Accelerated Prion Disease Pathogenesis in Toll-Like Receptor 4 Signaling-Mutant Mice

Abstract

Prion diseases such as scrapie involve the accumulation of disease-specific prion protein, PrP^Sc, in the brain. Toll-like receptors (TLRs) are a family of proteins that recognize microbial constituents and are central players in host innate immune responses. The TLR9 agonist unmethylated CpG DNA was shown to prolong the scrapie incubation period in mice, suggesting that innate immune activation interferes with prion disease progression. Thus, it was predicted that ablation of TLR signaling would result in accelerated pathogenesis. C3H/HeJ (Tlr4^Lps-d) mice, which possess a mutation in the TLR4 intracellular domain preventing TLR4 signaling, and strain-matched wild-type control (C3H/HeOuJ) mice were infected intracerebrally or intraperitoneally with various doses of scrapie inoculum. Incubation periods were significantly shortened in C3H/HeJ compared with C3H/HeOuJ mice, regardless of the route of infection or dose administered. At the clinical phase of disease, brain PrP^Sc levels in the two strains of mice showed no significant differences by Western blotting. In addition, compared with macrophages from C3H/HeOuJ mice, those from C3H/HeJ mice were unresponsive to fibrillogenic PrP peptides (PrP residues 106 to 126 [PrP_106-126] and PrP_118-135) and the TLR4 agonist lipopolysaccharide but not to the TLR2 agonist zymosan, as measured by cytokine production. These data confirm that innate immune activation via TLR signaling interferes with scrapie infection. Furthermore, the results also suggest that the scrapie pathogen, or a component(s) thereof, is capable of stimulating an innate immune response that is active in the central nervous system, since C3H/HeJ mice, which lack the response, exhibit shortened incubation periods following both intraperitoneal and intracerebral infections.

Prion diseases are a family of disorders that affect a variety of mammalian species in such forms as scrapie, bovine spongiform encephalopathy, chronic wasting disease, and Creutzfeldt-Jakob disease (7). To develop experimental prion disease models for research, naturally occurring sheep scrapie was adapted first to mice and then to hamsters a number of years ago (14, 15); scrapie strains such as ME7, 139A, 263K, and others were derived by this method.

Prion diseases are characterized by the deposition of the abnormal prion protein PrP^Sc, which is derived from conversion of the normal cellular prion protein, PrP^C. In prion diseases that occur following a peripheral infection, PrP^Sc is first deposited peripherally in lymphoreticular tissues and, via peripheral nerves, moves centrally into the spinal cord and, ultimately, to clinical target areas in the brain (7). Since PrP^Sc is resistant to proteolytic degradation, its clearance from the body is slow (53). The clearance of PrP^Sc that does occur probably results from innate immune cell activation, particularly that of phagocytes such as macrophages (MΦs), microglia, and dendritic cells (DCs) capable of ingesting and degrading PrP^Sc, which are mobilized during the course of prion infection (8, 12, 35, 38, 41, 50).

Signaling through members of the Toll-like receptor (TLR) family is a central process in host innate immune response. The TLRs function as sentinels that recognize and trigger responses to microbial constituents that include such molecules as bacterial lipopolysaccharide ([LPS]; TLR4) and unmethylated CpG-containing DNA (TLR9), among others (60). Such responses are immediate and promote the mobilization of phagocytes to the infected site (31).

