2a cells was used. Levels of PK-resistant PrP are expressed relative to DMSO-treated controls in Rosa26-targeted cells vs KO cells SEM: (C) 3.1 μM JZ107 against Htr6^0/0 cells: 53.0 ± 2.7% (p = 0.0012) vs 18.7 ± 4.4% (p = 0.0002); (D) 3.1 μM JZ107 against Drd4^0/0 cells: 46.6 ± 2.4 (p = 0.0010) vs 25.2 ± 6.6 (p = 0.0012); (E) 5.1 μM PB-28 against Drd4^0/0 cells: 33.1 ± 3.1% (p = 0.0138) vs 34.2 ± 8.6% (p = 0.0035); (F) 3.5 μM rimcazole against Hrh1^0/0 cells: 33.9 ± 1.4% (p = 0.0004) vs 20.0 5.9% (p = 0.0005); (G) 5.1 μM PB-28 against Hrh1^0/0 cells: 50.3 ± 2.5% (p = 0.0021) vs 44.9 ± 4.8% (p = 0.0034); (H) 12.3 μM BD1063 against Hrh1^0/0 cells: 24.4 ± 2.4% (p = 0.0070) vs 42.6 ± 6.7% (p = 0.0189); (I) 13.2 μM haloperidol against Hrh1^0/0 cells: 43.6 ± 2.9% (p = 0.0006) vs 31.7 ± 7.3% (p = 0.0022); (J) 19 μM BD1047 against Hrh1^0/0 cells: 56.2 ± 2.7% (p = 0.329) vs 36.4 ± 8.6% (p = 0.0051); (K) 39.5 μM ANAVEX2–73 against Hrh1^0/0 cells: 42.6 ± 1.7% (p = <0.0001) vs 25.7 ± 6.2 (p = 0.0008); (L) 3.1 μM JZ107 against Hrh3^0/0 cells: 53.0 ± 27.1 9.7% (p = 0.0401) vs 29.0 ± 6.5% (p = 0.0024) 2.7% (p = 0.0012) vs 25.9 ± 27.1 9.7% (p = 0.0401) vs 29.0 ± 6.5% (p = 0.0024) 2.6% (p = 0.0020); (M) 5.1 μM PB-28 against Hrh3^0/0 cells: 27.1 ± 9.7% (p = 0.0401) vs 29.0 ± 6.5% (p = 0.0024); (N) 3.1 μM JZ107 against Chrm40/0 cells: 46.6 27.1 ± 9.7% (p = 0.0401) vs 29.0 ± 6.5% (p = 0.0024) 2.4% (p = 0.0010) vs 40.4 ± 27.1 9.7% (p = 0.0401) vs 29.0 ± 6.5% (p = 0.0024) 3.6% (p = 0.0006). Assays shown in C and L and in D and N were done using the same western blot, necessitating the use of the same Rosa26 control images. For all experiments, n = 3 replicates, degrees of freedom = 4. p < 0.0001 = ****; p < 0.001 = ***; p < 0.01 = **; p < 0.05 = *; ns = not significant using the Holm?Sidak test for multiple comparisons.

Phospholipidosis as a Potential Antiprion Mechanism.

σ₁R and σ₂R were recently identified in a screen for host molecules that interact with SARS-CoV-2 proteins, and it was demonstrated that known sigma receptor ligands, including some under investigation here, possessed antiviral activity.⁴⁰ However, a subsequent study found that these effects were not correlated with binding affinity for σ₁R or σ₂R but instead depended on the ability of these molecules to induce phospholipidosis.⁴¹ A phenomenon that often confounds drug discovery efforts, phospholipidosis occurs due to the capacity of cationic, amphiphilic molecules to accumulate in endosomes and lysosomes, resulting in disruptions of lipid processing and alterations in the morphology of these compartments.⁴² Because lysosomal function is known to be important for PrP^Sc degradation, and some of the molecules under investigation here are reported inducers of phospholipidosis, we next explored this as a potential mechanism of action of these antiprion compounds.^41,43

Interestingly, in addition to being a potent inducer of phospholipidosis,⁴¹ the antiarrhythmic medication amiodarone has been previously identified as an antiprion compound⁴⁴ and a sigma receptor ligand.⁴⁵ In our hands, amiodarone has an EC₅₀ of 4.4 μM and an LC₅₀ of 6.3 μM (Figure 6A; Supplementary Figure 5), and K_i values of 85.3 and 171 nM for σ₁R and σ₂R, respectively (Table 2). Here, we used amiodarone as a positive control in phospholipidosis assays, which were monitored using LipidTox, a fluorescent reagent that stains intracellular lipid accumulations.⁴⁶ Similar results were obtained with an alternative reporter, NBD-PE (not shown,⁴¹). Of the most potent compounds, only rimcazole induced phospholipidosis at levels >50% that of amiodarone after incubation with N2a cells at 10 μM for 24 h (Figure 6B,C). Further, the level of phospholipidosis induced by these compounds did not correlate with their antiprion activity in ScN2a cells (r = −0.4436, p = 0.2709) (Figure 6D). These results indicate that phospholipidosis itself is unlikely to be a major mechanism underlying the antiprion effects of the sigma receptor ligands analyzed here although we cannot rule out that other alterations in lysosomal or autophagosomal pathways could be a contributing factor to the action of some of the compounds.

Figure 6.

