Follicular dendritic cell dedifferentiation reduces scrapie susceptibility following inoculation via the skin

Abstract

Transmissible spongiform encephalopathies (TSEs) are a group of subacute infectious neurodegenerative diseases that are characterized by the accumulation in affected tissues of PrP^Sc, an abnormal isoform of the host prion protein (PrP^c). Following peripheral exposure, TSE infectivity and PrP^Sc usually accumulate in lymphoid tissues prior to neuroinvasion. Studies in mice have shown that exposure through scarified skin is an effective means of TSE transmission. Following inoculation via the skin, a functional immune system is critical for the transmission of TSEs to the brain, but until now, it has not been known which components of the immune system are required for efficient neuroinvasion. Temporary dedifferentiation of follicular dendritic cells (FDCs) by treatment with an inhibitor of the lymphotoxin-β receptor signalling pathway (LTβR-Ig) 3 days before or 14 days after inoculation via the skin, blocked the early accumulation of PrP^Sc and TSE infectivity within the draining lymph node. Furthermore, in the temporary absence of FDCs before inoculation, disease susceptibility was reduced and survival time significantly extended. Treatment with LTβR-Ig 14 days after TSE inoculation also significantly extended the disease incubation period. However, treatment 42 days after inoculation did not affect disease susceptibility or survival time, suggesting that the infection may have already have spread to the nervous system. Together these data show that FDCs are essential for the accumulation of PrP^Sc and infectivity within lymphoid tissues and subsequent neuroinvasion following TSE exposure via the skin.

Introduction

The transmissible spongiform encephalopathies (TSEs), or prion diseases, are subacute neurodegenerative diseases that affect humans and both wild and domestic animals. Most TSEs, including natural sheep scrapie, bovine spongiform encephalopathy (BSE), chronic wasting disease in mule deer and elk, and kuru and variant Creutzfeldt–Jakob disease (vCJD) in humans, are acquired by peripheral exposure. Although the main route of transmission of BSE to cattle and other species is considered to be oral (ingestion), other routes of TSE transmission have been identified. Accidental iatrogenic transmissions of sporadic (s)CJD to patients have occurred through the transplantation of sCJD-contaminated tissues or via pituitary-derived hormones.¹ Recent evidence also indicates that vCJD in humans has been transmitted via blood transfusion.^2,3 Studies in mice have shown that skin scarification is an effective means of scrapie transmission, highlighting another potential route of accidental transmission.⁴ Therefore, it is possible that some cases of natural sheep scrapie might be transmitted through skin lesions either in the mouth⁵ or during close contact,⁶ or be passed from mother to offspring through sites of skin trauma at birth. Surgical instruments contaminated with sCJD infectivity have also been shown to have the potential to transmit disease.⁷ Together, these examples highlight important health and safety issues concerning risks to patients, health workers and scientists of acquiring disease. Biopharmaceutical and cosmetic products derived from sheep and cattle tissues might also have the potential to transmit disease when applied to skin lesions.^8,9 Understanding the immunobiology of scrapie transmission via the skin will aid the determination of risk and the development of therapeutic strategies.

The host prion protein (PrP^c) is critical for TSE agent replication and accumulates as an abnormal, detergent-insoluble, relatively proteinase-resistant isoform, PrP^Sc, in diseased tissues.¹⁰ PrP^Sc, or an intermediate between PrP^c and PrP^Sc, is considered to constitute a major, or possibly the sole, component of the infectious agent.¹¹ Once TSEs infect the central nervous system (CNS), the accumulation of PrP^Sc is accompanied by neurodegeneration and, ultimately, the death of the host. Following peripheral inoculation, TSE agents usually accumulate in lymphoid tissues prior to the dissemination of infection to the CNS. Within the lymphoid tissues of TSE-infected hosts,^12–16 PrP^Sc accumulation initially takes place in germinal centres in association with follicular dendritic cells (FDCs). Studies in rodents inoculated intraperitoneally with scrapie have shown that mature FDCs are critical for scrapie accumulation in lymphoid tissues, and in their absence neuroinvasion is significantly impaired.^12,13,17 From the lymphoid tissues, infectivity spreads to the CNS via peripheral nerves.^18,19

Previous studies have shown that a functional immune system is critical for scrapie neuroinvasion following inoculation by skin scarification, as mice with severe combined immunodeficiency (SCID) do not accumulate PrP^Sc and infectivity in their spleens, or develop clinical disease when inoculated with scrapie by this route.^4,20 SCID mice are indirectly deficient in FDCs as they require important stimuli from lymphocytes for their maturation.²¹ The induction of FDC development in SCID mice by bone marrow grafting restores the accumulation of scrapie in the spleen after inoculation via the skin.²⁰ However, whether FDCs or other components of the immune system are required for scrapie neuroinvasion following inoculation via the skin is not known. In an experimental system, migratory bone marrow-derived dendritic cells have been shown to have the potential to deliver scrapie infectivity directly to the nervous system.²² As skin is highly innervated, we considered that neuroinvasion might occur via an FDC-independent pathway. For example, lymphocytes or Langerhans' cells might acquire scrapie within the skin and transport it directly to local peripheral nerves. Therefore, in this study experiments were performed to investigate whether FDCs are required for scrapie neuroinvasion after inoculation via the skin.

