Exploring the cause of initially reactive bovine brains on rapid tests for BSE

Abstract

Bovine spongiform encephalopathy (BSE) is an invariably fatal prion disease of cattle. The identification of the zoonotic potential of BSE prompted safety officials to initiate surveillance testing for this disease. In Canada, BSE surveillance is primarily focused on high risk cattle including animals which are dead, down and unable to rise, diseased or distressed. This targeted surveillance results in the submission of brain samples with a wide range of tissue autolysis and associated contaminants. These contaminants have the potential to interfere with important steps of surveillance tests resulting in initially positive test results requiring additional testing to confirm the disease status of the animal. The current tests used for BSE screening in Canada utilize the relative protease resistance of the prion protein gained when it misfolds from PrP(C) to PrP(Sc) as part of the disease process. Proteinase K completely digests PrP(C) in normal brains, but leaves most of the PrP(Sc) in BSE positive brains intact which is detected using anti-prion antibodies. These tests are highly reliable but occasionally give rise to initially reactive/false positive results. Test results for these reactive samples were close to the positive/negative cut-off on a sub set of test platforms. This is in contrast to all of the previous Canadian positive samples whose numeric values on these same test platforms were 10 to 100 fold greater than the test positive/negative cut-off. Here we explore the potential reason why a sample is repeatedly positive on a sub-set of rapid surveillance tests, but negative on other test platforms. In order to better understand and identify what might cause these initial reactions, we have conducted a variety of rapid and confirmatory assays as well as bacterial isolation and identification on BSE positive, negative and initially reactive samples. We observed high levels of viable bacterial contamination in initially reactive samples suggesting that the reactivity may be related to bacterial factors. Several bacteria isolated from the initially reactive samples have characteristics of biofilm forming bacteria and this extracellular matrix might play a role in preventing complete digestion of PrP(C) in these samples.

Keywords: bacterial contamination;; bovine spongiform encephalopathy;; confirmatory testing;; diagnostics;; false positive; rapid testing,.

FIGURE 1.

Western blots using the anti-PrP anitbody 6H4. Sample 1 and 4 are a mid-range positive BSE sample. 2A and B are replicate lanes of a poor tissue quality weak BSE positive sample (P4). Lane 3 is the same weak BSE positive sample (P4) after SAF purification. 5A and B are replicate lanes of a poor tissue quality initially reactive sample (R4) and lane 6 is the same sample after SAF purification. The increase in immuno-reactivity is obvious when comparing lanes 2A and B to lane 3.There is no obvious increase when comparing lanes 5A and B to lane 6.

FIGURE 2.

Immunohistochemistry (IHC) and Gram stains of confirmed BSE positive and initially reactive surveillance brain tissue samples. Tissue sections were immuno-stained with anti-prion antibodies (6H4 or L42) or Gram stained to identify the presence of Gram positive and/or negative bacteria. The BSE positive IHC images (A, B) show the coarse, granular labeling of PrP^Sc in both fair (A) and poor quality (B) tissue samples. The initially reactive samples (E, F) do not have this staining pattern in IHC. Gram positive and Gram negative bacteria can be seen in the BSE + and initially reactive tissues of fair (C, G) and poor quality (D, H). In fair quality tissues, bacteria is usually of a single type and is limited to the tissue peripheries (C, G). Poor quality tissues have an abundant and diverse bacterial flora distributed throughout the sections (D, H).

FIGURE 3.

The viable aerobic bacterial populations cultured from surveillance brain tissue samples. A non-reactive sample (N9a) had the highest individual colony count. The average CFU/mg of tissue for the initially reactive samples was 436 (N1 to N5) while the autolysed non-reactive samples averaged 422 CFU/mg of tissue (N6a to N10a).

FIGURE 4.

Biofilm assay optical densities for bacteria isolated from initially reactive and negative samples. One bacterial isolate from each of the first 3 initially reactive samples formed significantly more biofilm in the 24 hour growth period as represented by the increased average OD of more than 4 times greater (R1-3, R2-3 and R3-3) than bacterial isolates from negative samples (N1-1 to N1-4 and N6a-1 to N6a-4). None of the bacterial isolates tested from initially reactive sample 4 (R4-1 to R4-4) produced significantly more biofilm than the negative sample bacteria.