The difference between “EPM” testing formats
One more time, there are no EPM tests! Tests detect antibody against S. neurona
A common question to us is why a horse is positive on IFAT, negative on EDS 2, 4/3 and positive for one, or all, of our SAG tests. Or, why is the EDS test positive and the S. neurona SAG 6 test negative?
Testing for “EPM” is a theme running through the body of our work spanning 23 years. If you understand that EPM is a rare disease, S. neurona sarcocystosis is common, that EPM is a syndrome (inflammation that leads to polyneuritis that is not treated with anti-protozoals), a final hurdle for you might be understanding tests that detect S. neurona antibodies and then interpreting the results. This blog discusses the test formats and some of the differences between them.
Our philosophy is that your horse matters. There are tools that are needed to make good decisions. Our toolbox includes a working knowledge about infection versus disease, strain differences, the unique host selectivity of Sarcocystis, the biology of Sarcocystis, how and why the organism changes in culture and finally, a familiarity with the literature. Find these topics in our blogs. Now, let’s get to the topic of understanding S. neurona antibody tests and results.
The ABC’s of protein antigens
For this discussion we define antigens as proteins. Living organisms have thousands of proteins that are antigens. Antigens can be as unique to a strain of an organism as a fingerprint is to a human. Antigens can be specific to a life cycle stage, and some antigens are so common they are shared among species, or even shared among all living things.
Scientists select the most specific antigen in order to identify the aspect of what they are trying to study. Some antigens are immunodominant, meaning they elicit a large host (horse) antibody response. An antibody response is detected multiple ways.
Species specific antigens SAG 1, 5, and 6 of S. neurona
Answers the question "Does the horse have antibody against S. neurona?"
A SAG is a surface antigen. The SAG 1, 5, 6 ELISAs are three, independent tests used to detect antibodies in horses that have or had infections caused by S. neurona. Antibodies are produced when an animal is infected with a strain of S. neurona. The surface antigens we detect are strain specific, our test ignores the non-specific ones. We focus on the mutually exclusive antigens displayed by strains (serotypes) of S. neurona. S. neurona doesn’t express more than one of these antigens. These antigens are the most specific to detect EPM causing strains of S. neurona.
The SAG 6 antigen of S. neurona is similar to the SAG 6 antigen found on S. falcatula. The difference is biology. S. falcatula doesn’t infect horses! Any antibody against SAG 6 found in horse sera must be from infection caused by S. neurona. The data that S. falcatula doesn’t infect horses was “proven” by the University of Florida in the lab run by Rob MacKay.
SAGs 2, 3, and 4 are common antigens. They are found in all the organisms in the genus Sarcocystis. SAG 2, 3, and 4 tests cannot serotype, or distinguish, the three strains of S. neurona. This means the antibodies produced and detected by SAG 2, 3, 4 tests cross react between species. You will detect SAG 2, 3, and 4 in samples from animals infected with S. cruzi (cattle)! And Neospora! Sarcocystis neurona SAGs 2, 3, 4 cross-react with antigens found in Toxoplasma. And don’t forget the most common Sarcocystis infection in horses, S. fayeri. A horse with S. fayeri sarcocystosis, EMS, will test positive on the SAG 2, 3, 4 tests, but will be negative on the SAG 1, 5, 6 ELISA tests.
The IFAT test format and Western Blot
IFAT tests use whole organisms to capture antibodies. This test format captures a mix of antibodies found in horse serum. You see, while SAG 1, 5, 6 are mutually exclusive, the other common antigens of the organism are there to be recognized by the horse and detected by the IFAT test. There are many antigens in this format, not a single one as found in most ELISA format tests. If the IFAT test uses a S. neurona SAG 1 strain for capturing antibody, the common antigens will be detected but SAG 5 and SAG 6 strain antibodies are missed.
The same is true for the Western Blot, WB. The WB format is a mix of antigens used to detect a mix of antibodies. Early on, the WB was the only game in town. Improvements were made when antigens were cloned and the resulting recombinant antigens allowed the ELISA format to exploit a single antigen to detect antibodies to all the reactive sites on that one protein. If a WB employs only one protein, it would be as useful as the ELISA, just resource heavy.