Materials and methods

Treatment with lymphotoxin β-receptor (LTβR) immunoglobulin

C57BL/Dk mice (8–12 weeks old) were given a single intraperitoneal (i.p.) injection of a fusion protein containing the soluble LTβR domain linked to the Fc portion of human immunoglobulin G1 (IgG1) (LTβR-Ig)²³ or with 100 µg of polyclonal human IgG (hu-Ig) (Sandoglobulin^®) as a control.

Scrapie inoculation

Mice were inoculated with the ME7 strain of scrapie by skin scarification of the medial surface of the right thigh. Briefly, prior to scarification ≈ 1 cm² of hair covering the scarification site was trimmed using curved scissors and then removed completely with an electric razor. Twenty-four hours later, a 23-gauge needle was used to create a 5-mm long abrasion in the epidermal layers of the skin at the scarification site. Care was taken to avoid damage to the dermis or to draw blood during scarification. Then, using a 26-gauge needle, one droplet (≈ 6 µl) of ME7 scrapie inoculum, from a 1% or 0·1% (w/v) terminal scrapie mouse brain homogenate in physiological saline, was applied to the abrasion and worked into the site using sweeping strokes. The scarification site was then sealed with OpSite (Smith & Nephew Medical Ltd, Hull, UK) and allowed to dry before the animals were returned to their final holding cages. Following challenge, the animals were coded, assessed weekly for signs of clinical disease and killed at a standard clinical end-point.²⁴ Scrapie diagnosis was confirmed by histopathological assessment of TSE vacuolation in the brain.

At the time-points indicated, some mice were killed and their spleens and inguinal lymph nodes (ILNs) were taken for further analysis. For bioassay of scrapie infectivity, two half spleens were pooled from each treatment group and prepared as 10% (w/v) homogenates in physiological saline. Likewise, the ILNs draining the inoculation site were pooled from two mice and prepared as a 10% (w/v) homogenate. Groups of 12 C57BL/Dk indicator mice were injected intracerebrally (i.c.) with 20 µl of each homogenate. The scrapie titre in each sample was determined from the mean incubation period in the assay mice, by reference to established dose/incubation period response curves for ME7 scrapie-infected spleen tissue, as previously described.²⁵

Immunohistochemical analysis

To monitor the effects of treatment on FDC status, ILNs and half spleens were taken from two mice in each group and snap-frozen at the temperature of liquid nitrogen. Serial frozen sections (thickness, 10 µm) were cut on a cryostat and FDCs were visualized by staining with either the FDC-specific rat monoclonal antiserum, FDC-M2 (AMS Biotechnology, Oxon, UK), or 8C12 monoclonal antiserum to detect CD35 (BD Biosciences PharMingen, Oxford, UK). Immunolabelling was carried out by using alkaline phosphatase coupled to the avidin–biotin complex (Vector Laboratories, Burlingame, CA, USA). Vector Red (Vector Laboratories) was used as a substrate.

For the detection of PrP in the brain, tissues were fixed in periodate–lysine–paraformaldehyde and embedded in paraffin wax. Sections (6-µm thickness) were deparaffinized and pretreated to enhance PrP immunostaining by hydrated autoclaving (15 min, 121°, hydration) and subsequent immersion in formic acid (98%) for 5 min.²⁶ Sections were then stained with the PrP-specific monoclonal antiserum, 6H4 (Prionics, Zürich, Switzerland), and immunolabelling was detected by using hydrogen peroxidase coupled to the avidin–biotin complex (Vector Laboratories), with diaminobenzidine (DAB) as a substrate. Glial fibrillary acid protein (GFAP) was detected on adjacent brain sections by using rabbit GFAP-specific antiserum (DAKO Ltd, Ely, UK), and immunolabelling was carried out by using alkaline phosphatase coupled to the avidin–biotin complex with Vector Red as a substrate.

All sections were counterstained with haematoxylin to distinguish cell nuclei.

