Abstract

Domoic acid toxicosis (DAT) is a major cause of California sea lion (Zalophus californianus) strandings along the west coast of the United States with more than 1,300 confirmed or suspected cases between 1998 and 2006.¹ Acute DAT in California sea lions manifests as seizures, often leading to death. Affected individuals may also develop chronic DAT, which is characterized by persistent seizures, as a result of cerebral anoxia and chronic hippocampal damage.² Markers of DAT are limited and diagnosis is typically uncertain until after imaging the hippocampus by MRI or after necropsy using histology. By developing markers of DAT, diagnosis and treatment could be greatly facilitated. Therefore, we focused on identifying peptide biomarkers in serum samples from 107 California sea lions from The Marine Mammal Center and the U.S. Navy Marine Mammal Program. Samples were chosen using strict quality selection criteria and were case controlled by age class and sex. Twenty individuals served as healthy controls and of the remaining 87 individuals, 34 had been diagnosed with acute DAT and 53 without DAT. Peptides were isolated from serum using C8-magnetic beads and were measured using matrix assisted laser desorption ionization time of flight (MALDI-TOF) mass spectrometry. Resulting peptide profiles (m/z 1,000 to 10,000) were aligned, peak heights were normalized to an internal standard (glufibrinopeptide; Glu-Fib) or total ion current, and 104 peaks were selected using Progenesis MALDI software. The performance of individual peptide masses as classifiers of acute DAT were evaluated using area under the receiver operator characteristic curves (AUC) and no single peptide mass was a good predictor of acute DAT. Data were modeled using a non-parametric support vector machine (nearest related neighbor; NRN) and artificial neural network (ANN). Using 104 peptide masses, a non-parametric support vector machine and 101 ANNs were trained with resulting AUCs of 0.9577 and 0.9412 (median ANN), respectively, suggesting that either is an excellent test to discriminate acute DAT from case controls and healthy animals. These models were then qualified using a blinded set of 20 sera consisting of 10 animals with acute DAT and 10 animals without DAT. Based on the qualification set, artificial neural networks were the best models. The performance measures for a single median ANN was: sensitivity, 100%; specificity, 60%; positive predictive value, 71%; negative predictive value, 100%. When all 101 ANNs were combined and allowed to vote for the outcome using Glu-Fib normalized data, the performance measures were: sensitivity, 30%; specificity, 100%; positive predictive value, 100%; negative predictive value, 59%. These results suggest that MALDI-TOF peptide profiling and neural networks together can perform as a highly sensitive (with 100% negative predictive value) or highly specific (with 100%positive predictive value) diagnostic tool for acute DAT.

Acknowledgements

Funding provided though the Office of Naval Research: N00014-08-1-0341. The authors would like to thank the US Navy Marine Mammal Program and Mr. Kevin Carlin of the National Marine Mammal Foundation as well as the veterinarians and technicians at The Marine Mammal Center. Also, thanks to the MUSC Nephrology Proteomics Laboratory and Jennifer Bethard (MUSC Mass Spectrometry Facility) for technical support.

References

1.  Bejarano AC, Gulland FM, Goldstein T, St Leger J, Hunter M, Schwacke LH, Van Dolah FM, Rowles TK. Demographics and spatio-temporal signature of the biotoxin domoic acid in California sea lion (Zalophus californianus) stranding records. Marine Mammal Sci 2008; 24: 899–912.

2.  Goldstein T, Mazet JAK , Zabka TS, Langlois G, Colegrove KM, Silver M, Bargu S, Van Dolah F, Leighfield T, Conrad PA, Barakos J, Williams DC, Dennison S, Haulena M, Gulland FMD. Novel symptomatology and changing epidemiology of domoic acid toxicosis in California sea lions (Zalophus californianus): an increasing risk to marine mammal health. Proc R Soc B 2008; 275: 267–276.