It was previously shown that the TLR9 agonist, unmethylated CpG DNA, slows scrapie pathogenesis when administered to infected mice (55, 56), suggesting that innate immune activation interferes with prion infection. However, MyD88, an obligatory signaling intermediate for several TLR family members including TLR9 was shown not to play a role in scrapie as infection in MyD88^−/− and wild-type (WT) mice exhibits similar pathogenesis (i.e., similar incubation periods, infection kinetics, and neuropathological profiles) (51). Since TLRs, such as TLR4, can utilize both MyD88-dependent and -independent signaling pathways (60, 67), we reasoned that associations between TLR signaling and scrapie pathogenesis might be revealed in mice ablated at the TLR itself. C3H/HeJ (Tlr4^Lps-d) mice, which possess a mutation in the TLR4 intracellular signaling domain preventing TLR4 signaling (29, 49), and strain-matched WT control (C3H/HeOuJ) (2) mice were infected intracerebrally (i.c.) or intraperitoneally (i.p.) with various doses of the mouse-adapted 139A or ME7 scrapie agent. As assessed by onset of clinical scrapie symptomatology, scrapie pathogenesis was significantly accelerated in C3H/HeJ versus C3H/HeOuJ mice regardless of the scrapie strain, the route of infection, or the dose administered. Our results indicate that innate immune activation inhibits prion pathogenesis and that stimulation of TLR4 may play a role in suppressing scrapie infection. The results also suggest for the first time that the scrapie pathogen, or a component(s) thereof, stimulates an innate immune response through TLR4 that is capable of combating the disease. In addition, the defensive response triggered through TLR4 signaling appears to be operative in the central nervous system (CNS) since, as measured by scrapie incubation period, it functions with similar potency following both i.p. and i.c. infections.

MATERIALS AND METHODS

Mice, scrapie, and brain histology.

All animal experimentation was formally approved by the New York State Institute for Basic Research in Developmental Disabilities Institutional Animal Care and Use Committee prior to initiation. Female mice (The Jackson Laboratory, Bar Harbor, ME) were inoculated by injection with scrapie or control inoculum (25 μl in phosphate-buffered saline i.c. or i.p.) at 5 weeks of age. Inocula were dilutions of 10% (wt/vol) brain homogenates from clinical 139A or ME7 scrapie or normal control female C57BL/6J mice. Inoculum dosages are expressed as dilutions of brain (e.g., 10% brain homogenate is a dose of 10⁻¹, or a 10-fold dilution of brain), with 1 g of clinical scrapie brain containing ∼10⁸ infectious doses, as previously determined by titration in i.c. infection (52, 62). During the described experiments, all scrapie inocula were derived from identical stock. Starting 3 months after inoculation, mice were monitored weekly for symptomatology as described previously (33). The incubation period was calculated as the time period (in days) between scrapie inoculation and the first sign of clinical symptoms; clinically positive mice were monitored for an additional 2 weeks (making a total of three positive scores) to confirm the diagnosis. Mice were then sacrificed by lethal pentobarbital (Nembutal) anesthesia, and brains were dissected, and one hemisphere was immersion-fixed in 4% neutral-buffered paraformaldehyde for neuropathological examination and the other was frozen for Western blot analyses. Fixed brain tissue was paraffin embedded, cut into 7-μm-thick coronal sections, and stained with hematoxylin and eosin.

Western blot analyses.

Brains were homogenized to 10% (wt/vol) in either 1% sarcosyl-Tris-buffered saline (TBS-S buffer) or 0.5% NP-40-0.5% sodium deoxycholate-100 mM NaCl-10 mM EDTA-10 mM Tris-Cl (pH 7.5; buffer H) and, where indicated, treated with 50 μg/ml proteinase K (PK) at 37°C for 1 h, followed by treatment with a cocktail of protease inhibitors. Sodium dodecyl sulfate-polyacrylamide gel electrophoresis, blotting, and staining were performed as previously described (56). Primary probes for PrP were either monoclonal antibody (MAb) 6D11 (56) or MAb 4C4 (32) at 0.1 μg/ml. Anti-β-actin control MAb (catalog no. A3853; Sigma, St. Louis, MO) was used at 0.8 μg/ml. Molecular masses of protein bands were indicated with either Kaleidoscope (Bio-Rad, Hercules, CA) or BlueRanger (Pierce, Rockford, IL) prestained standards. Densitometric comparisons of PrP bands were done using Un-Scan-It gel software (Silk Scientific; Orem, UT). For analysis of PrP^Sc levels, amounts of PrP^Sc in homogenate samples on Western blots were found to be in a superlinear range of response by densitometric analysis; thus, small changes in PrP^Sc present in the sample would necessarily lead to large changes in densitometric measurements. This allowed certainty of whether PrP^Sc levels were closely comparable among brain samples.