Paraffin-embedded tissue (PET) immunoblot detection of PrP^Sc

PrP^Sc was detected in PET sections of spleen and ILNs, as previously described.²⁷ Briefly, tissues were fixed in periodate–lysine–paraformaldehyde and embedded in paraffin wax. Serial sections (6-µm thickness) were mounted on poly(vinylidine difluoride) membrane (Bio-Rad, Hemel Hempstead, UK) and fixed by incubation at 55° overnight. Membranes were then deparaffinized and digested with proteinase K (20 µg/ml) for 16 hr at 55° (to confirm the presence of PrP^Sc), washed in TBS/Tween [10 mm Tris–HCl pH 7·8, 100 mm NaCl, 0·5% (v/v) Tween] and denatured in 3 m guanidine isothiocyanate (10 mm Tris–HCl, pH 7·8) for 10 min. Membranes were blocked in 2% casein, and PrP was detected with the PrP-specific rabbit polyclonal antiserum, 1B3,²⁸ followed by alkaline phosphatase-conjugated goat anti-rabbit immunoglobulin (Jackson ImmunoResearch Laboratories Inc., West Grove, PA). Bound alkaline phosphatase activity was detected by using SigmaFast™ nitroblue tetrazolium/5-bromo-4-chloroindol-2-yl phosphate (NBT/BCIP) solution (Sigma, Poole, UK). Immunostained membranes were assessed using an Olympus dissecting microscope.

Statistical analysis

Where indicated, incubation periods are presented as mean (days) ± standard error (SE). Significant differences between incubation periods in different groups were determined by using one-way analysis of variance (anova).

Results

Effect of LTβR-Ig treatment on FDC status in ILNs

The maintenance of FDCs in a differentiated state requires continual stimulation through the LTβR, as mature cells rapidly dedifferentiate when this signalling pathway is blocked.²⁹ Here, temporary blockade of the LTβR signalling pathway was achieved by a single i.p. injection of 100 µg of LTβR-Ig.²³ Within 3 days of treatment with LTβR-Ig, expression of the FDC markers FDC-M2 and CD35 (complement receptor 1) was undetectable in ILNs (Fig. 1) and in spleen (data not shown). The effects of treatment with LTβR-Ig on FDC status are temporary, lasting ≈ 28 days.^29,30 Treatment with 100 µg of hu-Ig, as a control, had no adverse effect on FDC status in ILNs (Fig. 1) or in the spleen (data not shown).

Figure 1.

Effect of treatment with lymphotoxin β-receptor immunoglobulin (LTβR-Ig) on follicular dendritic cell (FDC) status in the inguinal lymph nodes of uninfected mice. Tissues were taken 3 days after injection with polyclonal human immunoglobulin G (hu-Ig) (control) or LTβR-Ig, and adjacent frozen sections were stained with the FDC-specific monoclonal antiserum FDC-M2 (top row; red) and 8C12 antiserum to detect CD35 (bottom row; red). Expression of FDC-M2 and CD35 were undetectable in inguinal lymph nodes after treatment with LTβR-Ig. All sections were counterstained with haematoxylin (blue). Original magnification × 200.

Effect of LTβR-Ig treatment on the early accumulation of PrP^Sc within ILNs and the spleen

Following peripheral inoculation of mice with the ME7 strain of scrapie, high levels of PrP^Sc and infectivity accumulate in lymphoid tissues within the first few weeks postinoculation, and these levels are maintained throughout the course of infection.^12,13,31 In the present study, mice were treated with LTβR-Ig or hu-Ig (as a control), 14 or 42 days after inoculation with scrapie by skin scarification, and spleens and both ILNs were taken from two mice of each treatment group 3 days later (days 17 and 45, respectively). PrP^Sc accumulations within these tissues were detected by PET immunoblot analysis. PrP^Sc was detected within two or three lymphoid follicles of the draining (right) ILNs of hu-Ig-treated control mice, 17 days after inoculation with scrapie (Table 1, treatment on day 14). The cellular distribution of the PrP^Sc was consistent with its accumulation on FDCs.¹³ No PrP^Sc was detected in the non-draining (left) ILNs or spleen 17 days after inoculation (Table 1, treatment on day 14). By 45 days after inoculation with scrapie, PrP^Sc was present in a greater number of lymphoid follicles in the draining ILNs of control-treated mice (Table 1, treatment on day 42; Fig. 2a). Furthermore, PrP^Sc was also detectable in a single lymphoid follicle in the left ILN (Table 1) and in spleen (Table 1, treatment on day 42; Fig. 2c) at this time-point. These data demonstrate that following inoculation via the skin, PrP^Sc accumulates first upon FDCs in the draining ILN and subsequently spreads to other non-draining lymph nodes and spleen between 17 and 42 days after inoculation.