Phospholipidosis induction is not correlated with antiprion activity. (A) N2a cells chronically infected with RML prions were incubated with the indicated concentrations of amiodarone for a total of 7 days, before being lysed for PK digestion and analysis by western blot. Cultures treated in parallel were subjected to MTT assay. EC₅₀ = 4.4 μM, R2 = 0.9203; LC₅₀ = 6.3 μM, R2 = 0.9305. All data points represent three independent replicates. (B) Uninfected N2a cells were incubated in 10 μM of the indicated compound with LipidTox for 24 h. Green fluorescence represents LipidTox staining, blue represents Hoechst nuclear staining. (C) Total LipidTox fluorescence was normalized to the signal derived from Hoechest. (D) For each compound, the level of phospholipidosis induction relative to DMSO is plotted against the EC₅₀ determined against RML prions. r = –0.4436, p = 0.2709. Correlation analysis was performed using GraphPad software.

Sigma Receptor Ligands Do Not Inhibit Prion Seeding Activity in Vitro.

We next used Real-Time Quaking-Induced Conversion (RT-QuIC) assays⁴⁷ to explore the possibility that these antiprion compounds operate by directly inhibiting the ability of PrP^Sc to induce conformational changes in PrP^C. We tested all compounds at two concentrations, 1× and 2× the EC₅₀ determined against RML prions in ScN2a cells. At these concentrations, each the compounds was present in molar excess over the recombinant PrP^C substrate (4.4 μM), except for rimcazole at the 1× EC₅₀ concentration. With the exception of a slight increase in the lag phase observed in the presence of ANAVEX2–73 at 2× the EC₅₀ (79 μM), none of the compounds inhibited the seeding activity of RML prions at either concentration (Figure 7A–J; Table 2). As a positive control, we performed RT-QuIC reactions in the presence of Congo red, which is known to have several kinds of effects on the conformation, stability, and biochemical properties of both PrP^C and PrP^Sc48 (Figure 7K).

Figure 7.

RT-QuIC assay analysis of antiprion compounds. RT-QuIC assays were performed using full-length recombinant bank vole PrP as substrate⁴⁹ for 50 h. Uninfected (NBH) and prion-infected (RML) brain homogenates were diluted to 10⁻⁵ for use as seeds. For each compound, the following concentrations (1× EC₅₀ and 2× EC₅₀) were used: (A) PD 144418, 5.2 and 10.4 μM; (B) BD1047, 19 and 38 μM; (C) BD1063, 12.3 and 24.6 μM; (D) PB-28, 5.1 and 10.2 μM; (E) rimcazole, 3.5 and 7 μM; (F) haloperidol, 13.2 and 26.4 μM; (G) SA 4503, 27.2 and 54.4 μM; (H) ANAVEX2–73, 39.5 and 79 μM; (I) (+)-pentazocine, 35.8 and 71.6 μM; (J) DTG, 68.1 and 136.8 μM; (K) Congo red, 2.5 μM.

Sigma Receptor Ligands Do Not Alter the Levels or Subcellular Localization of PrP^C.

To further probe the mechanism of action of the sigma receptor ligands, we tested their effects on the total levels and subcellular localization of PrP^C, factors that could potentially influence the cellular content of PrP^Sc. Focusing on compounds with an EC₅₀ of <20 μM, we found that none significantly altered the total levels of PrP^C assessed by western blotting when tested at their EC₅₀ values for antiprion activity (Figure 8A; Supplementary Figure 6). The compounds also had no observable effect on the subcellular localization of PrP^C based on immunofluorescence staining (Figure 8B).

Figure 8.

Antiprion compounds do not alter PrP^C levels or subcellular localization. (A) Effect of compounds on total PrP levels in uninfected N2a cells. For each compound, the EC₅₀ determined against RML prions was used. Following incubation with each compound for 7 days, levels of total PrP, compared to DMSO treatment, were plotted as mean ± SEM: 5.2 μM PD 144418 = 97.5 ± 4.8% (p = 0.8295), 19 μM BD1047 = 94.9 ± 0.9% (p = 0.0553), 12.3 μM BD1063 = 85.8 ± 4.4% (p = 0.2239), 5.1 μM PB-28 = 85.3 ± 3.4% (p = 0.1276), 3.5 μM rimcazole = 99.9 ± 1.6% (p = 0.9735), 13.2 μM haloperidol = 92.0 ± 3.5% (p = 0.3608). (B) Compounds do not alter the cell-surface localization of PrP^C. For each compound, the EC₅₀ determined against RML prions was used. PrP^C was imaged by immunofluorescence staining using D18. 4′,6-Diamidino-2-phenylindole (DAPI) is used as a nuclear counter stain. Scale bar = 10 μm. For all experiments, n = 3 replicates. p < 0.0001 = ****; p < 0.001 = ***; p < 0.01 = **; p < 0.05 = *; ns = not significant using a two-tailed students t test.

Prion Strain Specificity of Sigma Receptor Ligands.

