Oral Prion Infection Requires Normal Numbers of Peyer’s Patches but Not of Enteric Lymphocytes

Animals

C57BL/6, Sv129 × BL/6, μMT, 27 RAG-1^−/−, 28 and TNFα^−/−/LTα^−/− 26 mice were bred at the Institute of Laboratory Animal Science of the University of Zurich and maintained under specific pathogen-free conditions. Wild-type mice were obtained originally from the Jackson Laboratory, Bar Harbor, ME.

Alkaline Phosphatase Detection

For detection of alkaline phosphatase (AP) activity, fixed whole PPs were studied. Segments of the small intestine containing one PP were excised, opened, and pinned on dental wax. Tissues were fixed for 10 minutes in Baker formalin calcium medium (pH 6) on ice. At this stage, villi were removed using a binocular microscope, and incubated with 125 nmol/L Tris buffer (pH 9.2) at 37^°C for 5 minutes and then incubated in Tris buffer containing naphthol AS-BI phosphate (Sigma, Deisenhofen, Germany) with constant shaking for 10 minutes. The reaction was stopped with ice-cold buffer. Goblet cells were detected by incubating tissue briefly in 1% Alcian blue dissolved in 3% acetic acid. Whole-mount FAEs were then used to enumerate AP-negative M cells as a percentage of the total cell population within the FAE. Statistical significance was determined using the unpaired Student’s t-test. Values of P < 0.05 were considered as significant.

Challenge with Scrapie Prions

Mice were fed with logarithmic dilutions of 10% heat- and sarcosyl-treated brain homogenate prepared from mice infected with the Rocky Mountain Laboratory (RML) scrapie strain (passage 4.1). 8 log LD₅₀ of the prion homogenate RML were administered via gastrolavage directly into the stomach. Smaller amounts (see Table 1 ▶ ) were administered by feeding a mixture of brain homogenate with food pellets. For intraperitoneal inoculations, 6 log LD₅₀ of scrapie inoculum was administered. Mice were monitored on alternate days, and scrapie was diagnosed according to standard clinical criteria.

Infectivity Bioassay

Bioassays were performed on 1% homogenates of spleens or of PBS-perfused small intestines (containing Peyer’s patches). Tissues were homogenized in PBS/BSA with a microhomogenizer and passed several times through 18-gauge and 22-gauge needles. Following preparation of a homogeneous suspension, centrifugation was performed at 500 × g for 5 minutes. Supernatants (30 μl) were inoculated intracerebrally into groups of at least four tga20 mice. 29 The titer of the standard inoculum (7.9 log 10 of 50% lethal dose (LD₅₀)/ml, corresponding to 8.9 log LD₅₀/g of brain tissue) was determined by the 50% endpoint calculation method. 30 The relationship y = 11.45 − 0.088 (y, log LD₅₀ × ml⁻¹ of homogenate; x, incubation time in days to terminal disease) was calculated by linear regression. 25,31 Indicator mice were sacrificed after development of terminal scrapie and infectivity titers were calculated.

Histology and Immunohistochemistry

Paraffin sections (2 μm) and cryostat sections (10 μm) from brain, spleen and PPs were stained with hematoxylin and eosin. Immunostaining for GFAP was performed using rabbit GFAP antisera (1:300 dilution; DAKO, Carpinteria, CA) and detected with biotinylated swine anti-rabbit serum (1:250 dilution; DAKO), avidin-peroxidase and diaminobenzidine (Sigma). Antibodies reactive against the follicular dendritic cell marker FDC-M1 (clone 4C11; 1:300 dilution) or the pan-B cell marker CD45R/B220 (BD Pharmingen, San Diego, CA) were used for immunohistochemistry of frozen sections as previously described. 32

Western Blot Analysis

Tissue homogenates were adjusted to protein concentrations of 5 mg/ml (brain) or 8 mg/ml (spleen) and treated with proteinase K (20 μg/ml, 30 minutes, 37°C). 50 μg (brain) or 80 μg (spleen) of total protein of each sample were electrophoresed through a 12% SDS-PAGE gel. Proteins were transferred to nitrocellulose membranes by semi-dry blotting. Membranes were blocked with Tris-buffered saline containing 0.1% Tween 20, pH 8.0 (TBST)/5% nonfat milk, incubated with antibodies 6H4 (for brain protein) or 1B3 (for spleen protein), and detected using alkaline phosphatase-linked anti-mouse (6H4) or anti-rabbit (1B3) sera followed by development by enhanced chemiluminescence (ECL; Amersham, Uppsala, Sweden).