Table 1. The effect of treatment with lymphotoxin β-receptor immunoglobulin (LTβR-Ig) on the early accumulation of PrP^Sc in the inguinal lymph nodes and spleens of scrapie-inoculated mice^*†.

	hu-Ig			LTβR-Ig
Day of treatment	Right ILN‡	Left ILN§	Spleen	Right ILN	Left ILN	Spleen
14	++	−	−	+/−	−	−
42	+++	+	+	+/−	−	+/−

^*Mice were given a single intraperitoneal (i.p.) injection (100 µg) of LT β R-Ig or polyclonal human immunoglobulin G (hu-Ig), as a control, on the days indicated after inoculation with scrapie via skin scarification of the right thigh.

^†Tissues were taken 3 days after treatment, and the number of PrP^Sc-containing lymphoid follicles in inguinal lymph nodes (ILNs) and spleen from two mice was scored as follows: +++, ≥ 4 positive follicles; ++, 2–3 positive follicles; +, 1 positive follicle; +/−, 1 positive follicle in only one of the samples; −, no PrP^Sc detected.

^‡Inguinal lymph node draining the site of inoculation.

^§Non-draining inguinal lymph node.

Figure 2.

Effect of treatment with lymphotoxin β-receptor immunoglobulin (LTβR-Ig) on the early accumulation of PrP^Sc in the draining inguinal lymph node and in the spleen. Mice were given a single intraperitoneal (i.p.) injection of LTβR-Ig or polyclonal human immunoglobulin G (hu-Ig) (control) 42 days after inoculation with scrapie via skin scarification on the right thigh. Tissues from two mice from each group were collected 3 days after treatment, and PrP^Sc accumulations were determined by paraffin-embedded tissue (PET) immunoblotting. Abundant PrP^Sc accumulations were detected in the lymphoid follicles of polyclonal human immunoglobulin G (hu-Ig)-treated animals (a and c, dark staining, arrowheads). In contrast, PrP^Sc accumulations were undetectable in tissues from LTβR-Ig-treated mice (b and d). Terminally scrapie affected brain tissue (e) and uninfected normal brain tissue (f) were included as controls to confirm the specificity of PrP^Sc detection.

When mice were treated with LTβR-Ig, the number of follicles containing PrP^Sc in the draining ILN and spleen were visibly reduced or completely absent 3 days after treatment (Table 1; Fig. 2b, 2d, respectively). Therefore, the temporary dedifferentiation of FDCs correlated with a rapid reduction in the number of PrP^Sc-positive lymphoid follicles in ILNs and the spleen.

Effect of LTβR-Ig treatment on the early accumulation of infectivity within ILNs and the spleen

The draining ILNs and spleens were taken from two mice of each control and LTβR-Ig-treatment group 70 days after scrapie inoculation by skin scarification. The scrapie infectivity titre in pooled (n = 2) tissue homogenates was estimated by bioassay in groups of indicator mice. As expected, ILNs from each group of hu-Ig-treated control mice contained high infectivity titres [6·3–6·7 log i.c. 50% infectious dose (ID₅₀)/g; Table 2]. In the draining ILNs of mice treated with LTβR-Ig 3 days before scrapie inoculation, infectivity was undetectable, suggesting a scrapie infectivity titre, if present, of < 3·5 log i.c. ID₅₀/g (at least 1000-fold less than the level detected in ILNs of control-treated mice assayed at the same time postinoculation; Table 2). Thus, temporary FDC dedifferentiation before inoculation with scrapie via the skin blocks the early accumulation of infectivity in the draining ILN. The mean incubation periods obtained following injection of pooled lysates of ILNs from mice treated with LTβR-Ig 14 or 42 days after inoculation, were significantly longer than those obtained following injection of lysates from control mice (P < 0·01 and P < 0·001, n = 9, anova, respectively; Table 2). Therefore, ILNs from mice treated with LTβR-Ig 14 or 42 days after inoculation, contained detectable but significantly lower levels of infectivity than those measured in ILNs from control-treated mice (Table 2).

