Bidirectional manipulation of syntaxin-6 expression reveals an inverse relationship with cell-associated prion infectivity

To validate a functional role for syntaxin-6 in prion-related phenotypes in cellular models and investigate its mechanistic role, we stably knocked down Stx6 in the prion-susceptible PK1 neuroblastoma cell line [26] by ~ 85–90% in multiple independent cell lines, relative to cell lines expressing a non-silencing control (NSC) scrambled shRNA (Supplementary Fig. 1a–b in supplementary file 2). We additionally generated two independent cell lines with ~ nine-to-tenfold overexpression of syntaxin-6 (Supplementary Fig. 1d–e in supplementary file 2). To explore whether syntaxin-6 plays a role in susceptibility to prion infection, we employed the scrapie cell assay (SCA) [26, 46], which is an ELISpot based method to quantify cell-associated prion infectivity. The output of this assay is “spot count” with the detection of foci of aggregated PrP (spots detected via the ELISpot assay) differentiating infected and non-infected cells, which allows prion titres to be calculated [26].

Infection of PK1 Stx6 knockdown cell lines with the mouse-adapted scrapie prion strain, RML [8], resulted in a statistically robust increase in the spot count, which was broadly consistent across split numbers and prion dilutions (Fig. 1a-b). This suggests that syntaxin-6 knockdown increases the number of infected cells. In contrast, the spot count was abolished in cell lines with diminished PrPC levels (PK1 Prnp KD1 and KD2), the substrate for conversion, confirming the expected performance of the assay (Fig. 1b). Furthermore, when we challenged PK1 Stx6 knockdown cell lines with a limiting dose of prions, cells became susceptible to prions, in contrast to controls which were resistant to infection (Supplementary Fig. 2a–b in supplementary file 2). Strengthening these findings, infection of PK1 Stx6 overexpression cell lines resulted in a robust reduction in the spot count (Fig. 1a, c), suggesting that overexpression may enhance clearance of prion infection. Taken together, there was a strong negative gene–dosage association between Stx6 levels and susceptibility to prion infection across all of the aforementioned cell lines (Spearman r = 0.85, P = 0.0062) (Fig. 1d).

Fig. 1

Bidirectional manipulation of syntaxin-6 expression reveals an inverse relationship with cell-associated infectivity. a Matrix summarising the effects of syntaxin-6 manipulation on the spot count of infected cell number in the scrapie cell assay (SCA) in PK1 neuroblastoma cells. The results of three independent PK1 Stx6 knockdown (KD) cell lines and two PK1 independent Stx6 overexpression (OE) cell lines are shown with each column representing a different passage number (P1–6). Each row details the infection paradigm, including different dilutions of 10% (w/v) RML-infected brain homogenate (BH) or a crude infected exosome preparation (exo). Statistics are based on one-way ANOVA followed by Fisher's LSD test of the pre-planned comparison of the relevant control (NSC or NEG OE) to each cell line with Stx6 manipulation on a plate-by-plate basis. All statistics are based on the raw spot count except for PK1 Stx6 overexpression cell lines, where the spot count was normalised to the haematoxylin total cell count. Nominal differences were defined as means which surpassed the threshold of the mean spot count of the negative control ± 0.5 standard deviations. b Representative example of the spot count of infected cell number at the 5th split in the SCA following infection with 3 × 10–4 RML prions (8 technical replicates/cell line). As the assay was conducted across multiple plates, spot counts of subsequent plates were normalised to the mean spot count of the non-silencing control (NSC1) on plate 1. NSC1 was technically replicated across two plates as indicated in separate colours. c Representative example of the spot count of infected cell number at the 3rd split in the SCA following infection with an infected exosome fraction (3 × 10–3; 24 technical replicates/cell line). The spot count was normalised to haematoxylin total cell count to determine the proportion of cells infected before being normalised to the mean of the relevant NEG OE cell line. d Graph illustrating the relationship between Stx6 protein level (normalised to the relevant negative control) and the spot count in the SCA shown in the representative examples in (b) and (c) in PK1 cells. The strength of correlation was assessed by the Spearman’s rank correlation coefficient (Spearman r =  − 0.85, P = 0.0062). e, f Matrix summarising the effects of syntaxin-6 manipulation on the spot count in the SCA in CAD5 catecholaminergic cells as in (a) with the additional interrogation of infection with other mouse-adapted prion strains, including 22L, MRC2, and ME7. g–j Representative examples of the spot count of infected cell number at the 4th split in the SCA following infection with 1 × 10–5 RML prions, 1 × 10–5 22L, 6 × 10–4 MRC2, and 6 × 10–4 ME7 (6 technical replicates/cell line/strain). As the assay was conducted across multiple plates, spot counts of subsequent plates were normalised to the mean spot count of non-silencing control (NSC1) on plate 1. NSC1 was technically replicated across two plates as indicated in separate colours. Statistical differences were assessed by one-way ANOVA followed by Fisher’s LSD test on planned comparisons. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001

