Report

Coincident Scrapie Infection and Nephritis Lead to Urinary Prion Excretion

See allHide authors and affiliations

Science  14 Oct 2005:
Vol. 310, Issue 5746, pp. 324-326
DOI: 10.1126/science.1118829

Abstract

Prion infectivity is typically restricted to the central nervous and lymphatic systems of infected hosts, but chronic inflammation can expand the distribution of prions. We tested whether chronic inflammatory kidney disorders would trigger excretion of prion infectivity into urine. Urinary proteins from scrapie-infected mice with lymphocytic nephritis induced scrapie upon inoculation into noninfected indicator mice. Prionuria was found in presymptomatic scrapie-infected and in sick mice, whereas neither prionuria nor urinary PrPSc was detectable in prion-infected wild-type or PrPC-overexpressing mice, or in nephritic mice inoculated with noninfectious brain. Thus, urine may provide a vector for horizontal prion transmission, and inflammation of excretory organs may influence prion spread.

The prion, the infectious agent of transmissible spongiform encephalopathies (TSEs), is detectable at extraneural sites long before clinical symptoms appear (1). PrPSc, a protease-resistant isoform of the host protein PrPC, accumulates mostly in central nervous system and lymphoid organs of infected organisms and may represent the infectious principle (2, 3). In addition to PrPC (4), splenic prion replication requires follicular dendritic cells (FDCs), the maintenance of which depends on B cells expressing lymphotoxins (LT) α and β (5). By activating local LTα/β signaling, which induces lymphoneogenesis, chronic inflammation enables ectopic prion replication (6). Inflammatory kidney conditions induced by bacteria, viruses, or autoimmunity are frequent in animals and humans, and urosepsis can occur in terminally demented patients (7). We therefore wondered whether renal inflammatory conditions might lead to urinary prion excretion.

To probe this possibility, we administered prions to RIPLTα and NZB × NZW F1 mice (henceforth termed NZBW) suffering from lymphocytic nephritis (figs. S1 and S2 and table S1), as well as NZW mice and milk fat globule–epidermal growth factor 8 (MFG-E8)–deficient mice, which develop glomerulonephritis but lack lymphofollicular inflammation (fig. S1).

After intraperitoneal (i.p.) prion inoculation [3 and 5 log LD50 (50% lethal dose) units of the Rocky Mountain Laboratory (RML) scrapie strain (passage 5, henceforth called RML5) (8)], brains and spleens of RIPLTα, NZBW, MFG-E8–/–, and control mice displayed similar prion and PrPSc loads (fig. S3, A to C). Whereas RIPLTα and NZBW kidneys progressively accumulated PrPSc and prion infectivity at 60 to 90 days postinoculation (dpi), presymptomatic (66 dpi) and terminally sick MFG-E8–/– mice lacked renal PrPSc (fig. S3D). Histoblot and immunohistochemical analysis identified PrPSc in renal lymphofollicular infiltrates of RIPLTα and NZBW mice (6).

RIPLTα, AlbLTαβ, C57BL/6 (4 to 6 months old), NZW, NZB, NZBW, MFG-E8–/–, tga20, and 129Sv × C57BL/6 mice (8 to 16 weeks old) were inoculated i.p. with 3 or 5 logLD50 scrapie prions. We dialyzed and purified urinary proteins from pools of three to six mice of each genotype at 30, 45, 60, 85, 110, 120, and 130 dpi (all presymptomatic) and from terminally scrapie-sick mice (Fig. 1). Each urine donor was confirmed to contain brain or spleen PrPSc and/or infectivity upon necropsy (fig. S3, A to C).

Fig. 1.

Transmission of prions through urine. Urine samples were collected from individual donors (horizontal lines) at time points after inoculation, denoted by vertical lines, and pooled (intersections between lines, arrows). Squares represent individual tga20 mice inoculated i.c. with urinary proteins. White squares: no scrapie symptoms; red squares: histopathologically confirmed scrapie; green squares: positive PrPSc immunoblot. Numbers within squares: days to terminal disease. Clinical disease: red line. Prion incubation time is expressed in days. Asterisk: intercurrent death without clinical scrapie signs.