Table 2. Effect of treatment with lymphotoxin β-receptor immunoglobulin (LTβR-Ig) on the accumulation of scrapie infectivity in the draining inguinal lymph node 70 days after inoculation with scrapie via the skin^†.

	hu-Ig			LTβR-Ig
Day of treatment	Incidence‡	Mean incubation period (days) ± SE	Titre§	Incidence	Mean incubation period (days) ± SE	Titre
−3	7/7	196 ± 5	6·6	0/9	9× > 300	< 3·5
+14	7/7	193 ± 6	6·7	8/9	217 ± 3*	5·5
+42	8/8	203 ± 4	6·3	6/8	249 ± 4**	4·1

^†Mice were given a single intraperitoneal (i.p.) injection (100 µg) of LT β R-Ig or polyclonal human immunoglobulin G (hu-Ig) as a control on the days indicated before or after inoculation with scrapie via skin scarification of the right thigh. Inginal lymph nodes draining the site of inoculation were pooled from two mice and infectivity levels were determined by intracerebral (i.c.) injection of lysates into groups of C57BL/Dk indicator mice.

^‡Incidence = number of animals affected/number of animals tested. The notation ‘n × > 300’ means that mice were free of the signs of scrapie up to at least this time-point after inoculation.

^§Scrapie infectivity titres expressed as log i.c. 50% infectious dose (ID₅₀)/g.

^*P < 0·01, when compared to the mean incubation period for hu-Ig control tissues.

^**P < 0·001, when compared to the mean incubation period for hu-Ig control tissues.

Similarly, spleens taken from each group of control-treated mice 70 days after inoculation contained high levels of scrapie infectivity (6·0–7·1 log i.c. ID₅₀/g; Table 3). However, after treatment with LTβR-Ig 3 days before or 14 days after inoculation with scrapie, infectivity was undetectable in the spleen, suggesting a scrapie infectivity titre, if present, of < 3·5 log i.c. ID₅₀/g (at least 1000-fold less than the level detected in spleens of control-treated mice assayed at the same time postinoculation; Table 3). However, comparisons of mean incubation periods obtained following injection of spleen lysates from mice treated 42 days after inoculation, suggested that spleens from LTβR-Ig-treated mice contained detectable, but significantly lower, levels of infectivity (at least 100-fold less; P < 0·02, n = 9, anova) than those measured in spleens from control-treated mice (Table 3).

Table 3. Effect of treatment with lymphotoxin β-receptor immunoglobulin (LTβR-Ig) on the accumulation of scrapie infectivity in the spleen 70 days after inoculation with scrapie via the skin^†.

	hu-Ig			LTβR-Ig
Day of treatment	Incidence‡	Mean incubation period (days) ± SE	Titre§	Incidence	Mean incubation period (days) ± SE	Titre
−3	9/9	187 ± 3	7·1	0/9	9× > 300	< 3·5
+14	8/8	188 ± 4	7·0	0/8	9× > 300	< 3·5
+42	9/9	202 ± 9	6·0	6/9	243 ± 5*, 3× > 300	≤ 4·1

^†Mice were given a single i.p. injection (100 µg) of LT β R-Ig or polyclonal human immunoglobulin G (hu-Ig) as a control on the days indicated after inoculation with scrapie via skin scarification of the right thigh. Spleens were pooled from two mice and infectivity levels were determined by intracerebral (i.c.) injection of lysates into groups of C57BL/Dk indicator mice.

^‡See Table 2, footnote^‡

^§See Table 2, footnote ^§

^*P < 0·02, when compared to the mean incubation period for hu-Ig control tissues.

PrP^Sc accumulation within ILNs and the spleen at the terminal stage of disease

As expected, abundant PrP^Sc was detected by immunoblot analysis in ILNs and spleens from all hu-Ig-treated control animals that developed clinical signs of scrapie (data not shown). The effects of LTβR-Ig treatment on FDC status are temporary, and mature networks reappear ≈ 28 days later.^29,30 Therefore, the detection of abundant PrP^Sc in ILNs and spleens of LTβR-Ig-treated animals that developed clinical signs of scrapie (data not shown) is consistent with the replication of residual infectivity on recovered FDC networks in these tissues.

Effect of LTβR-Ig treatment on scrapie susceptibility

Mice were given a single i.p. injection of LTβR-Ig or hu-Ig (as a control) at one of the following three time-points relative to scrapie inoculation by skin scarification: 3 days before scrapie inoculation (so that mature FDCs would be absent in lymphoid tissues at the time of challenge); 14 days after inoculation (i.e. soon after the onset of scrapie accumulation in the draining ILN; Table 1); or 42 days after inoculation, when abundant PrP^Sc (Table 1; and Fig. 2a) and infectivity (J. Mohan et al. unpublished) are detectable in the draining ILN.