In our paper, Characterizing the immunodominant 29 kDa surface antigen of S neurona, we reveal an important fact relating to the detection of antibodies. IFAT procedures don't change the structure of proteins, however WB protocols commonly use reducing conditions. WB tests will not have the same results when testing an animals sera using non-reducing conditions! Figure 7 of the 29 kDa paper clearly shows the decrease in detection of the immunodominant SAG 1 protein when reducing conditions are used in WB analysis.
This is common knowledge to molecular biologists and the laboratories running WBs. By chemically changing the protein and denaturing it, not only are bonds broken, but only linear reactive areas (epitopes) of the antigen are detected. Avoiding reducing conditions allows the analysis of conformational epitopes that are important in detecting disease. Conformational epitopes are paramount when detecting disease due to Toxoplasma gondii, check out the literature on this topic on PubMed.
What is 2, 4/3 anyway?
The SAGs 2, 3, 4 are found in nature, they are represented by recombinant SAGs 2, 3, 4. These formed the first generation of tests using recombinant antigens representing common Apicomplexan antigens. The recombinant protein SnSAG 4/3 is a synthetic protein. Parts of the S. neurona SAG 4 protein were fused to a SAG 3 protein, forming chimeras. The best fusion product was selected by testing a small group of selected horses' sera and commercialized as the SAG 2, 4/3 ELISA tests.
Different antigens detect different antibodies
The 4/3 is a chimeric protein that doesn’t exist in nature. Some SAG 3 and 4 linear epitopes will be detected by a 4/3 chimera, but not all of them, and not for each protein, SAG 3 and SAG 4. The test development process selected the "best" fit for their test using the selected screening sera and accepting one 4/3 chimera.
Please look at work done by Howe. The group recognized that there is quite a variation in the SAG 4 protein for strains of S. neurona isolated from horses. The SAG 4 is also variable in S. falcatula. You can read the work done by Marsh, Characterization of a Sarcocystis neurona isolate from a Missouri horse with equine protozoal myeloencephalitis, showing the strain differences in S. falcatula demonstrated by WB, conducted under reducing and non-reducing conditions. It is in this work she shows that a horse with a S. neurona SAG 5 infection has antibodies against one strain of S. falcatula SAG 4 protein, but not another S. falcatula SAG 4 strain, Figure 3.
It would be interesting to know if the 4/3 chimera uses a falcatula SAG 4 reactive sequence. More important information is Marsh's analysis of the horse infected with a SAG 5 S. neurona. The horse is WB seropositive to the SAG 4 in the Missouri strain of falcatula and it's own strain of S. neurona, but not the UCD1 or the SN2 S. neurona SAG 4. Note the detected SAG 4 protein was to conformational epitopes! In her figure 4, it is clear there are differences detected under non-reducing conditions between three strains of falcatula. Figure 5 shows the blotting pattern for the UF S. falcatula that did not infect horses in the UF study.
Take a step back, the SAG 2, 4/3 ELISA’s
The 2, 4/3 scientific paper describes a “defined sample set of equine sera and cerebrospinal fluids that were “characterized by WB and/or were from confirmed EPM horses”. If one were diagnosing EPM based on histopathology using inflammation and not presence of parasites, the jury would be out on the definitive diagnosis on this sample set. Recall the issues with WB, reduced conditions and lack of specificity if WB is the gold standard for selecting the defined samples.
The Trivalent Chimera
A trivalent test was developed based on the two tests, SAG 2, 4/3. The lab fused the recombinant SAG 2 with the chimeric SAG 4/3.
The Trivalent Chimera. Here is what the paper said:
“To achieve reliable antibody detection with a single ELISA instead of 2 separate ELISAs, rSnSAG2 was fused with rSnSAG4/3 into a single trivalent protein, designated rSnSAG2/4/3. Paired serum and CSF from 163 horses were tested with all 3 ELISAs. When the consensus antibody titers obtained with the rSnSAG2 and rSnSAG4/3 ELISAs were compared to the single SAG2/4/3 ELISA titers, Spearman rank correlation coefficients of ρ = 0.74 and ρ = 0.90 were obtained for serum and CSF, respectively, indicating strong agreement between the tests. When the rSnSAG2 and rSnSAG4/3 consensus serum-to-CSF titer ratio was compared to the rSnSAG2/4/3 serum-to-CSF titer ratio, the Spearman correlation coefficient was ρ = 0.87, again signifying strong agreement. Importantly, comparing the diagnostic interpretation of the serum-to-CSF titer ratios yielded a Cohen kappa value of 0.77. These findings suggest that the single ELISA based on the trivalent rSnSAG2/4/3 will provide serologic and diagnostic results that are highly comparable to the consensus of the 2 independent ELISAs based on rSnSAG2 and rSnSAG4/3.”