The efficacy of some antiprion compounds has been shown to be strain dependent.^8,50 To examine the potential for strain specificity, we used N2a cells chronically infected with a second strain of prions (22L), again focusing on compounds that displayed an EC₅₀ of < 20 μM against RML prions. We observed that three of these compounds have efficacies against 22L prions that are comparable to those against RML prions: When tested at their EC₅₀ values for RML prions, BD1063 reduced PK-resistant PrP to 63.3 ± 1.2% (p = <0.0001), PB-28 by 52.6 ± 2.6% (p = <0.0001), and haloperidol by 43.0 ± 8.6% (p = 0.0007) of that observed in vehicle-treated cells (Figure 9A, Supplementary Figure 7). Two molecules were more effective against 22L prions than RML prions, resulting in reductions in PK-resistant PrP by BD1047 to 19.1 ± 10.5% (p = 0.0004) and by rimcazole to 13.5 ± 7.8% (p = <0.0001) of that seen in controls (Figure 9A, Supplementary Figure 7). PD 144418, however, was ineffective against 22L prions at the EC₅₀ determined against RML prions (Figure 9A; Supplementary Figure 7). However, a dose–response curve using 22L-infected ScN2a cells revealed that PD 144418 is effective at higher concentrations, with an EC₅₀ of 12.6 μM against this strain compared to an EC₅₀ of 5.2 μM against RML prions (Figure 9B; Supplementary Figure 8).

Figure 9.

Antiprion compounds are efficacious against multiple prion strains. (A) Compound efficacy against 22L prions in chronically infected N2a cells. For each compound, the EC₅₀ determined against RML prions was used. Levels of PK-resistant PrP, relative to DMSO treatment, were plotted as mean ± SEM: 5.2 μM PD 144418 = 92.2 ± 7.5% (p = 0.2144), 19 μM BD1047 = 19.1 ± 10.5% (p = 0.0004), 12.3 μM BD1063 = 63.3 ± 1.2% (p = <0.0001), 5.1 μM PB-28 = 52.6 ± 2.6% (p = <0.0001), 3.5 μM rimcazole = 13.5 ± 7.8% (p = <0.0001), 13.2 μM haloperidol = 43.0 ± 8.6% (p = 0.0007). (B) Dose–response curves of PK-resistant PrP remaining after incubation with increasing concentrations of PD 144418. RML and MTT values are replotted from Figure 1A. Against 22L prions, PD 144418 has an EC₅₀ = 12.6 μM (R² = 0.9622). For all experiments, n = 3 replicates. p < 0.0001 = ****; p < 0.001 = ***; p < 0.01 = **; p < 0.05 = *; ns = not significant using a two-tailed students t test.

Sigma Receptor Ligands Prevent Prion-Induced Synaptotoxicity.

We previously demonstrated that exposure of cultured hippocampal neurons to prion-infected brain homogenates or purified PrP^Sc causes a rapid (<12 h) and dramatic retraction of dendritic spines.^16,51,52 This phenomenon, which we have used as an assay of the synaptotoxic activity of prions, is dependent on the presence of cell-surface PrP^C and is blocked by NMDA receptor antagonists and p38 MAPK inhibitors. Our results suggest that rapid conversion of PrP^C to PrP^Sc on the cell surface triggers an intracellular signaling cascade that ultimately leads to the collapse of the actin cytoskeleton within dendritic spines. In this scenario, compounds that inhibit the conversion of PrP^C to PrP^Sc would be predicted to prevent spine collapse upon prion exposure. To test this prediction, hippocampal neurons (21 DIV) were preincubated with one of the six most potent compounds or DMSO vehicle for 2 h before treatment with either purified RML prions or mock-purified material from uninfected, age-matched control brains (Supplementary Figure 9). Following 24 h of prion exposure in the continued presence of the test compound, spine density was determined by immunofluorescence staining for F-actin. As expected, the application of purified prions to these cultures in the absence of test compound caused dendritic spine collapse, reducing spine density to less than 40% of that observed in cultures treated with mock-purified material (Figure 10; mock vs prion). Strikingly, all six sigma receptor ligands tested in this assay were able to significantly prevent spine loss (Figure 10). Further, PD 144418, PB-28, and haloperidol (Figure 10A,D,F, respectively) restored spine density to levels that were not statistically different from those of neurons treated with a mock-purified material.

Figure 10.