Prnp genotype determination.

Genomic DNA isolated from the tail tissue of mice was amplified by PCR using forward and reverse primers for the Prnp coding region within exon 3 (SenseB3, 5′-TCA TGG CGA ACC TTG GCT AC-3′; AntisenseB4, 5′-CCA CGA GAA TGC GAA GGA AC-3′) with 1.5 mM MgCl₂; the PCR program consisted of 94°C for 2 min, with 35 cycles of 94°C for 30 s, 67°C for 30 s, and 72°C for 1 min, followed by 72°C for 10 min. Amplified products were sequenced by Genemed Synthesis (San Antonio, TX) in the forward and reverse direction with primers S1 (5′-CCT TGG TGG CTA CAT-3′) and A2 (5′-TGG CCT GTA GTA CAC-3′).

MΦ isolation and cytokine production.

Female mice, 8 to 12 weeks old (n = 3 per mouse strain for each experiment) were injected i.p. with 1 ml of 4% thioglycolate; 4 days later MΦs were collected by peritoneal lavage at greater than 98% purity, with those from each mouse strain pooled separately, and cultured as previously described (34). After cells were incubated overnight, the medium was replaced with medium containing either a peptide consisting of human PrP residues 106 to 126 (PrP_106-126; 40 μM), PrP_106-126 scrambled (scrPrP_106-126; 40 μM), or PrP_118-135 (80 μM) (Bachem, King of Prussia, PA); LPS (Escherichia coli strain 055:B5; 1 ng/ml); zymosan (10 μg/ml), or no added reagents and incubated for an additional 24 h. Prior to addition to MΦ cultures, all PrP peptides (at 1.0 mM) had been aged at ambient temperature for 72 h in RPMI base medium to generate fibrils (PrP_106-126 and PrP_118-135 only), according to previously described methods (23, 47). Endotoxin was not detected in any of the above reagents (except for LPS) using a Limulus amebocyte lysate assay (Sigma). Interleukin-6 (IL-6) and tumor necrosis factor alpha (TNF-α) secreted into culture medium were measured using corresponding Duoset enzyme-linked immunosorbent assay kits (R&D Systems, Minneapolis, MN), and concentrations of IL-6 and TNF-α produced by MΦs were derived by comparison with corresponding recombinant murine standards.

Statistical analysis.

Statistical significance of differences between groups was determined using the Statistica software package (StatSoft, Tulsa, OK). Differences between incubation periods and between densitometric quantities of brain PrP were determined by Student's t test analysis. For MΦ cytokine secretion, differences were evaluated by post hoc Tukey honestly significant difference (HSD) analysis. All differences were considered significant at a P value of ≤0.05.

RESULTS

Scrapie incubation period is accelerated in TLR4 signaling mutant mice.

TLR4 signaling-mutant C3H/HeJ and WT (C3H/HeOuJ) mice were injected with a range of doses of 139A or ME7 scrapie brain inoculum by either the i.c. or i.p. route. Mice were then monitored for scrapie symptomatology. All mice injected with scrapie brain inoculum developed clinical disease symptoms, with all inoculations leading to significantly shorter incubation periods in C3H/HeJ than in C3H/HeOuJ mice, regardless of the route of inoculation (Table 1). Scrapie incubation periods in groups of C3H/HeJ mice were accelerated by 9.7 to 14.5% (14.5 to 28.0 days) versus their WT counterparts (mean acceleration over all agent strain-inoculation route dilutions, 11.4%). Remarkably, incubation periods in C3H/HeJ mice were similar to those in C3H/HeOuJ mice infected with a more than 10-fold higher titer of scrapie agent (Table 1).

TABLE 1.