All control mice treated with hu-Ig 3 days before scrapie inoculation succumbed to disease with a mean incubation period of 320 ± 3 days (n = 6; Table 4). In contrast, LTβR-Ig-treatment 3 days prior to scrapie inoculation reduced disease susceptibility and significantly extended the survival time (Table 4). Four of seven LTβR-Ig-treated mice remained free of the signs of scrapie for up to at least 480 days postinoculation. However, three of seven LTβR-Ig-treated mice did succumb to scrapie after individual incubation periods of 350, 381 and 402 days. These incubation periods were beyond the range seen in the hu-Ig-treated control mice (309–327 days), and the mean incubation period (378 ± 15 days, n = 6) was significantly longer than the mean incubation period for control mice (P < 0·001, anova). Characteristic spongiform pathology, disease-specific PrP accumulation and reactive astrocytes expressing high levels of GFAP were detected in the brains of all control and LTβR-Ig-treated mice that developed clinical scrapie (Fig. 3). In contrast, spongiform pathology, disease-specific PrP accumulation or reactive astrocytes were not detected in the brains of the surviving LTβR-Ig-treated mice (Fig. 3), indicating that infection had not spread to the nervous system in these surviving mice.

Table 4. Effect of treatment with lymphotoxin β-receptor immunoglobulin (LTβR-Ig) on scrapie susceptibility following inoculation via the skin^†.

	hu-Ig		LTβR-Ig
Day of treatment	Incidence‡	Mean incubation period (days) ± SE	Incidence	Mean incubation period (days) ± SE
−3	6/6	320 ± 3	3/7	378 ± 15*, 4× > 480
+14	8/8	316 ± 3	8/8	351 ± 4**
+42	7/7	328 ± 6	8/8	343 ± 8***

^†Mice were given a single intraperitoneal (i.p.) injection (100 µg) of LT β R-Ig or human immunoglobulin G (hu-Ig) as a control on the days indicated before or after inoculation with scrapie via skin scarification of the right thigh.

^‡See Table 2, footnote ^‡

^*P < 0·001, when compared to hu-Ig controls.

^**P < 0·000005, when compared to hu-Ig controls.

^***P = 0·158, when compared to hu-Ig controls.

Figure 3.

Immunohistological analysis of brain tissue from mice treated with polyclonal human immunoglobulin G (hu-Ig) (control) or with lymphotoxin β-receptor immunoglobulin (LTβR-Ig) 3 days before scrapie inoculation by skin scarification. Large disease-specific PrP accumulations (upper row; brown), and reactive astrocytes expressing high levels of glial fibrillary acid protein (GFAP) (middle row; red) and spongiform pathology (haematoxylin & eosin, lower row) were detected in the hippocampi of all mice showing clinical signs of scrapie. In contrast, in the brains of LTβR-Ig-treated mice that remained free of the clinical signs of disease, no evidence of PrP accumulation, reactive astrocytes or spongiform pathology was detected 480 days after inoculation. All sections were counterstained with haematoxylin (blue). Original magnification × 200. dpi, days postinoculation on which the tissues were analysed; pos., mice that developed clinical signs of scrapie; neg., mice that were free of the clinical signs of scrapie.

All mice treated with LTβR-Ig 14 days after inoculation developed clinical disease with a significantly extended incubation period when compared to control-treated mice (P < 0·000005, n = 8, anova; Table 4). However, treatment with LTβR-Ig 42 days after scrapie inoculation had no significant effect on the survival time when compared to control-treated mice (P = 0·158, n = 8, anova; Table 4).

Discussion

In order to determine the involvement of FDCs in scrapie pathogenesis following inoculation via the skin, these cells were temporarily dedifferentiated through the blockade of the LTβR signalling pathway, either before, or shortly after, scrapie inoculation. Data presented here show that treatment with LTβR-Ig blocked the early accumulation of scrapie in the draining ILN and spleen. These effects coincided with the temporary dedifferentiation of FDCs in lymphoid tissues. When given 3 days before scrapie inoculation, LTβR-Ig reduced disease susceptibility and extended the survival time when compared to control-treated mice. No pathological signs of scrapie were detected in the brains of surviving mice treated with LTβR-Ig, confirming that neuroinvasion does not occur by the direct uptake of infectivity by peripheral nerves in the skin. Thus, FDCs are critical for the transmission of scrapie from the skin to the CNS. Whereas treatment with LTβR-Ig 14 days after scrapie inoculation also significantly extended the survival time, treatment 42 days after scrapie inoculation did not, suggesting that infectivity had already spread to the peripheral nervous system by this time.