Sigma receptor ligands prevent prion-induced synaptotoxicity. Following a 2 h preincubation with the indicated compound, or DMSO vehicle, hippocampal neurons were exposed to PrP^Sc purified from the brains of RML-infected mice (prion) or to mock-purified material from uninfected brains (mock). Dendritic spine number was assayed 24 h later by imaging F-actin in spines using Alexa Fluor-488 conjugated phalloidin. The mean number of dendritic spines/μm ± SEM is plotted: (A) 10 μM PD 144418: mock = 0.74 ± 0.03, prion = 0.16 ± 0.01, mock + PD 144418 = 0.77 ± 0.01, prion + PD 144418 = 0.73 ± 0.02; mock vs prion, p = <0.0001; mock vs mock + PD 144418, p = 0.3644; mock vs prion + PD 144418, p = 0.7448; prion vs prion + PD 144418, p = <0.0001; mock + PD 144418 vs prion + PD 144418, p = 0.2986. (B) 0.5 μM BD1047: mock = 0.97 ± 0.04, prion = 0.37 ± 0.01, mock + BD1047 = 0.74 ± 0.02, prion + BD1047 = 0.78 ± 0.04; mock vs prion, p = <0.0001; mock vs mock + BD1047, p = <0.0001; mock vs prion + BD1047, p = 0.0033; prion vs prion + BD1047, p = <0.0001; mock + BD1047 vs prion + BD1047, p = 0.2986. (C) 1 μM BD1063: mock = 0.74 ± 0.03, prion = 0.16 ± 0.01, mock + BD1063 = 0.69 ± 0.03, prion + BD1063 = 0.63 ± 0.03; mock vs prion, p = <0.0001; mock vs mock + BD1063, p = 0.219; mock vs prion + BD1063, p = 0.0129; prion vs prion + BD1063, p = <0.0001; mock + BD1063 vs prion + BD1063, p = 0.1841. (D) 0.1 μM PB-28: mock = 0.74 ± 0.03, prion = 0.16 ± 0.01, mock + PB-28 = 0.86 ± 0.03, prion + PB-28 = 0.79 ± 0.03; mock vs prion, p = <0.0001; mock vs mock + PB-28, p = 0.0091; mock vs prion + PB-28, p = 0.3115; prion vs prion + PB-28, p = <0.0001; mock + PB-28 vs prion + PB-28, p = 0.0754. (E) 0.5 μM rimcazole: mock = 0.97 ± 0.04, prion = 0.37 ± 0.01, mock + rimcazole = 0.86 ± 0.02, prion + rimcazole = 0.82 ± 0.02; mock vs prion, p = <0.0001; mock vs mock + rimcazole, p = 0.0176; mock vs prion + rimcazole, p = 0.0018; prion vs prion + rimcazole, p = <0.0001; mock + rimcazole vs prion + rimcazole, p = 0.1672. (F) 2.5 μM haloperidol: mock = 0.85 ± 0.04, prion = 0.24 ± 0.01, mock + haloperidol = 0.85 ± 0.03, prion + haloperidol = 0.73 ± 0.04; mock vs prion, p = <0.0001; mock vs mock + haloperidol, p = 0.9440; mock vs prion + haloperidol, p = 0.1368; prion vs prion + haloperidol, p = <0.0001; mock + haloperidol vs prion + haloperidol, p = 0.0398. p < 0.0001 = ****; p < 0.001 = ***; p < 0.01 = **; p < 0.05 = *; ns = not significant using a two-tailed student’s t test.

DISCUSSION

Identification of Sigma Receptor Ligands as Antiprion Agents.

Using the DBCA, we previously identified phenethylpiperidines as a novel class of antiprion compounds, capable of reducing the levels of PK-resistant PrP in RML and 22L-infected ScN2a cells at low micromolar concentrations.^14,15 To identify the target of these molecules relevant to their antiprion effects, we submitted JZ107, a structurally optimized phenethylpiperidine, and JZ103, an inactive analogue, to the PDSP. In this screen, the highest affinity receptors for JZ107 were σ₁R and σ₂R, both of which were highly expressed in N2a cells (Table 1). Based on this result, we tested known ligands for σ₁R and σ₂R and identified ten novel antiprion compounds: PD 144418, BD1047, BD1063, PB-28, rimcazole, haloperidol, SA 4503, ANAVEX2–73, (+)-pentazocine, and DTG (Figure 1; Supplementary Figure 1; Table 2). Each of these molecules was found to significantly reduce PrP^Sc levels in ScN2a cells and to bind σ₁R and σ₂R with K_i values of 1.1 nM–2.5 μM (Table 2).

Exploration of the Mechanism of Action of Sigma Receptor Ligands.

σ₁R and σ₂R, originally thought to be a novel class of opioid receptors, were eventually shown to be a pharmacologically and molecularly distinct pair of receptors that reside primarily in ER membrane domains. Their endogenous ligands and physiological functions are uncertain.^36,53,54 There is evidence that σ₁R functions as a ligand-gated chaperone for other receptors, and that σ₂R plays a role in cholesterol exchange between the ER and lysosomes. Both receptors have an evolutionary relationship to sterol isomerases. Sigma receptors are being actively investigated as therapeutic targets for treatment of cancer as well as neurological disorders, including neuropathic pain, amyotrophic lateral sclerosis, Alzheimer’s disease, and Parkinson’s disease.⁵⁵

Given their localization in the ER and their involvement in sterol metabolism, it seemed plausible that σ₁R and σ₂R could play a role in the formation or degradation of PrP^Sc in ScN2a cells and thereby represent the molecular targets responsible for the antiprion effects of the compounds identified in this study. However, a comparison of the affinity for σ₁R or σ₂R and the EC₅₀ of these compounds against RML prions reveals that these parameters are not correlated (Figure 2). Further, dual knockout of the genes encoding σ₁R and σ₂R in ScN2a cells did not alter either basal levels of PK-resistant PrP (Figure 3; Supplementary Figure 2) or blunt the susceptibility of the cells to the antiprion effects of the compounds (Figure 4; Supplementary Figure 3).These results conclusively demonstrate that σ₁ R and σ₂ R are not the direct molecular targets responsible for the antiprion effects of these compounds. This was surprising, given that we chose to test the antiprion activity of these compounds based on the fact that they are well-known ligands for the sigma receptors. However, σ₁R and σ₂R are known to be promiscuous in their ligand-binding profiles, and binding to the sigma receptors does not rule out the possibility that these ligands may act via other molecular targets. To explore the possibility that these compounds may be acting through other, distinct, interaction partners, we applied a similar experimental approach to test five other target proteins identified by the PDSP (Htr6, Drd4, Hrh1, Hrh3, and Chrm4). However, similarly to what we observed with the sigma receptors, these additional target knockouts did not diminish the observed antiprion effects (Figure 5; Supplementary Figure 4).