Normal Numbers of PPs in β7^−/− Mice

To evaluate the extent to which the absence of α₄β₇ integrin, TNF/lymphotoxin, and B and T cells affects the presence of PPs, mice were examined anatomically. PPs of wild-type mice of the C57BL/6 and 129Sv×BL6 genetic background were easily visible to the naked eye, whereas in β7^−/−, RAG-1^−/−, and μMT mice PPs were undetectable by gross inspection. However, when small intestines were inflated with saline buffer (PBS) and transilluminated under a dissecting microscope (16× magnification), numerous small Peyer’s patches could be discerned (Figure 1) ▶ . The number and location (data not shown) of Peyer’s patches in β7^−/− mice (n = 6–10, mean 8.0 ± 1.5) were similar to wild-type mice (n = 6–11, mean 7.3 ± 1.5). A comparison with Student’s t-test, however, yielded a P value of >0.05, indicating that the number of Peyer’s patches in the two groups was not significantly different. In contrast, the number of PPs in B cell-deficient μMT mice were significantly reduced (0–7, mean 2.5 ± 2.2; P < 0.01 if compared to the respective wild-type control). Combined B cell and T cell deficiency in RAG-1^−/− mice further reduced the number of PPs (0–5, mean 2.0 ± 1.7; P < 0.01). The absence of TNF-α and lymphotoxin-α (LTα) completely prevented the formation of visible Peyer’s patches (P < 0.01).

The size of the M cell-containing subepithelial domes of β7-deficient mice were reduced by at least 70% compared to those of wild-type mice (approximately 250 μm versus up to 800 μm for full length of equatorial sections). A similar reduction in the size of FAE was found in B cell-deficient PPs of RAG-1^−/− and μMT mice.

Peyer’s Patches of β7^−/−, RAG-1^−/−, and μMT Mice Are Atrophic but Contain FDCs

To determine whether the absence of α₄β₇ integrin and lymphocytes alters the cellular composition of secondary lymphoid organs, we studied spleen and PPs by immunohistochemistry using antisera reactive against the B lymphocyte marker B220 and with the FDC marker M1 (Figure 2) ▶ . The germinal center histology of spleens of β7^−/− mice was indistinguishable from wild-type spleens. B cell zones were distinct and contained FDC-M1-positive cells of normal shape and localization.

In wild-type PPs, FDC networks occupied large regions of the germinal centers. Although there was a significant reduction in the number of B cells in β7^−/− PPs, some FDCs were still detectable. In contrast, no PPs were observed in TNF-α/LTα-deficient mice, although a small number of B cells were still present in the small intestine of these mice. Lymphoid organ structure of spleens of TNF-α/LTα-deficient mice was severely impaired, without any separation between T and B cell zones, and FDCs were undetectable.

No FDC-M1 positive cells were detected in the spleens of RAG-1^−/− and μMT mice, in agreement with earlier reports. 22 Unexpectedly, clusters of cells reactive with FDC-M1 and CD35 (CR1, not shown) were consistently detected in the small atrophic Peyer’s patches of RAG-1^−/− and μMT mice. Immunostaining also revealed a number of B220 positive cells in spleens and PPs. The majority of these B220-positive cells are most likely NK cells, because the antibody used (CD45R/B220) reacts with an epitope on the extracellular domain of the CD45 glycoprotein, which is expressed on lytically active subsets of lymphokine-activated killer cells such as NK cells and their progenitors in addition to B lymphocytes. 33 The smaller B220⁺ cell population in the PPs of RAG-1^−/− and μMT mice most likely consist of immature B cells which are still able to produce functional immunoglobulins of the A class even in the absence of surface, membrane-bound IgM or IgD, therefore representing a developmental pro-B cell block in these strains. 34

β7^−/− Intestinal Epithelium Contains M Cells Despite Depleted Intraepithelial B Cells

M cells of the FAE are believed to be important antigen sampling sites, and can act as entry sites for pathogens. To determine whether the absence or reduction of lymphocytes alters the composition of the FAE, we performed immunostaining of PPs of wild-type, β7^−/−, RAG-1^−/− and μMT with antibodies reactive against specific markers for M cells (Figure 3, A–D) ▶ . M cells, visualized histochemically by their reduced AP activity, 35 were readily detectable in the FAE of wild-type mice (Figure 3A) ▶ . In β7^−/−, μMT, and RAG-1^−/− (Figure 3, B, C, and D ▶ , respectively), both PPs and AP-negative M cells were apparent. The presence of PPs and AP-negative M cells was further confirmed by scanning electron microscopy, which allowed visualization of the disorganized brush border of these cells (data not shown).

In contrast to PPs of wild-type mice, in which M cells were located at the periphery of the FAE, the distribution of M cells appeared to be random over the entire surface of FAE in PPs of β7^−/−, RAG-1^−/−, and μMT mice.

We then estimated the relative number of intestinal M cells per total epithelial cells within the FAE by counting AP-negative M cells under 400× magnification by light microscopy (Figure 3E) ▶ . The relative number of M cells was increased in the FAE of β7^−/− (mean: 15.4 ± 4.0 vs. 6.7 ± 1.8 for 129Sv × BL/6 mice, P < 0.01), RAG-1^−/− (mean: 13.5 ± 2.3 vs. 7.2 ± 2.7 for BL/6, P < 0.05) and μMT (mean: 14.2 ± 2.8 compared to BL/6, P < 0.05) tissues. These data indicate that the differentiation program of the FAE, including M cells, is maintained even in the absence of proper B cell numbers.