As we had found profoundly altered prion-related phenotypes in PK1 cells with syntaxin-6 manipulation, we additionally manipulated the expression of syntaxin-6 in a different, widely used prion-susceptible cell line, CAD5 catecholaminergic cells [40], to increase the generalisability of our findings. Therefore, 5 independent CAD5 cell lines with ~ 59%-80% knockdown of syntaxin-6 were generated (Supplementary Fig. 1g–i in supplementary file 2), as well as 3 independent cell lines with ~ 8–14-fold overexpression of syntaxin-6 (Supplementary Fig. 1k–l in supplementary file 2). Although we found no consistent effect in Stx6-overexpressing CAD5 cells (Fig. 1e), knockdown in CAD5 cells consistently increased the spot count after RML infection, corroborating the PK1 data (Fig. 1f–g). Strengthening these findings, this broadly extended to other mouse-adapted prion strains, including 22L (Fig. 1h), MRC2 (Fig. 1i), and ME7 (Fig. 1j) prions. Taken together, these results suggest that syntaxin-6 modifies susceptibility to prion infection in cellular models, with confidence strengthened by consistent effects across different paradigms, prion strains, and cell types.

To discount confounding factors underlying these results, we confirmed broadly comparable growth rates of the cell lines (Supplementary Fig. 1c, f, j, m in supplementary file 2) and determined that the differing spot counts were not reflective of differences in cell viability (Supplementary Fig. 2g–i in supplementary file 2). There were no consistent differences in total PrP levels (Supplementary Fig. 2j, l, m in supplementary file 2) with the exception of a subtle reduction in PK1 cells with Stx6 overexpression (Supplementary Fig. 2k in supplementary file 2). Although this is unlikely to be a driver of the altered spot count given the modesty of the difference and the evidence that the rate of prion propagation is not a function of PrP expression in these cells [4, 14, 30, 48], we explored the underlying driver of this potential epiphenomenon. There were no differences in Prnp mRNA levels (Supplementary Fig. 2n in supplementary file 2) or PrPC degradation kinetics (Supplementary Fig. 2o in supplementary file 2) arguing against a role for syntaxin-6 in PrPC synthesis or its intracellular degradation in cellular models, respectively. Therefore, we found no evidence for confounders underlying the SCA results.

Syntaxin-6 knockdown increases the perinuclear accumulation and alters the aggregate morphology of disease-related PrP in prion-infected cells

Due to the prominent effects on the spot count, we performed complementary phenotypic characterisation of the infected Stx6 knockdown cell lines by confocal microscopy using two antibodies, 5B2 and 6D11, which each detect a distinct disease-associated PrP staining profile in this cellular system [42]. 5B2 immunostaining recognises elongated disease-related PrP aggregates at the plasma membrane and extracellular matrix in infected cells, whereas 6D11 immunostaining is seen as punctate, perinuclear staining in addition to some plasma membrane and extracellular matrix staining. There was a statistically significant increase in the accumulation of perinuclear 6D11-positive disease-related PrP in PK1 Stx6 KD1 and KD3 (Fig. 2a–b) as well as a statistically robust reduction in 5B2-positive plasma membrane staining in all Stx6 knockdown cell lines (Fig. 2c). This demonstrates that syntaxin-6 knockdown redistributes disease-related PrP, consistent with a trafficking mechanism. Confocal imaging of infected Stx6 overexpression cells revealed a universal reduction in 6D11 and 5B2 staining (Supplementary Fig. 2c–f in supplementary file 2), suggesting that syntaxin-6 overexpression was sufficient to almost clear the infection, corroborating the SCA data.