Other Potential Mechanisms of Antiprion Activity.

During the course of this study, the sigma receptors and their ligands were thrust into the spotlight for a potential role in the pathogenesis and treatment of SARS-CoV-2 infection.⁴⁰ However, the affinity of these compounds for the sigma receptors did not correlate with their antiviral activity.⁴¹ It was subsequently determined that the antiviral effects were due to the ability of these compounds to induce phospholipidosis, a poorly understood phenomenon that often confounds drug discovery efforts. Because these findings included some of the compounds under examination in the present work, and because similar mechanisms could plausibly disrupt prion propagation, we next investigated phospholipidosis as a potential mechanism of action for these antiprion compounds (Figure 6; Supplementary Figure 5). While we found no correlation between phospholipidosis induction and the antiprion effects of the molecules currently under investigation, we cannot rule out the possibility that phospholipidosis is a mechanism whereby some antiprion compounds exert their effects, such as amiodarone, or that other kinds of disruptions of lysosomal or autophagosomal pathways contribute to reductions in PrP^Sc.

We next investigated the possibility that the antiprion effects of the molecules are due to a direct interaction with either PrP^C or PrP^Sc. Arguing against this possibility, we found that none of the compounds inhibited the ability of PrP^Sc to seed conversion of recombinant bank vole PrP^C (a universal substrate) in the RT-QuIC assay (Figure 7).^47,49 We note that previous RT-QuIC assays using recombinant hamster PrP as a substrate demonstrated an inhibitory effect of JZ107 but only at concentrations exceeding 150 μM, which is ∼50× its EC₅₀ against RML prions.¹⁴We consider these latter effects to be nonspecific. We also determined that none of the six most potent sigma receptor ligands identified here had a significant effect on the total levels or cellular localization of PrP^C (Figure 8; Supplementary Figure 6). This observation argues against an inhibitory mechanism dependent on reducing or redistributing the PrP^C substrate, as has been proposed for some antiprion compounds.^11,56 Although there were some variations in the potency of the compounds when tested on ScN2a cells infected with the RML and 22L strains, all of them significantly reduced PrP^Sc levels in cells infected with either strain, suggesting lack of a strong preference for particular PrP^Sc conformations of mouse PrP (Figure 9; Supplementary Figures 7 and 8). The efficacy of these compounds against other murine prion strains and strains from other species has not yet been determined.

A striking observation that may provide mechanistic insight is that all six of the most potent sigma receptor ligands, like JZ107,¹⁴ significantly reduced PrP^Sc-induced retraction of dendritic spines on cultured hippocampal neurons (Figure 10; Supplementary Figure 9). We have previously demonstrated that spine retraction and decrements in synaptic transmission occur rapidly (<12 h) after exposure to PrP^Sc, and these events are entirely dependent on the expression of full-length PrP^C by target neurons.^16,51,57 We have also shown that the synaptotoxic effects of PrP^Sc depend on the activation of an NMDA receptor/p38 MAPK-mediated signal transduction cascade, likely initiated by the formation of PrP^Sc at the cell surface. The ability of the sigma receptor ligands to block this cascade could result from their effects on PrP^C or PrP^Sc at the neuronal plasma membrane or on subsequent steps of the synaptotoxic signaling pathway.

Therapeutic Implications.

The compounds identified here have several properties that make them attractive candidates as therapeutics for prion disease. In addition to their ability to inhibit prion propagation in ScN2a cells infected with multiple prion strains, the most potent compounds prevent prioninduced retraction of hippocampal neuron dendritic spines. This event is one of the earliest pathologies observed over the course of prion disease.⁵⁸ Further, these compounds are all known to be blood–brain barrier penetrant, and five of them, rimcazole, haloperidol, SA 4503, ANAVEX2–73, and (+)-pentazocine, have a history of use in humans.^{29,54,59–62} The preclinical molecules SA 4503 and ANAVEX2–73, known, respectively, as cutamesine and blarcamesine, were well tolerated in phase II trials: SA 4503 for ischemic stroke,⁶¹ and ANAVEX2–73 for Alzheimer’s and Parkinson’s diseases.⁶² Despite the negative side effects associated with high and/or prolonged dosing with rimcazole, haloperidol, and (+)-pentazocine,^29,60 all five molecules make excellent therapeutic candidates for testing in preclinical studies

MATERIALS AND METHODS

Psychoactive Drug Screening Program.

All information related to the PDSP, including detailed protocols, can be accessed at https://pdsp.unc.edu/pdspweb/. σ₁R and σ₂R binding assays were done using membrane fractions prepared from HEK 293T cells expressing the respective receptors. [³H]-pentazocine was used to determine K_i values for σ₁R, and [³H]-DTG was used to determine K_i values for σ₂R.

Cell Culture.

For all experiments, N2a cells were grown in Opti-MEM (Gibco) supplemented with 10% FBS (Gemcell) and 100 U/ml each of penicillin/streptomycin (Gibco) in a humidified atmosphere of 5% CO₂ at 37 °C. Chronic infection with RML or 22L prions was achieved by exposure of confluent cultures infected brain homogenate (1% final concentration) for 24 h before removal and subsequent passage.

Compound Treatment of Cells.