Incubation periods of scrapie-infected C3H/HeJ and C3H/HeOuJ mice

Strain and infection route	Dosea	log10 infectious U/dosea	Incubation period characteristicsb				P valuee
			Length ± SEM (days)		Difference in periods (days)c	Ratio of periodsd
			C3H/HeJ (n)	C3H/HeOuJ (n)
139A
i.c.	10−2	5.4	121.7 ± 1.9 (15)	137.1 ± 2.1 (20)	15.4	0.888	<0.0001
	10−3	4.4	128.3 ± 1.4 (13)	145.3 ± 3.0 (12)	17.0	0.883	<0.001
	10−4	3.4	144.9 ± 2.7 (11)	165.5 ± 3.5 (14)	20.6	0.876	<0.001
	10−5	2.4	169.6 ± 4.2 (16)	191.6 ± 3.4 (17)	22.0	0.885	<0.001
i.p.	10−1	6.4	155.0 ± 2.0 (22)	176.6 ± 1.7 (22)	21.6	0.878	<0.0001
	10−2	5.4	164.5 ± 1.6 (14)	181.5 ± 2.7 (14)	17.0	0.906	<0.0001
	10−3	4.4	180.4 ± 2.9 (17)	208.4 ± 6.6 (17)	28.0	0.866	<0.001
ME7
i.c.	10−2	5.4	134.8 ± 2.7 (13)	149.3 ± 2.8 (13)	14.5	0.903	<0.01
i.p.	10−1	6.4	156.8 ± 3.7 (13)	183.5 ± 1.9 (14)	26.7	0.855	<0.0001

^aDose of the inoculum was calculated per homogenate dilution.

^bIncubation periods for 139A i.c. 10⁻² and 10⁻⁵ doses and i.p. 10⁻¹ and 10⁻³ inoculum doses are based on pooled data from two separate experiments. n, number of mice.

^cCalculated as the number of days for C3H/HeOuJ minus the number of days for C3H/HeJ mice.

^dRatio of the number of days for C3H/HeJ mice to the number of days for C3H/HeOuJ mice.

^eDetermined by Student's t test analysis.

Mice that developed clinical signs of scrapie were found to have levels of PrP^Sc in the brain detectable by Western blotting. Brains were also analyzed for the presence of vacuolation, which was found only in the mice injected with scrapie (Fig. 1). Collectively, these results indicate that scrapie pathogenesis occurs more rapidly in mice mutant for TLR4 signaling and furthermore suggest that TLR4 signaling slows the progression of scrapie pathogenesis. In addition, since the scrapie incubation period is prolonged equally following infections through both the i.p. and i.c. routes (based on the ratio of incubation periods of C3H/HeJ to C3H/HeOuJ mice) (Table 1), the mechanism(s) involved in the delay of clinical scrapie onset is likely active in the CNS.

FIG. 1.

Brain sections from clinical scrapie-infected C3H/HeJ and C3H/HeOuJ mice show similar levels of vacuolation. Coronal sections stained with hematoxylin and eosin were photographed at a magnification of ×10 from mice inoculated i.c. with either 139A or normal mouse brain homogenate at a 10² dilution. Arrowheads mark the positions of select vacuoles indicating spongiform degeneration. Note the absence of spongiform change in control sections. SN, substantia nigra; TH, thalamus. Contrast in all images was adjusted equally to clearly visualize the presence or absence of pathology.

Investigation of mechanism(s) of accelerated scrapie pathogenesis.

A number of mechanisms can be envisaged that explain the phenomenon of the accelerated scrapie incubation period in C3H/HeJ versus C3H/HeOuJ mice. Possible mechanisms for accelerated scrapie pathogenesis in C3H/HeJ mice are that ablation of TLR4 signaling either reduces the threshold for neurological response to PrP^Sc or accelerates PrP^Sc deposition (by increasing PrP^Sc accumulation and/or decreasing PrP^Sc clearance). To rule out contributions from effects unrelated to impaired TLR4 signaling, we first investigated possible differences, described below, between C3H/HeJ and C3H/HeOuJ mice in endogenous PrP^C expression or expressed PrP amino acid sequence that might lead to differences in the scrapie incubation period.

(i) Differences in PrP^C expression or sequence.