Fig. 2

Syntaxin-6 knockdown increases the perinuclear accumulation and alters the aggregate morphology of disease-related PrP in prion-infected PK1 cells. a Representative images of three independent Stx6 knockdown cell lines and one non-silencing control (NSC1) stained with the discriminatory anti-PrP antibody pair, 5B2 and 6D11, which preferentially immunolabel disease-related PrP assessed by confocal laser-scanning microscopy. DAPI, nuclear stain, blue; 6D11, red; 5B2, green. Scale bar, 5 μm (Stx6 KD1) or 2 μm (Stx6 KD3-4, NSC1). b Quantification of the 6D11 signal area in the perinuclear space normalised by total cell count as indicated by DAPI. Statistical differences were assessed by one-way ANOVA on log-transformed data followed by Fisher’s LSD test on planned comparisons. Each dot represents an individual image (n = 47–49/cell line). One outlier (8.28) was excluded in the graph from the Stx6 KD1 group for visual clarity but was included in the statistical analysis of log-normal data. c Quantification of the total 5B2 signal area coverage at mid-cell level normalised by total cell count as indicated by DAPI. Each dot represents an individual image (n = 47–49/cell line). Statistical differences were assessed by one-way ANOVA on log-transformed data followed by Fisher’s LSD test on planned comparisons. d Representative images of iS7 cells with stable Stx6 knockdown, or control iS7 cells expressing an NSC shRNA construct, co-stained with two anti-PrP antibodies (6D11, red and 5B2, green) as well as DAPI nuclear stain (blue) at mid-cell level. Scale bar, 5 µm. e Quantification of 6D11-positive area of disease-related PrP in the perinuclear space with the line representing mean ± SEM and individual dots representing an individual cell (19–21 cells/cell line). Statistical differences were assessed by one-way ANOVA followed by Fisher’s LSD test on planned comparisons. f Quantification of the median length of elongated disease-related PrP aggregates derived from maximum intensity projections at the plasma membrane and extracellular matrix level (9–10 images/cell line). Statistical differences were assessed on log-normal data using one-way ANOVA to test for the effect of Stx6 expression level on length. g High magnification images of elongated disease-related PrP aggregates at the plasma membrane and extracellular matrix level displayed as maximal intensity projections of z-stack images to capture the length across all planes of the aggregate. Scale bar, 5 µm. For each experiment, brightness/contrast was adjusted similarly across conditions. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0. 0001

To more thoroughly explore phenotypic differences in the distribution and aggregate morphology of disease-related PrP, we additionally stably knocked down syntaxin-6 in chronically infected PK1 cells (a subclone called iS7) [42] allowing us to explore a role for syntaxin-6 in prion accumulation isolated from infection (Supplementary Fig. 3a–b in supplementary file 2). Recapitulating our observations with freshly infected cells (Fig. 2a-b), we found a prominent accumulation of perinuclear 6D11-positive disease-related PrP with Stx6 knockdown in chronically infected cells (Fig. 2d–e). We also observed longer 5B2-positive elongated disease-related PrP aggregates at the plasma membrane and extracellular matrix with Stx6 knockdown (P = 0.0062) (Fig. 2f–g) but not total load (Supplementary Fig. 3i in supplementary file 2). Taken together, these results show that syntaxin-6 knockdown causes cellular redistribution and structural reorganisation of disease-related PrP in chronically infected cells, further supporting a role for syntaxin-6 in trafficking of disease-related PrP.