Cells were incubated in the presence of each compound or DMSO vehicle (0.1% final DMSO concentration) for 3 days, before being split 1:5 and incubated for an additional 4 days in the continued presence of compound or DMSO vehicle. Cells were then washed with phosphate-buffered saline (PBS) and lysed using cell lysis buffer containing 10 mM Tris pH 7.8, 100 mM NaCl, 0.5% NP-40, 0.5% Na deoxycholate, and 0.1% SDS. JZ107 was synthesized and characterized previously;¹⁴ ANAVEX2–73 was purchased from Selleckchem; (+)-pentazocine was purchased from Cayman Chemical; all other compounds were purchased from Tocris. All compounds are >95% pure.

MTT Assay.

Cells were washed 2× with PBS and incubated in the presence of 0.5 mg/mL MTT at 37 °C for 30 min. This solution was removed, and cells were incubated in DMSO at 37 °C for 10 min. Absorbance was then read at 570 nm using a Synergy H1 plate reader (BioTek). The signal was normalized to DMSO treatment, and curves were fit by least-squares regression using GraphPad software.

Proteinase K Treatment

100 μg of protein, as determined by BCA assay (Pierce), was exposed to 10 μg/mL proteinase K (Roche) in a final volume of 250 μL lysis buffer at 37 °C for 1 h with shaking at 750 rpm. 30 μL of 10× protease inhibitor (Pierce) was added, and samples were centrifuged at 21,130g for 1 h at 4 °C. Supernatant was removed, and pellets were resuspended in 1× Laemmli buffer (BioRad).

Western Blot.

Protein samples were boiled in the presence of 1× Laemmli buffer (BioRad) and loaded into 12% Criterion TGX Precast Protein Gels and run at 200 V for 42 min. Proteins were transferred to poly(vinylidene fluoride) (PVDF) membranes for 45 min at 115 V before gentle washing in 0.1% TBST and blocking in 5% nonfat milk in 0.1% TBST for 1 h. The following primary antibodies were used: D18⁶³ (anti-PrP); σ₁R (B-5, Santa Cruz Biotechnology); σ₂R (26444-1-AP, Proteintech); β-actin (AC-74, Millipore Sigma). All HRP-conjugated secondary antibodies were purchased from BioRad. Signals were visualized using ECL (Millipore). The signal was normalized to DMSO treatment following quantification using ImageJ and plotted using GraphPad software.

Immunofluorescence.

For immunofluorescence staining of PrP^c , cells were plated on coverslips during the passage following 3 days of treatment with compounds. Cells were washed with PBS and fixed using 4% paraformaldehyde in PBS pH 7.4 for 15 min at room temperature. Cells were then washed 3 × 5 min with PBS before incubation in 2% bovine serum albumin (BSA) and 20 mg/mL glycine in PBS with 0.1% Tween-20 for 30 min. Cells were then incubated in 10 μg/mL D18 for 1 h at room temperature by inverting the coverslip over 30 μL of antibody solution spotted on parafilm, washed 3 × 5 min with PBS, and incubated with Alexa Fluor-488 goat anti-human secondary (Invitrogen) for 1 h at room temperature. Cells were then washed 3 × 5 min with PBS and mounted on slides using mounting media with DAPI. Imaging was performed using a Zeiss Axio Observer Z1 microscope equipped with a digital camera (C10600/ORCA-R2 Hamamatsu Photonics). Images were taken using Zen software.

CRISPR/Cas9-Mediated Gene Disruption.