Overexpression of PrP^C in the brain is known to accelerate the scrapie incubation period in transgenic murine models (11, 37). To compare brain levels of PrP^C in C3H/HeJ and C3H/HeOuJ mice, Western blot analyses were performed on brain homogenates and analyzed by semiquantitative densitometry (Fig. 2). These analyses failed to identify a significant difference in brain PrP^C expression between C3H/HeJ and C3H/HeOuJ mice; in fact, although the difference was not significant, there was a trend toward increased PrP^C in C3H/HeOuJ mice, which display the longer scrapie incubation period. Therefore, the acceleration of scrapie pathogenesis in C3H/HeJ mice is unlikely to be attributable to increased levels of PrP^C in these mice.

FIG. 2.

PrP^C is expressed at similar levels in the brains of C3H/HeJ and C3H/HeOuJ mice. (A) Western blotting analysis of brain homogenates (prepared in buffer H) from naïve mice using anti-PrP MAb 4C4 and anti-β-actin for comparison. Each lane was loaded with 50 μg of protein. (B) Densitometric analysis comparing levels of PrP^C (normalized to corresponding endogenous β-actin level) in samples from panel A (n = 3 per mouse strain). Values are expressed as means ± standard deviations. The difference between values for C3H/HeJ and C3H/HeOuJ mice is not statistically significant (P = 0.14).

Previous analysis of highly variable mouse strain-specific single nucleotide polymorphisms showed no differences between C3H/HeJ and C3H/HeOuJ mice (45). This indicates the high degree of relatedness of these two mouse lines, despite the TLR4 mutation carried in C3H/HeJ mice. Single nucleotide polymorphisms and mutations within the PrP sequence are known to significantly affect prion disease incubation period in mice (68) and other mammals. To determine if any such differences in PrP sequence exist between C3H/HeJ and C3H/HeOuJ mice that could explain the differences in the scrapie incubation period, the nucleotide sequence of the Prnp coding region within exon 3 was analyzed in both mouse strains (n = 3 of each strain). The Prnp sequence was found to be the same in C3H/HeJ and C3H/HeOuJ mice (data not shown); the deduced PrP amino acid sequence was identical to that of the previously described version of PrP (with Leu at residue 108 and Thr at 189) with a short incubation time (the Sinc s7s7 gene product) (68). Thus, the accelerated scrapie pathogenesis observed in C3H/HeJ compared to C3H/HeOuJ mice cannot be attributed to differences in PrP primary structure.

(ii) Differences in PrP^Sc.

In light of the above results, the explanation(s) for the accelerated incubation period in the TLR4-mutant mice most likely relates to one or more of the other possible mechanisms described earlier. To investigate these possibilities, levels of disease-associated PK-resistant PrP (i.e., PrP^Sc) in the brains of clinical mice infected with scrapie via the i.p. and i.c. routes (inoculum dosages of 10⁻¹ and 10⁻², respectively) were analyzed by Western blotting (Fig. 3A). Mice inoculated with normal brain homogenate had no detectable levels of PrP^Sc in the brain (n = 6 per group) (data not shown). For the scrapie-infected mice, PrP^Sc levels detected in Western blots were analyzed semiquantitatively by densitometry (Fig. 3B). Levels of PrP^Sc detected in scrapie-infected C3H/HeJ and C3H/HeOuJ mice were not significantly different from each other for either route of inoculation, regardless of whether the inoculum was 139A or ME7. The finding of generally similar levels of PrP^Sc in brains from clinical C3H/HeJ and C3H/HeOuJ mice suggests that PrP^Sc accumulation may be accelerated in C3H/HeJ mice.

FIG. 3.

PrP^Sc levels are similar in brains from C3H/HeJ and C3H/HeOuJ mice identically infected with scrapie. (A) Western blotting analysis using anti-PrP MAb 6D11 on PK-treated brain homogenates (prepared in TBS-S buffer) from clinically positive mice infected either i.c. (dose, 10⁻²) or i.p. (dose, 10⁻¹) with 139A or ME7 scrapie brain inoculum. Each lane represents 7.1 μg of protein (measured prior to PK treatment) from different mice. ME7 i.p. gels for C3H/HeJ and C3H/HeOuJ mice are identically image gamma-enhanced for visualization only. (B) Densitometric analysis comparing levels of PrP^Sc in samples from mice shown in panel A. Values are expressed as means ± standard deviations. Differences in PrP^Sc levels between C3H/HeJ and C3H/HeOuJ mouse groups are not statistically significant in either the i.c. or i.p. comparisons (P > 0.05).