Syntaxin-6 promotes prion export in chronically infected cells

To test whether a role for syntaxin-6 in prion export could explain the altered spot counts, we transiently knocked down syntaxin-6 in chronically infected iS7 PK1 cells followed by collecting conditioned media to infect PK1 reporter cells (Fig. 3a). We observed a reduction in secreted infectivity after syntaxin-6 knockdown, measured by the SCA (18.2% ± 3.7% reduction, mean ± SEM) (Fig. 3b), suggesting that syntaxin-6 mediates prion export. Corroborating this, when we harvested media from the infected stable Stx6 knockdown and overexpression PK1 cell lines described previously, syntaxin-6 knockdown resulted in reduced relative secreted infectivity titres, with the converse being observed with syntaxin-6 overexpression, following correction for baseline differences in cell-associated infectivity (Supplementary Fig. 2p, q in supplementary file 2).

Fig. 3

Syntaxin-6 promotes prion export in chronically infected PK1 cells. a Experimental design for assessing a modifying effect of syntaxin-6 knockdown on secreted infectivity from chronically infected cells (iS7 subclone). b The proportion of infected reporter PK1 cells (spot count normalised to haematoxylin total cell count) following the application of conditioned media harvested from iS7 cells in which Stx6 had been transiently knocked down or a non-silencing control (NSC) shRNA had been employed. c Representative images of iS7 cells co-labelled with 6D11 and markers of intracellular compartments to assess for altered distribution of disease-related PrP with syntaxin-6 knockdown in the early endosome (EEA1), lysosomes (LAMP1) and in the trans-Golgi (TGN46). Scale bar, 2 µm. d–f Assessment of levels of colocalisation between 6D11-positive PrP and organelle markers, expressed as Pearson’s correlation coefficient for EEA1 (n = 19–21 cells/cell line), LAMP1 (n = 20–27 cells/cell line) and TGN46 (n = 20–22 cells/cell line). Line represents mean ± SEM with one-way ANOVA followed by Fisher’s LSD test being used to test statistical differences. For each experiment, brightness/contrast was adjusted similarly across conditions. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0. 0001

In further support for a role of syntaxin-6 in prion export, as opposed to involvement in an intracellular trafficking step, we observed a generalised increase in the degree of colocalisation of 6D11-positive disease-related PrP with markers of intracellular organelles (Fig. 3c–f). Syntaxin-6 knockdown resulted in increased colocalisation of 6D11 with the early endosome marker, EEA1, (Fig. 3d) and the lysosome marker, LAMP1, (Fig. 3e) with some additional evidence for the TGN marker, TGN46 (Fig. 3f).

Collectively, this provides evidence for prion export being the syntaxin-6-driven molecular susceptibility mechanism.

Syntaxin-6 modifies prion pathogenesis in vivo by modulating the risk of disease development

As we had established a role for syntaxin-6 in modulating prion-related phenotypes in cellular models, we next wanted to validate a functional role for syntaxin-6 in prion pathogenesis in vivo and explore the disease stage at which it acts. The “gold standard” paradigm for studying prion disease pathogenesis is prion transmission in mice. Mice are naturally susceptible to prion infection, developing bona fide disease with faithful recapitulation of the clinical and neuropathological hallmarks of human disease when experimentally inoculated with prions such as the mouse-adapted scrapie prion strain, RML [8]. To determine whether Stx6 knockout modified the risk of mice developing prion disease, we intracerebrally infected Stx6+/+ and Stx6−/− mice (n = 90/genotype) with a tenfold serial dilution series of 10% (w/v) RML prion-infected brain homogenate and assessed disease development using both clinical and neuropathological diagnoses. The focus of this study was to examine prion doses with a partial attack rate (defined as < 90% across both arms of the study: 10–5, 10–6, 10–7, 10–8), where the likelihood of disease development was uncertain, providing a paradigm to assess whether syntaxin-6 modulated that risk.