Next, we quantified the recovery of spiked PrPSc and infectivity from urinary proteins (fig. S4). Scrapie cell endpoint assay (9) revealed a higher prion titer in dialyzed samples (fig. S4, C and D), possibly because dialysis removed biocontaminants inhibiting infection of PK1 cells.

Urinary proteins were purified by ultrafiltration followed by dialysis (∼600 μg pooled from groups of three to six mice), or by dialysis followed by ultracentrifugation, and inoculated intracerebrally (i.c.) into groups of three to eight tga20 mice that overexpress PrPC (10). We found prion infectivity within pools of presymptomatic (120 dpi, n = 3) and scrapie-sick RIPLTα (n = 6) and NZBW mice (n = 16). However, we did not find infectivity in C57BL/6 (n = 18), MFG-E8–/– (n = 8), 129Sv × C57BL/6 (n = 4), NZW (n = 12), or NZB (n = 4) urine at any time point after prion inoculation (Fig. 1). Urine from terminally scrapie-sick NZBW, NZW, and NZB mice could not be collected because the incubation time of scrapie exceeded the natural life span of these mice.

All clinically unaffected tga20 indicator mice were killed at ≥200 dpi. Histopathological and immunoblot analyses confirmed scrapie in all clinically diagnosed tga20 mice and excluded it from all others (Fig. 2, A to C, and fig. S5C). Phosphotungstate-mediated concentration of PrPSc from 1000 μg of protein did not reveal PrPSc in brains of clinically healthy urine-inoculated tga20 mice (fig. S5B). Thus, two pathogenetically distinct chronic inflammatory conditions of the kidney, in concert with prion infection, result in prionuria well before the onset of clinically overt prion disease.

Fig. 2.

Scrapie pathology in mice exposed to urine of nephritic mice. (A and B) Brain sections of tga20 mice that succumbed to scrapie after i.c. inoculation with urinary proteins from RIPLTα (terminal) (A) or NZBW mice (130 dpi) (B), showing gliosis (GFAP, glial fibrillary acidic protein) and PrP deposition (SAF84). Tga20 brains inoculated with urine from terminally sick C57BL/6 or presymptomatic NZW mice showed little or no astrogliosis and no PrP deposition. (C) (Upper panels) PrPSc in brains of tga20 mice inoculated i.c. with NZBW urinary proteins (130 dpi). Ten micrograms (left) or 20 μg (right) of tga20 brain were digested with proteinase K and immunoblotted. (Lower left panel) PrPSc in brains of tga20 mice inoculated i.c. with NZBW or RIPLTα urinary proteins. Lanes 4 to 7: Inoculation with NZBW urinary proteins at 60 dpi (lanes 4 and 5) and 110 dpi (lanes 6 and 7). Positive controls: scrapie-sick tga20 brain homogenate (left two lanes of each blot). Negative control: brain homogenate of a healthy tga20 mouse. (Lower right panel) Inoculation with RIPLTα urinary proteins at 120 dpi. (D) Prions were detected in tga20 mice exposed to urine from mice with lymphocytic nephritis (18.2%), but not in mice without kidney pathology or with isolated glomerulonephritis.

Whereas RIPLTα and NZBW mice suffer from combined interstitial lymphofollicular inflammation and glomerulonephritis, MFG-E8–/–, NZW, and NZB mice display glomerulonephritis but lack lymphofollicular foci (figs. S1 and S2). Hence, prionuria necessitates intrarenal organized inflammatory foci (6) and is not elicited by isolated glomerulonephritis (Fisher's exact test, P = 0.031). Urinary proteins from presymptomatic and terminal RIPLTα mice induced similar attack rates, suggesting similar urinary prion infectivity titers in presymptomatic and scrapie-sick mice. The consistent lack of infectivity in urine from noninoculated mice and prion-sick wild-type mice makes it unlikely that infectivity found in urine of nephritic mice represents a contaminant.