(iii) Differences in immune cell response.

Immune cells, particularly MΦs and DCs, have been reported to play a protective role during scrapie infection by clearing PrP^Sc/scrapie agent (8, 12, 35, 41), thereby reducing the infectious load. We thus analyzed whether MΦs from C3H/HeJ and C3H/HeOuJ mice respond equally to the PrP^Sc-mimetic peptides PrP_106-126 (24) and PrP_118-135 (13). Peritoneal MΦs were isolated and exposed to aged PrP peptides for 24 h; levels of TNF-α and IL-6 output were then measured. MΦs from C3H/HeOuJ mice responded to the PrP peptides by secreting moderate levels of cytokines (compared with those treated with LPS), while scrPrP_106-126 failed to induce cytokine secretion (Table 2). In contrast, PrP_106-126 and PrP_118-135 elicited no cytokine secretion in MΦs from C3H/HeJ mice. To confirm that C3H/HeJ mice lack and that C3H/HeOuJ mice possess functional TLR4 signaling, MΦs from the mice were treated with a high concentration of TLR4 agonist LPS for 24 h, and cytokine production was measured. While C3H/HeOuJ-derived MΦs responded with robust cytokine production, those from C3H/HeJ mice failed to respond to this stimulus (Table 2). To test whether loss of PrP peptide-induced cytokine production in C3H/HeJ MΦs is attributable exclusively to ablated TLR4 signaling, the capacity of C3H/HeJ MΦs for signaling through other TLR family members was probed with the TLR2-specific agonist zymosan. As shown in Table 2, both C3H/HeJ and C3H/HeOuJ MΦs responded to TLR2 stimulation, indicating functionality of signaling pathways initiated through other TLRs. Taken together, these data suggest that acceleration of pathogenesis in TLR4-signaling mutant mice relative to WT may be, at least partially, attributable to a decreased response by immune cells such as MΦs and microglia to the scrapie pathogen PrP^Sc.

TABLE 2.

Cytokine production by PrP peptide-challenged peritoneal macrophages from C3H/HeJ and C3H/HeOuJ mice

Stimulus	Cytokine production for the indicated group (pg/ml±SD)a
	IL-6		TNF-α
	C3H/HeOuJ	C3H/HeJ	C3H/HeOuJ	C3H/HeJ
Control	81.5 ± 2.9	<20	80.3 ± 5.5	<30
LPS (1 ng/ml)	1,120.0 ± 150b	<20	675.2 ± 58.0b	<30
PrP106−126 (40 μM)c	252.0 ± 43.7b	<20	215.4 ± 25.1b	<30
scrPrP106−126 (40 μM)	132.4 ± 18.2d	<20	97.8 ± 39.4d	<30
PrP118−135 (80 μM)	240.2 ± 20.2b	<20	181.7 ± 28.5b	<30
Zymosan (50 μg/ml)	6,040.8 ± 433.1b	1,175.4 ± 87.3	2,929.3 ± 204.8b	2,794.7 ± 62.9

^aValues are the means from samples measured in triplicate (data shown are representative of four independent experiments).

^bSignificantly different from the corresponding C3H/HeOuJ untreated control and scrPrP-treated MΦs (P < 0.05, Tukey HSD).

^cAll reagents (except for LPS) tested endotoxin negative by Limulus amebocyte lysate assay.

^dNot significantly different from the corresponding C3H/HeOuJ untreated control (P > 0.05, Tukey HSD) but significantly different from all other treated MΦs (P > 0.05, Tukey HSD).