For each of the 10–5, 10–6, 10–7, and 10–8 concentrations, the proportion of prion disease cases relative to the animals surviving to the end of the study (attack rate, see methods) was consistently lower in Stx6−/− mice relative to Stx6+/+ mice, suggesting that they were more resistant to disease development (Table 1; Fig. 4). Logistic regression analysis demonstrated a significant effect of dose (P < 0.0001) and genotype (P = 0.05) on prion disease diagnosis, with infected Stx6+/+ mice having 2.19 [95% CI 1.01–4.56] times higher odds of developing prion disease compared to Stx6−/− animals at concentrations 10–5 and lower. This suggests syntaxin-6 knockout reduces susceptibility to prion infection, which is further supported by calculations estimating the “effective” dose administered to Stx6+/+ and Stx6−/− mice, which was reduced in Stx6−/− mice by a log order of magnitude relative to Stx6+/+ mice (105.82 LD50/mL and 106.61 LD50/mL, respectively) by the Spearman–Karber method or ~ 0.5 log difference with the Reed and Muench method.

Table 1 Assessment of susceptibility differences in Stx6+/+ and Stx6−/− mice inoculated with RML prion dilutions with a partial attack rateFig. 4

Syntaxin-6 knockout renders mice less susceptible to prion disease when administered low doses of prions. a Experimental design of intracerebral inoculation of Stx6+/+ and Stx6−/− mice with a tenfold dilution series of 10% (w/v) RML prion-infected brain homogenate (n = 15/genotype/dose) with prion doses with a partial attack rate (< 90%) being of interest. Created in BioRender. One, S. (2025). https://BioRender.com/t36g169. b Bar chart showing the attack rates of prion disease in mice administered 10–5, 10–6, 10–7, or 10–8 concentrations of RML prions

Therefore, in addition to a role for syntaxin-6 in altering prion-related phenotypes in cellular models, this established a role for syntaxin-6 in modulating early stages of prion pathogenesis in vivo.

Syntaxin-6 does not alter prion propagation or prion-induced neurotoxicity in vivo during established disease

This modifying effect of syntaxin-6 on disease susceptibility in vivo could either be acting through directly modulating the establishment of prion infection, or alternatively, by altering subsequent prion propagation or prion-induced neurotoxicity. To systematically interrogate a role for syntaxin-6 in prion replication and neurotoxicity, we infected Stx6+/+ and Stx6−/− mice (n = 110/genotype) with 1% (w/v) RML prion-infected brain homogenate, with animals subsequently being culled at multiple predefined time points or at the onset of clinical disease (Fig. 5a). This allowed age-matched, cross-sectional analyses of prion-related phenotypes in the evolving stages of disease, with timed culls of PBS-inoculated mice providing the negative control (Fig. 5b).

Fig. 5

Syntaxin-6 knockout has no effect on prion propagation kinetics or levels of disease-related PrP in mice infected with RML prions. a Experimental design mapped onto the two-phase kinetics model whereby RML-infected Stx6+/+ and Stx6−/− mice were culled at predefined time points to assess for differences in prion propagation and neurotoxicity-related outcome measures. b Timed culls of PBS-inoculated controls. c Prion titres [log tissue-culture infectious units (TCIU) per gram brain] of RML-infected Stx6+/+ and Stx6−/− mice (n = 5–10/genotype/time point), across the incubation period (dpi, days post-inoculation). All PBS-inoculated brain homogenates were negative for infectivity (not shown). Curves were fitted using the logistic growth model [goodness of fit, r2: 0.928 (Stx6+/+), 0.926 (Stx6−/−)]. Bars indicate mean ± SEM with dotted lines representing 95% confidence intervals. d 10% (w/v) brain homogenates from RML-infected Stx6+/+ and Stx6−/− mice at 140 dpi were analysed by immunoblotting with the anti-PrP antibody, ICSM35, after digestion with proteinase K (50 µg/mL, 37 °C, 1 h). Semi-quantification of the PrPSc signal was performed using densitometry with each sample being normalised to the average of RML-infected Stx6+/+ mice. Bar graphs represent mean ± SEM of 4 biological replicates/genotype. C, positive control RML sample. Statistical differences were tested with a Student’s t test (P = 0.693). The image brightness/contrast was optimally adjusted. e Spatiotemporal differences of PrP deposition were assessed by immunohistochemistry using the anti-PrP antibody, ICSM35, at 70, 90, and 140 days post-inoculation (dpi) (n = 9–10/genotype). A schematic is shown to represent the overall staining pattern in the respective groups (pale pink: mild PrP deposition; pink shading: moderate PrP deposition; red shading: intense PrP deposition). Representative images of the whole brain section as well as magnified images from the cortex, hippocampus, and brainstem are shown. Scale bar, 2.5 mm (overview) and 0.5 mm (zoom). Numbers shown next to the schematics report the number of animals in the group positive for the pathology shown as a fraction of the total number of animals in each group