Lymphocytes provide essential cytokine signals for FDC development and maturation,^21,32,33 as immunodeficient mice that lack expression of lymphotoxin (LT)α³⁴ or LTβ³⁵ lack mature FDC networks. Lymphocytes express these cytokines as a membrane-bound heterotrimer (LTα₁β₂), which signals through LTβR on FDCs or their precursors.³⁶ FDCs likewise do not develop in LTβR-deficient mice.³⁷ FDC networks require continual LTβR stimulation as they rapidly collapse from their mature state when LTβR signalling is specifically blocked by treatment with LTβR-Ig (Fig. 1).^23,29 A single treatment with LTβR-Ig temporarily dedifferentiates FDCs for ≈ 28 days.²⁹ FDCs trap and retain antigens on their surfaces through interactions between complement components and cellular complement receptors.^38,39 Furthermore, complement components C1q and C3, and cellular complement receptors are considered to play an important role in the localization of scrapie infectivity to FDCs.^40,41 In the present work, treatment with LTβR-Ig resulted in the temporary loss of expression of the FDC-associated molecules complement receptor 1 (CD35) and FDC-M2 in ILNs and the spleen. The epitope recognized by the FDC-M2-specific monoclonal antiserum has been identified as complement component C4.⁴² Therefore, the temporary loss of FDC-M2-specific immunostaining after treatment with LTβR-Ig suggests that any remaining immature FDCs, if present, would be unable to trap and retain complement-opsonized antigens. These data suggest that these cells would also have lost their ability to capture scrapie during the period of dedifferentiation.

Within 17 days after scrapie inoculation by skin scarification, PrP^Sc-containing FDC networks were detected in the draining ILN of control mice, but not in the non-draining lymph node or spleen. By 45 days after inoculation, the number of FDC networks containing PrP^Sc had visibly increased in the draining ILN and were also detectable in a few networks in the non-draining ILN and spleen. The detection of low levels of PrP^Sc in the non-draining ILN and spleen at the later time-point suggests that following accumulation in the draining lymph node, PrP^Sc is disseminated to other lymphoid tissues via the bloodstream. In contrast, mice treated with LTβR-Ig at 42 days post-TSE inoculation had visibly reduced or undetectable accumulations of PrP^Sc within 3 days of treatment. The reduced detection of PrP^Sc coincided with the loss of FDCs following treatment with LTβR-Ig. The rapid reduction in PrP^Sc after treatment with LTβR-Ig is probably a result of the release of PrP^Sc from dedifferentiating FDCs and its uptake and clearance by phagocytic cells, such as macrophages.^43–45

Following scrapie inoculation via skin scarification, high levels of infectivity begin to peak in the draining ILN around 50 days postinoculation and subsequently plateau (Mohan et al. unpublished). In this study, ILNs and spleens from control mice contained high levels of scrapie infectivity at 70 days postinoculation. However, when mice were treated with LTβR-Ig 3 days before inoculation, scrapie infectivity was undetectable within the ILNs and spleen at 70 days postinoculation, ≈ 40 days after the expected reappearance of mature FDCs. Furthermore, in the temporary absence of FDCs at the time of inoculation, disease susceptibility was reduced and survival time in those treated mice that did develop clinical disease was significantly extended. Thus, these data suggest that scrapie infectivity from the inoculum is unable to replicate in the draining lymphoid tissue in the temporary absence of FDCs at the time of inoculation. The reduced disease susceptibility of LTβR-Ig-treated mice and absence of infectivity in lymphoid tissues at 70 days postinoculation suggests that the original inoculum is cleared by macrophages.^43–45 However, three LTβR-Ig-treated mice did develop clinical disease, albeit with incubation periods beyond the range observed in control, treated mice. Therefore, in some LTβR-Ig-treated mice, a fraction of the inoculum was able to persist until the FDCs reappeared.

The levels of infectivity responsible for natural TSE transmissions are unknown but are probably much lower than the moderate dose used in this study. SCID mice, which have a permanent absence of FDCs, do not develop clinical scrapie when inoculated via the skin with the same dose of scrapie used in the current study.^4,20 Therefore, the dose of scrapie infectivity administered in previous^4,20 and in the current work is not taken up directly by peripheral nerves, bypassing a need for replication in lymphoid tissues. The lack of detection of disease-specific PrP accumulations in the brains of surviving LTβR-Ig-treated mice demonstrates that, in our study, neuroinvasion following inoculation via the skin does not occur by direct uptake via nerves in the skin. These data also suggest that infection will have reached the CNS in the LTβR-Ig-treated mice that did develop clinical disease following replication on regenerated FDCs, and not by direct uptake by peripheral nerves. However, these data do not exclude the possibility, following inoculation with a higher dose of scrapie than the one used here, that neuroinvasion might occur via an FDC-independent pathway, such as uptake by peripheral nerves.