DISCUSSION

The C3H/HeJ mouse strain has been studied extensively for its hyporesponsiveness to LPS at both the cellular and physiological levels (25, 26). The phenomenon of LPS hyporesponsiveness was shown to result from a TLR4 signaling defect (29, 49). TLR4 signaling confers complete or partial host protection against a variety of pathogens, including different classes of bacteria and mono- and multicellular parasites (1, 34, 66). The evolutionary importance of TLR4 functionality has recently been highlighted by the suggestion that TLR4 polymorphisms shaped genetic variation and function of the immune system of modern humans (21).

In the current study we investigated whether the effect of TLR4 signaling on microbial host defense also extends to the prion pathogen. We have shown that C3H/HeJ mice develop clinical scrapie at an accelerated rate following infection relative to their WT (C3H/HeOuJ) counterparts, which have intact TLR4 signaling. The degree of incubation period shortening in the C3H/HeJ mice (group mean, 11.4%), equating to a more than 10-fold increase in infectious dose of scrapie relative to that in C3H/HeOuJ mice, was independent of the route (i.p. or i.c.), the dosage of infection, or the infecting strain of agent (139A or ME7). These results suggest that TLR4 signaling interferes with scrapie pathogenesis. Furthermore, since the only shared requirement for i.c. and i.p. scrapie infections is replication of the agent in the CNS (57), our findings strongly suggest that the anti-scrapie mechanism(s) induced by TLR4 signaling is functional in the CNS.

It is generally accepted that PrP^Sc levels in the brain are similar at the onset of clinical disease for prion infections that result from similar infection parameters. Therefore, the possibility that the PrP^Sc produced during scrapie progression induces neurological sequelae earlier in C3H/HeJ than in C3H/HeOuJ mice was contraindicated as both mouse strains appear to possess equivalent concentrations of PrP^Sc in the brain at the clinical stage of disease (given an identical dose of a particular scrapie strain and inoculation route). Thus, an alternative possibility, i.e., an increased rate of PrP^Sc accumulation in C3H/HeJ mice, is supported by our experimental findings; however, a more detailed investigation comparing PrP^Sc levels throughout the incubation period in C3H/HeJ and C3H/HeOuJ mice would be highly informative and provide additional support that PrP^Sc accumulates at a significantly accelerated rate in C3H/HeJ mice.

Studies have suggested that the TLR4 ligand LPS and fibrillogenic PrP peptides share a common signaling pathway leading to cytokine generation (3, 6, 46). Immune cells that express TLR4 such as MΦs, microglia, and DCs are activated by PrP^Sc (3, 5, 6, 38, 46, 64) and are capable of ingesting and degrading PrP^Sc (8, 12, 35, 41). During scrapie pathogenesis, bone marrow-derived MΦs enter the CNS, where they are likely involved in limiting the accumulation and spread of PrP^Sc/scrapie agent throughout the brain (50). This suggests that the absence of TLR4 signaling in C3H/HeJ mice may lead to reduced immune recognition of and response to PrP^Sc. To address this possibility in the current study, we tested MΦs from C3H/HeJ and C3H/HeOuJ mice and found that those from C3H/HeJ mice fail to respond to the PrP^Sc mimicking peptides, PrP_106-126 and PrP_118-135, but do respond to the TLR2 agonist zymosan. Thus, we conclude that accelerated scrapie pathogenesis in C3H/HeJ mice may be attributable, at least in part, to reduced innate immune responses to PrP^Sc in this mouse strain. In C3H/HeOuJ mice prion disease is slowed but not prevented, indicating that functional TLR4 signaling impedes the scrapie agent.

In addition to decreased phagocyte activation, other possible mechanisms contributing to the accelerated scrapie pathogenesis in TLR4-mutant C3H/HeJ mice can be envisaged. An additional link between TLR4 signaling and protective host responses to scrapie infection may be provided in the fact that TLR4 signaling induces anti-inflammatory cytokines such as IL-10 (10). IL-10, which is significantly increased in the cerebrospinal fluid of Creutzfeldt-Jakob disease patients (58), is a protective factor in the host response to prion infection, as clearly demonstrated by the significant acceleration of scrapie pathogenesis in transgenic IL-10^−/− mice (63). IL-10 production has been found to be impaired in C3H/HeJ mice (28, 42).