The automated SCA (ASCA) [26, 46] was used to provide measurement of prion titres throughout the disease course (Fig. 5c). Prion titres in both RML-infected Stx6+/+ and Stx6−/− mice increased rapidly at comparable rates before reaching a similar maximal prion titre of ~ 108.5 infectious units/g at ~ 90 days post-inoculation (dpi), in line with the two-phase kinetics model [43, 44]. This suggests that syntaxin-6 does not alter prion propagation kinetics in established disease. This was further supported by biochemical assessment of disease-related proteinase K (PK)-resistant PrP (PrPSc) at 140 dpi, which was comparable in infected Stx6+/+ and Stx6−/− mice, with indistinguishable electrophoretic mobility and glycosylation patterns (Fig. 5d). This was also corroborated by immunohistochemical detection of disease-related PrP, where we found no differences in the appearance, extent and distribution of PrP deposits with both the onset and evolution of deposition being broadly comparable in RML-infected Stx6+/+ and Stx6−/− mice (Fig. 5e). Taken together, these results suggest that syntaxin-6 is not involved in prion propagation nor in modulating the levels of disease-related PrP during established disease.

Subsequently, we explored whether there were any differences in the onset and/or progression of markers of neurotoxicity or neurodegeneration in RML-infected Stx6+/+ and Stx6−/− mice. There were no differences in the extent or distribution of intraneuronal vacuoles (“spongiosis”) (Supplementary Fig. 4a–b in supplementary file 2), synaptic integrity (Supplementary Fig. 4c in supplementary file 2), or the spatiotemporal evolution of astrogliosis and microgliosis (Supplementary Fig. 4d–e in supplementary file 2) across the disease course. Furthermore, we observed comparable levels of disease-associated PK-sensitive PrP species (Supplementary Fig. 5a–b in supplementary file 2) and toxicity levels in prion-infected brain homogenates using a validated neurotoxicity assay [3] (Supplementary Fig. 5c–d in supplementary file 2). There were also no consistent differences in serum neurofilament light-chain (NfL) levels (Supplementary Fig. 5e in supplementary file 2), which is a sentinel neurodegeneration biomarker in prion disease [33]. Finally, RML-infected Stx6+/+ and Stx6−/− mice exhibited comparable neurological phenotypes, time to first symptom (Stx6−/− median [95% confidence interval] = 129 days [126–129] vs. Stx6+/+  = 129 days [126–131]), incubation times (Stx6−/− median [95% confidence interval] = 139 days [134–141] vs. Stx6+/+  = 140.5 days [139–144]), and clinical progression (Supplementary Fig. 5f–h in supplementary file 2). These results suggest that syntaxin-6 is not involved in prion-induced neurotoxicity.

Brain transcriptomic analyses of Stx6+/+ and Stx6−/− mice suggested that compensatory mechanisms were not at play at the RNA level with there being no significant upregulation of genes encoding other syntaxins/trafficking proteins (Supplementary Table 1 in supplementary file 3). Therefore, taken together with the positive results of the titration study, these findings suggest syntaxin-6 modifies the establishment of disease, with no discernible effect on prion propagation nor prion-induced toxicity in established disease.