The trivalent test is highly comparable to the consensus of the 2 independent ELISAs that are based on the SAG 2, 4/3 assay.
They don't say how many horses in the US are seropositive on the trivalent test,
They did not say they diagnose EPM,
They did not say they detected native SAG 1, SAG 3, or SAG 4,
They do not detect conformational epitopes.
When comparing results of the trivalent chimera to the SAG 1, 5, 6 ELISAs there are large considerations. SAG 1, 5, 6 are serotype specific S. neurona antigens found in nature, on the disease causing organisms and the test detects both linear and conformational epitopes. The seroprevalence to SAG 1, 5, 6 in the United States are published.
Sarcocystis fayeri is a common equine infection
that causes neuromuscular disease, rarely
Sarcocystis fayeri infects almost all horses in the United States, yet only 6-10% of these infected horses get neuromuscular disease. Other tests, and testing formats, detect antigens of Sarcocystis sp. and can’t distinguish diseased from non-diseased animals.
In one published paper, the chimeric ELISA testing format justifies dilution in order to decrease detecting of S. fayeri antibodies. Reducing cross-reactivity in the neurona test is why samples are diluted to run the test. Samples are diluted 1:250, not the lowest dilution that detects antibody but the sweet spot to un-muddy the results. Yes, S. fayeri has surface antigens and the SAG 4 can be quite variable and cross reactive. And yes, S. fayeri displays common antigens cross reactive to neurona SAG 2 and SAG 3. Apparently, dilution is the solution.
By the way, the IFAT uses a 1:40 dilution and is subjectively interpreted. It has been suggested in the literature that when testing horses for disease one should use a lower dilution, if testing in research one should dilute to 1:80. That never made sense to us.
Our testing format doesn’t test for SAGs of S. fayeri to determine antibody against fayeri. We test horses with neuromuscular disease associated with S. fayeri by detecting anti-toxin. You see, only the S. fayeri strains that excrete a toxin are likely to be pathogenic and the toxin has been associated with disease. Our goal is to differentiate the S. fayeri EMS diseased horses from those that do not have disease.
We did not, nor will we ever say, any of these tests diagnose EPM. We differentiate the horses with infections by species using specific neurona SAGs and a disease-associated anti-toxin of S. fayeri. These tests don’t measure the same antigens, the formats are ELISAs.
Results between testing labs are explained based on antigen selection and the format of the test. For example, you say the WB result is negative but seropositive for SAG 1 ELISA. We suggest the WB used reducing conditions, eliminating the SAG 1 reactive band. You say the chimera test is negative and the SAGs 1, 5, 6 are positive. Perhaps the answer is in conformational epitopes we detect that are not detected in the chimera test. They don't detect SAG 6, while the SAG 1, 5, 6 ELISAs serotype the exposure. Or, it could be the dilution used to run the chimera test. They watered down the antibody at a 1:250 dilution while the antibodies we detect are present in the 1:4 dilution. S. fayeri isn't a problem in our assays.
Another possibility with the the trivalent chimera is the lab didn’t use fused domains that would be useful in your horse. The trivalent paper did not say that the chimera was 100% the same as their SAG 2 and 4/3 ELISA. The quote is: These findings suggest that the single ELISA based on the trivalent rSnSAG2/4/3 will provide serologic and diagnostic results that are highly comparable to the consensus of the 2 independent ELISAs based on rSnSAG2 and rSnSAG4/3. Were you asking the test to tell you if the animal was exposed to S. neurona or if the results would be highly comparable to the consensus of 2 ELISAs based on SAG 2 and the chimera 4/3?
Three independent serotype specific SAG ELISAs detect antibody against three strains of S. neurona in our analysis. S. fayeri is a useful test we use to detect anti-toxin in disease. If your sample is S. neurona negative we can use the Sidewinder test and detect demyelinating polyneuropathy. The bottom line is how testing helps you make treatment decisions for your clinically ill horse.