Scrapie-infected hamsters and Creutzfeldt-Jakob disease (CJD) patients were reported to excrete urinary PrPSc (UPrPSc) (11). However, these findings were not reproduced (12) and were deemed artifactual (13, 14). We attempted to detect UPrPSc in presymptomatic and terminally sick RIPLTα, MFG-E8–/–, tga20, C57BL/6, and 129Sv × C57BL/6 mice, as well as in presymptomatic NZW, NZB, and NZBW mice. Overnight dialysis did not affect the quantitative recovery of spiked PrPSc from urine (fig. S4, A and B); the detection threshold was ≥100 ng of terminal brain homogenate per milliliter of urine (Fig. 3, B and D), equivalent to 103 median infectious dose (ID50) units/ml. Under these conditions, we failed to reveal any UPrPSc, even in prionuric mice (Fig. 3 A, C, and D). These negative findings are not unexpected, because urinary infectivity titers were typically ≤1 ID50 units per 2 ml of pooled urine (Fig. 1), which is below the detectability of PrPSc (Fig. 3B).

Fig. 3.

Failure to detect urinary PrPSc. (A) Immunoblot analysis of urinary proteins from terminally scrapie-sick C57BL/6 mice. No PrPSc was found after ultracentrifugation. For control, Prnpo/o urine was spiked with scrapie brain homogenate. (B) Threshold of PrPSc detection in urinary proteins purified by dialysis and ultracentrifugation. C57BL/6 urine was spiked with serial dilutions of brain homogenate. Assay sensitivity: ≥100 ng of terminal brain homogenate per milliliter of urine (≅103 ID50 units/ml). (C) Immunoblot analysis of urinary proteins after ultracentrifugation. Scrapie-sick tga20 mice lacked UPrPSc. PK, proteinase K digestion; ICSM-18, primary antibody to PrP. Omission of primary antibody (right) abolished all signals. (D) Immunoblot analysis of urinary proteins from presymptomatic [NZB, NZW, and NZBW (100 dpi)] and terminally scrapie-sick mice. No PrPSc was detected after ultracentrifugation (long exposure). Controls: scrapie brain homogenate used for spiking (lane 1); urine spiked with brain homogenate from scrapie-sick (lanes 2 to 5) or healthy mice (lane 6).

We then tested whether inflammation of nonexcretory organs leads to prionuria. We administered prions to AlbLTαβ mice, which lack nephritis but develop hepatitis (6). Urine from AlbLTαβ and appropriate wild-type control mice (four pools of n = 4 mice, 120 dpi) lacked prion infectivity and UPrPSc (Figs. 1 and 3D; fig. S5, B and C). Thus, extrarenal inflammation, though enabling prion accumulation at the site of inflammation, does not induce prionuria.

Because PrPC is necessary for prion replication (4), its expression may be rate-limiting for urinary prion excretion. We assessed prionuria in tga20 mice, whose renal PrPC content is six to eight times that of wild-type mice (fig. S3F). Pooled urinary proteins (600 μg each) from six terminally scrapie-sick tga20 mice were inoculated i.c. into tga20 mice (Fig. 1). None of the recipient tga20 mice developed scrapie. Upon necropsy (>200 dpi), no scrapie histopathology was detected (fig. S5C). Thus, PrPC overexpression does not induce prionuria. The PrPC content of RIPLTα, NZBW, and MFG-E8–/– kidneys was similar to those of wild-type controls (fig. S3, G and H). RIPLTα and NZBW kidneys contain FDC-M1+ cells with high, focal levels of PrPC (6), which may facilitate local prion replication (5). Inoculation of urinary protein from noninfected mice did not elicit any abnormality in tga20 mice (fig. S5C).

How do prions enter the urine? Upon extrarenal replication, blood-borne prions may be excreted by a defective filtration apparatus. Alternatively, prions may be produced locally and excreted during leukocyturia. Although prionemia occurs in many paradigms of peripheral prion pathogenesis (15, 16), the latter hypothesis appears more likely, because prionuria was invariably associated with local prion replication within kidneys.

Urine from one CJD patient was reported to elicit prion disease in mice (17, 18), but not in primates (19). Perhaps unrecognized nephritic conditions may underlie these discrepant observations. Inflammation-associated prionuria may also contribute to horizontal transmission among sheep, deer, and elk, whose high efficiency of lateral transmission is not understood.

Supporting Online Material

www.sciencemag.org/cgi/content/full/310/5746/324/DC1

Materials and Methods

Figs. S1 to S5

Table S1

References

References and Notes

View Abstract

Navigate This Article