We utilized a highefficiency gene editing kit from Synthego (Gene Knockout Kit v2), which employs three chemically modified sgRNAs for each targeted locus (Rosa26 is targeted by a single guide): Sigmar1 5′-CCGUGUACAACUGUCUCUCC-3′, 5′-CCAGGAGAGACAGUU-GUACA-3′, 5′-CAGGAGAGACAGUUGUACAC-3′; Tmem97 5′-UGCAGUUCAGCAACCUGUUG-3′,5′-ACUGUACCAACCUUU-GAAGA-3′, 5′-CAUAUGCCUUCUUCAAAGGU-3′; Hrh1 5′-CU-CACACAUCUUGUCUUCGG-3′, 5′-GGUUCUAAGUAGUAU-CUCCC-3′, 5′-GAGCGCAAGCUACACACCGU-3′; Hrh3 5′-GCU-C A U G G C G C U G C U C A U C G –3 ′ ,5 ′ -G C C U G C A G C C G C C G C C U C U C - 3 ′ , 5 ′ -A U G -_ GAGCGCGCGCCGCCCGA-3′; Htr6 5′-UUAGACGUGUUGCG-CAGCGC-3′, 5′-GUGCGUGGUCAUCGUGCUGA-3′, 5′-AACA-GUAGCACCCCAGCCUG-3′; Chrm4 5′-GACUGUGGUGGGUAA-CAUCC-3 ′ , 5 ′-CUGCACACGCCAGGCUGAAC-3 ′ , 5 ′ -UGAUGAUGUAUAAGGUGUAA-3′; Drd4 5′-AGCAGCGCUACU-GAGGACGG-3′, 5′-CUGGGGACUGGCGCCGGGCU-3′, 5′-GGGCGUGCUGCUCAUCGGCU-3 ′ ; Rosa26 5 ′-ACTC-CAGTCTTTCTAGAAGA-3′. Recombinant Cas9 containing two nuclear localization signals (Synthego) is then mixed with sgRNAs to form ribonucleoprotein complexes that are then delivered by electroporation. This procedure allowed us to propagate gene-edited cells as a population, without the need for cloning. Electroporation was performed using the Amaxa Cell Line Nucleofector kit V (Lonza Bioscience) with the manufacturer’s settings for N2a cells. To evaluate knockout efficiency, genomic DNA was extracted, and targeted regions were amplified using Q5 high fidelity polymerase (New England Biolabs) using the following primers: Sigmar1-F 5′-AAAGGCCA-GAAGAGGGCATC-3′, Sigmar1-R 5′-CTGGGTGGATGTGAGTG-CAT-3′, Tmem97-F 5′-TGCAGCTCTACAGTAGCATGT-3′, Tmem97-R 5′-GCTGGACACCAAACTGAGGT-3′, Hrh1-F 5′-ACTCCCAGTCTGACCACCAT-3′, Hrh1-R 5′-CGTGCTGGCCA-CATAATCCA-3′, Hrh3-F 5′-GACTCTGGCAGCGGACA-3′, Hrh3-R 5′-CAGATTCCGATACCAGCCCC-3′, Htr6-F 5′-AGCACAT-CACGTCGAAGGC-3′, Htr6-R 5′-AACCCTGTTTTGCCACC-TAC-3′, Chrm4-F 5′-GTCTCCAGTTGGGTTCAGCA-3′, Chrm4-R 5 ′ -GCGGGCTGGATAGGTGA-3 ′ , Drd4 -F 5 ′ -AG-CATTTCTCCCTCTGCCAA-3′, Drd4-R 5′-AGGCTCACCTCG-GAGTAGAC-3′, Rosa26-F 5′-GAGGCGGATCACAAGCAATA-3′,and Rosa26-R 5′-GGGAGGGGAGTGTTGCAATA-3′. Sanger sequencing was performed with the following primers: Sigmar1 5′-TCAAGCAGCTTGGCCAGTAGGGTAG-3′; Tmem97 5′-CAGTTGGAGCATGTGACTGCTGCC-3 ′ ; Hrh1 5 ′ -TGCTGGCCACATAATCCATAGAG-3 ′ ; Hrh3 5 ′ - TCTGGGGGCTTTACCCACGAGGAAGTCGGA-3′; Htr6 5′-GTCCAAAGCAGACAGAGGCCGCGAGCTAGC-3′; Chrm4 5′-GCGGGCTGGATAGGTGAGGG-3′; Drd4 5′-GCTCACCTCG-GAGTAGACAAAG-3′; Rosa26 amplicons were sequenced with the forward primer listed above. Editing efficiency was determined with the Synthego Inference of CRISPR edits (ICE) tool (https://ice.synthego.com).³⁹

RNA-seq and RT-PCR.

Total RNA was extracted using the RNeasy mini plus extraction kit (Qiagen) following the manufacturer’s instructions. RNA purity and concentration were measured using a nanodrop (Thermo). Library preparation and sequencing were performed by Novogene (Beijing, China). Genes with FPKM > 0.5 were considerably expressed. For RT-qPCR analysis, 1 μg of total RNA was converted to cDNA using the iScript cDNA synthesis kit (BioRad). RT-qPCR reactions were performed using Fast Sybr Master Mix (ABI) on a ViiA7 Real-Time PCR system. Relative mRNA expression was determined after normalization to the housekeeping gene Actb. The following primers were used: Sigmar1-F: 5′-GTCTGAGTACGTGCTGCTCTTC-3′, Sigmar1-R: 5′-GAAGACCTCACTTTTCGTGGTGC-3′, Tmem97-F: 5′-TCACGCTGTTCATCGACCTGCA-3′, Tmem97-R: 5′-GGAAGGACTTGAACCACACTGG-3′. Prnp-F: 5′-CAGCAACCAGAACAACTTCGTGC-3′, Prnp-R: 5′-CGCTCCATCATCTTCACATCGG-3′, Actb-F: 5′-CATTGCTGACAGGATGCAGAAGG-3 ′ , Actb -R: 5 ′ -TGCTGGAAGGTGGACAGTGAGG-3′. Relative fold changes were calculated using the delta–delta Ct method.

LipidTox Assay.

Uninfected N2a cells were seeded in black 96-well glass bottom plates and incubated with 20 μM of compound or DMSO alone with 1× HCS LipidTOX Green neutral lipid stain (Invitrogen) for 24–48 h. Cells were then rinsed with 3× PBS before fixation with 4% paraformaldehyde in PBS with 10 μg/mL Hoechest 33342 (Tocris) for 30 min at RT. Cells were then washed with 3× PBS before being read on a plate reader at the following wavelengths: ex/em 495/525 nm for LipidTox; 350/450 nm for Hoechest 33342. LipidTox signal was normalized to the Hoechest 33342 signal for plotting. Fluorescence was also visualized with a Zeiss Axio Observer Z1 microscope using a 10× objective and Zen software.

Real-Time Quaking-Induced Conversion (RT-QuIC) Assay.