Treatment with LTβR-Ig 14 days after scrapie inoculation by skin scarification likewise significantly extended the survival time, but had no effect on disease susceptibility to the dose of scrapie used. PrP^Sc had already begun to accumulate upon FDC networks in the draining ILNs prior to treatment with LTβR-Ig at 14 days postinoculation. This suggests that the level of infectivity which had accumulated prior to treatment with LTβR-Ig was sufficient to avoid substantial clearance by phagocytic cells during the period of FDC dedifferentiation. The extended incubation period in these LTβR-Ig treated mice is probably related to the time required for the FDC networks to restore and initiate replication of the retained scrapie infectivity. This would delay the subsequent transfer of infectivity to peripheral nerves. Our data are also consistent with the hypothesis that the action of macrophages on TSE infectivity is concentration-dependent: the low concentrations derived from the original inoculum may be easily destroyed, whereas higher concentrations, such as those present in the ILN 14 days after inoculation, are less easily cleared and a fraction is retained.

Treatment with LTβR-Ig has no effect on disease pathogenesis once infection is established within the peripheral or central nervous systems.^30,46 The lack of any observable effect of LTβR-Ig treatment on disease pathogenesis, when given 42 days after scrapie inoculation, suggests that neuroinvasion had already occurred prior to treatment. Neuroinvasion may have occurred directly from the draining ILN, which had a heavy deposition of PrP^Sc at the time of treatment, whereas only limited amounts of PrP^Sc were detected in the spleen (Table 1). Removal of the spleen before i.p. scrapie inoculation significantly extends the survival time, suggesting that the spleen plays an important role in neuroinvasion via this route.⁴⁷ In contrast, removal of the spleen before subcutaneous scrapie inoculation has no effect on pathogenesis.⁴⁸ Data from these studies and the current study indicate that the major route of neuroinvasion following inoculation via the skin is not via the spleen.

Treatment with LTβR-Ig inhibits or prevents the development of experimental autoimmune encephalomyelitis by impairing T-lymphocyte responses and migration.⁴⁹ LIGHT is a transmembrane protein produced by activated T lymphocytes that also binds to LTβR.⁵⁰ However, the effects of treatment with LTβR-Ig on scrapie pathogenesis are unlikely to be a result of impaired LTβR- or LIGHT-mediated T-lymphocyte responses or migration, as pathogenesis is unaffected in T-lymphocyte-deficient mice.^51–54 Signalling via LTβR has been shown to be important for the presence of migratory dendritic cells in the spleen.⁵⁵ Therefore, it is plausible that blockade of the LTβR-signalling pathway might have affected cell trafficking or the transport of scrapie infectivity to the draining ILN. However, as treatment with LTβR-Ig 14 days after inoculation significantly extended the survival time, the effects of treatment on scrapie pathogenesis are unlikely to be the result of effects on cell trafficking, as dendritic cells migrate to draining lymphoid tissues within the first few hours of antigen encounter.⁵⁶ Migratory Langerhans' cells might also transport TSEs from the epidermis to the draining lymph node. However, as LTα- or LTβ-deficiency does not affect the distribution of Langerhans' cells in the skin, or their ability to take up antigen and migrate to the draining lymph node,⁵⁷ the effects of LTβR-Ig treatment on scrapie pathogenesis are also unlikely to be caused by impaired transportation by Langerhans' cells. Collectively, these observations suggest that it is highly unlikely that the major effects of LTβR-Ig treatment on scrapie pathogenesis are independent of its effects on FDC maturation.

Our data demonstrate that mature FDCs are critical for the accumulation of scrapie in the draining lymph node following inoculation by skin scarification. Furthermore, in the temporary absence of FDCs, disease susceptibility is reduced. Treatments that inactivate FDCs may have therapeutic potential against peripherally acquired TSEs. However, such treatments have no effect on pathogenesis once disease has spread from the FDCs to peripheral nerves.⁵⁸ While little is known about the precise timing of these events, comparisons of the effects of LTβR-Ig treatment on scrapie pathogenesis following inoculation by different peripheral routes indicate that this period varies considerably according to the route of exposure. Neuroinvasion will probably occur rapidly from the gastrointestinal tract because treatment with LTβR-Ig, 14 days after oral inoculation, is ineffective.³⁰ In the current study, our data suggest that neuroinvasion occurs between 14 and 42 days after inoculation via the skin. However, treatment with LTβR-Ig remains effective, even when administered up to 42 days after i.p. scrapie inoculation.^17,46 Thus, if manipulation of FDCs were to be used therapeutically against TSEs, the time interval available for intervention would depend critically on the route of TSE exposure.