Unimpaired Scrapie Pathogenesis in β7-Deficient Mice

To determine the contribution of different components of the GALT to the intestinal uptake of orally-delivered prions, mice were challenged with logarithmic dilutions of scrapie inoculum. Attack rate and time to terminal disease were measured. Following high-dose oral infection (8 log LD₅₀) all β7^−/− mice, Sv129 × C57BL/6 wild-type mice, and 9 of 10 C57BL/6 wild-type mice developed clinical signs of scrapie with similar incubation times of 213 to 230 days (Table 1) ▶ . These results indicate that β7^−/− mice are highly susceptible to disease via oral administration of prions. Accordingly, accumulation of PK-resistant PrP^Sc was readily detectable in brains and spleens of clinically sick β7^−/− and wild-type mice (Figure 4) ▶ . Moreover, topography and intensity of spongiosis and gliosis were similar in clinically sick β7^−/− mice compared to clinically sick wild-type mice (Figure 5) ▶ . Importantly, the use of lower prion inocula via the oral route significantly decreased the susceptibility to disease in these mice compared to i.p. challenge. A 10-fold dilution of the scrapie inoculum reduced the scrapie attack rate to 50% in β7^−/− and to 50 to 25% in the different wild-type mice. Further dilution of the inoculum did not evoke scrapie in mice of any genotype.

After i.p. prion challenge (6 log LD₅₀) β7^−/− and wild-type mice had similar incubation times of 210 to 217 days, indicating normal peripheral prion pathogenesis. As described earlier, 8,14 B cell-deficient RAG-1 and μMT mice were completely resistant, and TNFα^−/−/LTα^−/− mice partially resistant, to prions administered i.p. Wild-type mice can be experimentally infected via the i.p. route with prion inocula of less than 3 intracerebral log LD₅₀, yet the minimal dose for oral infection is 7 intracerebral log LD₅₀ of scrapie inoculum. Therefore, these results confirm previous observations that the oral route of prion infection is relatively inefficient in mice.

In contrast to wild-type or β7^−/− animals, mice deficient for LTα and TNF-α and mice lacking mature B cells were virtually resistant to orally-administered prions, even when extremely large inocula were used. This provides evidence that neither small intestines from TNFα^−/− × LTα^−/− mice nor atrophic PPs of RAG-1^−/− and μMT mice are sufficient as entry sites of prions.

Although PPs of the small intestine are potential entry sites for pathogens, draining lymph nodes and the spleen provide lymphoid sites with a mechanism for pathogen clearance. However, RAG-1^−/−, μMT, and TNFα^−/− × LTα^−/− mice were highly resistant to i.p. infection with prions (Table 1) ▶ , 8 and therefore the phase of peripheral prion replication is also impaired in these mice. Accordingly, PrP^Sc was undetectable in brains and spleens of clinically healthy RAG-1^−/−, μMT, and TNFα^−/− × LTα^−/− mice at the indicated postinfection time points.

Peyer’s Patches of β7^−/− but Not RAG-1^−/− and μMT Mice Are Sites of Peripheral Prion Replication

We then investigated peripheral prion replication in PPs and spleens of β7^−/−, RAG-1^−/−, and μMT mice following oral challenge. PPs (Figure 6 ▶ , red symbols) and spleens (Figure 6 ▶ , blue symbols) were collected at various time points after oral (8 log LD₅₀) or i.p. (6 log LD₅₀) infection until terminal disease. Infectivity titers were determined by transmission of homogenized tissues to indicator mice. For those TNFα^−/− × LTα^−/− mice which lacked all PPs, selected portions of the small intestine were dissected and used for transmission studies.

Infectivity titers in spleens of wild-type (crosses) and β7^−/− mice (circles) after oral challenge were generally determined to be lower than after intraperitoneal challenge. Additionally, the kinetics of prion replication differed in PPs and spleens after oral and i.p. challenge. Following i.p. inoculation, high titers of infectivity were already present at approximately 50 dpi in both genotypes, whereas following oral challenge, infectivity was only detected at 110 days.

With the exception of trace levels of infectivity in spleens and PPs of RAG-1^−/− and TNFα^−/− × LTα^−/− mice at latter time points, we found no evidence for infectivity in any spleen and PP of challenged RAG-1^−/− (diamonds), μMT (squares), and TNFα^−/− × LTα^−/− (triangles) mice. Therefore, the absence of PPs in mice deficient in TNF-α and LTα prevents gastroenteric invasion of prions.

Even more surprisingly, PPs of RAG-1^−/− and μMT mice were devoid of infectivity despite the presence of FDC-M1-positive cell cluster (Figure 2) ▶ . As previously mentioned, it is probable that these cells are FDC-like since they are CD35-positive and have also been reported to be dependent on LTβ signaling. 21 However, it is not obvious whether these cells represent functionally inactive FDCs. The data presented here appear to suggest that ablation of mature B cells, which are the primary sources of LTα and LTβ, abolishes the induction of FDC-M1-positive cells in the spleen, but not in PPs.