Recombinant bank vole PrP (residues 23 to 230; Methionine at position 109; accession no. AF367624) was expressed and purified as described previously.⁴⁹ A molar extinction coefficient at 280 nm of 62,005/M/cm was used to determine the concentration of recombinant PrP. RT-QuIC reactions were performed in black clear bottom 96-well plates using a reaction mixture composed of 10 mM phosphate buffer (pH 7.4), 300 mM NaCl, 0.001% SDS, 1 mM EDTA, 10 μM ThT, and 0.1 mg/mL PrP in a final volume of 98 μL, containing the indicated concentration of the compound in a final concentration of 0.02% DMSO. Reactions were seeded with 2 μL of a 10⁻⁵ dilution of RML-infected (RML) or -uninfected (NBH) brain homogenate. Plates were sealed and incubated in a BMG Polarstar plate reader at 42 °C with cycles of 1 min of shaking (700 rpm double orbital) and 1 min of rest. ThT fluorescence measurements were taken every 15 min (450 ± 10 nm excitation and 480 ± 10 nm emission; bottom read; 20 flashes per well; manual gain of 2000; 20 s integration time). Data were analyzed using GraphPad Prism.

Dendritic Spine Retraction Assay.

Cultures of hippocampal neurons were performed as described previously.¹⁶ Neurons were preincubated for 2 h with compound (or DMSO vehicle for the controls), followed by treatment for 24 h with 4.4 μg/mL of purified PrP^Sc or an equivalent volume of mock-purified control sample from noninfected brains. Neurons were then fixed in 4% paraformaldehyde and stained with Alexa Fluor 488-phalloidin (Thermo Fisher) to visualize actin in dendritic spines. Images were acquired using a Zeiss 700 confocal microscope with a 63× objective using Zen software. The number of spines was normalized to the measured length of the dendritic segment to give the number of spines/μm using ImageJ software. All animal studies were approved by the Boston University Chobanian & Avedisian School of Medicine Institutional Animal Care and Use Committee and performed in accordance to the United States Department of Agriculture Animal Welfare Act and the National Institutes of Health Policy on Humane Care and Use of Laboratory Animals.

Prion Purification.

Prion purification was performed as described previously.⁵² Briefly, 200 μL of 10% (w/v) brain homogenate from an RML-inoculated C57BL6/J mouse at the terminal phase of the disease was treated with 100 μg/mL of Pronase E (Sigma) for 30 min at 37 °C with shaking in a thermal mixer at 800 rpm. EDTA was added to a final concentration of 10 mM (pH 8) before adding sarkosyl in D-PBS and Benzonase to final concentrations of 2% (w/v) and 50 U/mL, respectively. After incubation at 37 °C for 10 min, samples were brought to 0.3% (w/v) NaPTA (pH 7.4) and incubated at 37 °C for 30 min. Samples were then mixed with iodixanol to a final concentration of 35% (w/v) adding enough additional NaPTA to maintain the final concentration of 0.3% (w/v). Samples were then centrifuged at 16,100g for 90 min at room temperature. The upper layer was removed and filtered with a 0.45 μm pore size Durapore membrane Ultrafree-HV microcentrifuge filtration unit (Millipore). Samples were mixed with an equal volume of 0.3% NaPTA and 2% sarkosyl and incubated at 37 °C for 10 min before centrifugation for 90 min at 16,100g. The supernatant was discarded and the pellet was resuspended in D-PBS with 17.5% iodixanol and 0.1% sarkosyl. Following two washes, samples were resuspended in D-PBS with 0.1% (w/v) sarkosyl, pooled and stored as aliquots at –80 °C. Samples were analyzed bysilver stain (per manufacturer’s instruction; Pierce) and immunoblotting with and without PK digestion.

Safety.

There are no unexpected, new, or significant hazards or risks associated with the reported work.

Supplementary Material

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acschemneuro.4c00095.

Additional figures of uncropped western blots presented in the main figures, silver stain and western blot of purified prion preparations, and a table of complete PDSP analysis (PDF)

ABBREVIATIONS USED

BCA

bicinchoninic acid protein assay

BSA

bovine serum albumin

CJD

Creutzfeldt–Jakob disease

CNS

central nervous system

CRISPR

clustered regularly interspaced short palindromic repeats

DAPI

4′,6-diamidino-2-phenylindole

DBCA

drug-based cellular assay

DIV

days in vitro

DMSO

dimethyl sulfoxide

ECL

enhanced chemiluminescence substrate

ER

endoplasmic reticulum

F-actin

filamentous actin

HEK293T

human embryonic kidney cell line

ICE

inference of CRISPR edits

MAPK

mitogen-activated protein kinase

MTT

3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide

NaPTA

sodium phosphotungstic acid

NIMH

National Institute of Mental Health

NMDA

N-methyl-D-aspartic acid

N2a

mouse neuroblastoma cell

PBS

phosphate-buffered saline

PCR

polymerase chain reaction

PDSP

Psychoactive Drug Screening Program

PK

proteinase K

PrP

prion protein

PrP^C

cellular prion protein

PrP^Sc

scrapie-associated prion protein

PVDF

poly(vinylidene fluoride)

RML

Rocky Mountain Laboratories strain of prions

SARS-CoV-19

severe acute respiratory syndrome coronavirus 2

ScN2a

prion-infected mouse neuroblastoma cell

sgRNA

single guide RNA

TBST

tris-buffered saline with 0.1% Tween 20

σ₁R

sigma-1 receptor

σ₂R

sigma-2 receptor

22L

strain of mouse prions

Data Availability Statement

Reasonable requests to access the data used in these analyses can be made to David A. Harris.