In addition to genetic mutations that affect TLR4 signaling, as occurs in C3H/HeJ mice, agents that function in similar fashion also alter prion pathogenesis. Adenovirus, which inhibits TLR4 signaling (18), induces a dose-dependent shortening of the scrapie incubation period in mice (20). Amphotericin B, which stimulates TLR4 signaling (54), prolongs the scrapie incubation period in mice after i.p. and i.c. infections (40, 48). Although the specific mechanisms by which these agents act to modulate prion pathogenesis have not been determined, these findings further support the hypothesis that TLR4 activation interferes with prion pathogenesis.

Implications for prion disease therapy.

In contrast to the effects of TLR inhibition observed in C3H/HeJ mice, TLR activation clearly interferes with prion disease progression. The TLR9 agonist CpG DNA significantly delays the onset of scrapie symptoms in i.p.-infected mice when administered peripherally at early postinfection time points (55, 56). Peripheral administration of complete Freund's adjuvant, essentially a cocktail of TLR stimulators (9), also significantly prolongs the scrapie incubation period in mice infected with scrapie agent either i.p. or i.c. (61). Since TLR agonists increase microglial phagocytosis and degradation of amyloid in vitro and in vivo (36, 59), these effects are most likely attributable to increased clearance by microglia or similar cell types.

Results of the current study suggest that TLR4 signaling interferes with prion disease pathogenesis. Might TLR4 activation represent an effective target for prevention or treatment of prion disease? Since PrP^Sc is insoluble and deposited in specific regions of the periphery and the CNS, the TLR4 activation that normally occurs during the course of a prion infection is likely to be triggered locally by PrP^Sc. TLR4 can be activated by treatment with LPS, synthetic LPS mimetics (30), or stimulatory anti-TLR4 MAbs (16). While administration of TLR4 ligands directly to the CNS may effectively induce clearance of amyloid (19), in practice this route would be complicated to exploit in humans. The alternative, systemic treatment with these agents by peripheral injection, could lead to lethal side effects induced by high levels of evoked proinflammatory mediators, such as TNF-α and IL-1β (17, 44). The negative consequences of TNF-α and IL-1β generation, which result from TLR4 activation, might be counteracted without dampening prion infection-protective phagocyte activation by coadministering specific antagonists to neutralize these cytokines (4, 27, 39).

As a preventative or therapeutic modality for fatal prion disease, stimulation of TLR4 may be optimally achieved during early stages of infection. First, to limit the spread of prion agent throughout the lymphoreticular tissues and the CNS, it would be beneficial to prevent the agent from replicating as early as possible. Second, TLR4 stimulation and subsequent phagocyte activation may be diminished later during the course of prion disease; diminished phagocyte activity has been observed in Alzheimer's disease (22), and we have observed in our own studies that peritoneal MΦs from preclinical and clinical scrapie-infected mice are hyporesponsive to LPS challenge (E. Park et al., unpublished data). This hyporesponsiveness to LPS may result from continuous TLR4 stimulation imposed by the ever-present TLR4 ligand, PrP^Sc, during scrapie infection; chronic stimulation of TLR4 leads to its downregulation and consequent hyporesponsiveness to LPS (43). One might then speculate that only a short burst of TLR4 activation early in the course of prion infection would be effective in arousing host responses that persist for the entire prion-infected host's life span. Therefore, early intervention with TLR4 stimulation may be most beneficial.

While additional studies in TLR4^−/− mice, in which the TLR4 gene is absent, would be desirable for exploring the relationship between scrapie pathogenesis and TLR4 signaling, the current studies in C3H/HeJ mice strongly suggest that TLR4 signaling interferes with scrapie pathogenesis and hint at possible underlying mechanisms. It would also be interesting to examine scrapie pathogenesis in TLR4-overexpressing mice (65), which, based on our results, may be predicted to exhibit a prolonged scrapie incubation period. Finally, the results presented here on TLR4 signaling and prion disease progression provide an additional lead for pursuing immune-based therapeutics for this currently untreatable neurodegenerative disease.