In this herd of horses, which one has EPM? Is there a test that can tell us?
The definition of pathognomonic is “a specific measure or observation that is unique to a condition”. In this discussion, we want a test that is pathognomonic for EPM.
The hunt for a pathognomic test for early diagnosis of EPM
Everyone would like an ante-mortem, definitive diagnostic test for EPM, a neurodegenerative disease caused by Sarcocystis neurona. The reason we don’t think we will find that EPM-litmus test is that EPM is a syndrome, inflammatory responses are a significant part of post-infection disease.
S. neurona seropositive horses don't necessarily have EPM
Antibody tests will not be useful for finding acute disease
An overwhelming majority of horses in the United States have antibody against S. neurona. We already developed specific and precise tests that detect the serotype of S. neurona. Unfortunately, the mere presence of antibody won’t denote a horse that has active or acute disease. A change in titer to SAG 1, 5, and 6 over time is very important and our tests can give you that information. One use of our antibody testing is to determine the titer in response to treatment. These values are used clinically. And this is one reason why treating animals without first determining antibody level can be frustrating.
Our statistics show that 55% of horses with a presumptive diagnosis of EPM do not have EPM! They don't have EPM because they don't have specific S. neurona antibodies. They have inflammation, a better diagnosis is polyneuropathy. It would be great if these horses could get carte blanche diagnostics such as MRI’s, CAT scans, radiographs, and ultrasounds until the diagnosis was definitive. Remember that post-mortem evaluation isn’t always definitive, but it is permanent. We need a panel of tests that includes an inflammatory marker that can pinpoint the horse with S. neurona EPM versus those with alternate etiologies, that also cause neurological disease.
A good test is quick and cheap...and accurate
An EPM test must be available, quick, precise, and affordable. One problem with acutely ataxic horses is that they are too sick to travel. These horses are often so ill that euthanasia is an early option—we need a test that will facilitate a decision to continue treating just a few days, because just a few days can be life-saving.
We explored possible diagnostic tests using our S. neurona challenge model. Putative tests included lymphocyte proliferation assays, gene expression signatures, and serum amyloid A. Each test had limitations that included difficulty for end use, failure to perform, and lack of specificity. The lack of specificity has been a difficult hurdle because specificity, accuracy, and precision are required.
Sorting out tests for infection by experiment
As previously stated, inflammation is a large part of EPM-associated pathology. Inflammatory responses are innate and non-specific. There are many points in the cascade of inflammatory reactions that can be tested, alas none are pathognomonic for inflammation caused by S. neurona. We measured changes in non-specific mediators of inflammation, before and after infection using the Trojan Horse model.
Initially, we wanted to concentrate our efforts on acute phase reactive proteins. These proteins are usually pro and anti-inflammatory, and their effects depend on the reactive (target) tissues. Care was used in interpreting the presence of pro-inflammatory molecules that are present in the central nervous system versus the periphery. Innate immune response proteins up-regulate cytokines. Failure to identify a useful acute phase reactive protein sent us after a downstream cytokine, a cytokine that is turned on by these proteins such as IL-6, TGF-beta, or IFN-gamma. These downstream molecules are measurable by ELISA testing.
The in vivo controlled, blinded, challenge study
We teamed up with Anne E Schwab and Timothy Gerry (they are Canadian researchers) to identify candidate biomarkers associated with infection with Sarcocystis neurona in horses. Here is the synopsis of the work.
Consecutive testing from 1 to 12 weeks after infection in multiple horses with expensive equipment
Multiple serum samples from infected horses and control animals were taken 1 week prior and again at 9 times over 12 weeks following experimental infection. The sera were separated into 6 fractions using anion-exchange chromatography. Fractions 1, 3, 5 and 6 were then bound to weak cationic (CM-10) protein chip arrays in order to assess proteins of intermediate (10-30 kDa) and large (30-100 kDa) mass. Chips were read using SELDI-TOF mass spectrometry (MS) with a PBSII protein chip reader. Each chip underwent two series of MS analysis to obtain reproducible spectra for these two mass ranges. To analyze peptides (2-15 kDa), samples were bound to CM10 chips and an a-cyano-4-hydroxy-cinnamic acid (CHCA) matrix was applied. Spectra were obtained with a PBSII protein chip reader. We repeated our analysis on two different PBSII protein chip readers.
Study results and interpretation
Proteins were analyzed by size, only one small protein was significant
Minimal differences were observed in the intermediate and large protein ranges over the course of infection, so we concentrated on small protein and peptide profiles.
Differences between samples prior to infection and 11 weeks post-infection were observed in fraction 1, but in none of the other fractions. Analysis from the two different machines yielded similar serum protein profiles. Intensities of peaks that were significantly different (p= 0.05) at week 11 post infection compared to pre-infection were seen.
We did not meet our objective for an early diagnostic, the change wasn't apparent until 11 weeks after infection
Oh no! The differences were observed pre-infection and at 11 weeks post infection. We're going for a diagnostic that would be useful in hours to days…not three and a half months.
OK, what did we find
Despite the overwhelming disappointment that this small molecule wouldn't detect acute disease, at this point we were curious about it's identity.
The small molecule was serum amyloid A, SSA
Week 1 and week 11 samples from horses were fractionated using anion exchange chromatography, desalted using acetone precipitation and electrophoresed through a Novex 4-16% Bis-Tris gel, as well as a 16% Tris-Tricine gel. Bands of interest were excised for MS-MS sequence analysis at Genome Quebec. A band of 10-12 kDa appeared to be more abundant in week 11 samples and was identified to be serum amyloid A.
Data provides more questions than answers or dollars!
In this analysis none of the candidate biomarkers were postulated as specific indicators of S. neurona infection in horses. All the markers identified have been previously associated with inflammation or infection. Furthermore, none of the candidate biomarkers appeared to be detectable prior to the onset of symptoms in these horses, which is a necessary goal for an early diagnostic. Additional SELDI-ToF analyses that would generate more information are not readily apparent at this point. The project was shelved.
SSA is a late phase, acute marker in S. neurona infections in horses
Revisiting this molecule, SAA, isn't productive because we showed, by some pretty expensive tests, in a blinded controlled experiment, that SAA is a late phase response when associated with S. neurona. We will hold that information in the back of our minds just in case a future panel of tests would require a late phase marker.