Faculty Advisor

Dave Adams

Faculty Advisor

Elizabeth Ryder

Faculty Advisor

Dan Gibson


"Ependymin (EPN) is a goldfish neurotrophic factor (NTF) that is one of the most abundant secreted glycoprotein components of brain extracellular fluid (ECF) and cerebrospinal fluid. This protein was first discovered due to its enhanced turnover following learning events, but has since been found to function in other important cellular events such as long-term memory formation and optic nerve elongation (Shashoua, 1976; Shashoua, 1977; Shashoua, 1985). Goldfish EPN has several demonstrated effects on mammalian cells, and immuno-reactive EPN-like proteins have been observed in a variety of organisms ranging from invertebrates (Limulus) to mice. Some NTFs have been shown to alleviate oxidative stress, one of the primary mediators of cell damage in neurodegenerative conditions. One mechanism by which they accomplish this is to increase cellular levels of anti-oxidative enzyme superoxide dismutase (SOD). In fact, our lab recently showed that a synthetic EPN fragment (CMX-8933) increases SOD mRNA and protein levels in rat primary cortical cultures (Parikh, 2003). Transgenic mice and rabbits that overexpress SOD are resistant to ischemia, while mice that lack SOD present with worse ischemic damage. Thus, due to this important SOD activating NTF-like feature of EPN may have potential therapeutic applications for treating neurodegenerative conditions such as Alzheimer’s disease, Parkinson’s or stroke. Because full-length NTFs do not efficiently cross the blood brain barrier (BBB) when administered intravenously, our lab, in collaboration with Ceremedix, Inc. (Boston, MA), is interested in designing short peptides that mimic the action of full-length NTFs, especially EPN. Due to proteases that naturally exist in ECF, EPN is partially cleaved to release the 8 aa peptide KKETLQFR. Other brain proteins, like encephalins and endorphins, also release similar short active components, so we hypothesized that this 8 aa peptide may represent an active component of EPN. Indeed, our lab demonstrated that administration of this sequence in synthetic form (termed CMX-8933) to rat primary cortical cells increases cellular titers of SOD (Parkih, 2003). This thesis was divided into three parts. The first part investigated which signal transduction pathway is responsible for CMX-8933’s ability to upregulate SOD. Because our lab also showed that CMX-8933 activates the MAPK pathway (Hasson, 1998; El-Khishin, 1999; Adams et al., 2003) we hypothesized that CMX-8933 may use this pathway to upregulate SOD. Inhibition experiments were performed to test three known components of the MAPK pathway, and a member of an unrelated pathway. Six independent SOD immunoblot experiments demonstrated that pre-treatment of rat primary cortical cultures with specific inhibitors for the protein kinase-C family (PKC), protein tyrosine kinases (PTKs), or MEK protein kinases (MEKK), completely blocked (p = 0.0001) CMX-8933’s average 15-fold upregulation of SOD. Thus, these three critical components of the MAPK pathway appear to be involved in the CMX-8933-induced upregulation of SOD. An inhibitor of transcription factor NF-êB in an unrelated pathway had no significant effect (p = 0.901). The second part of this thesis tested whether treatment of rat primary cortical cultures with CMX-8933 increases the cellular titers of mRNAs related to translation. Previous observations indicated that treatment of these cells with with CMX-8933 induces neurite sprouting (Shashoua, unpublished), and that EPN plays a role in optic nerve elongation (Schmidt and Shashoua, 1988), two processes related to growth. So we hypothesized that EPN, or CMX-8933, may stimulate the transcriptioin of mRNAs related to growth. We tested mRNAs for translation factor EF-2, and ribosomal proteins S12 and L19 based on previous observations in our lab with hybridization arrays (Parikh, 2003). RT-PCR experiments indicated that treatment of rat primary cortical cultures with 10 ng/ml CMX-8933 for 5 hrs increased the mRNAs for S12 an average of 12-fold relative to untreated cultures (N = 3, p = 0.02), L19 an average of 9-fold (N = 3, p = 0.048), and EF-2 an average of 11-fold (N = 3, p = 0.045). Levels of housekeeper polyubiquitin remained unchanged. Thus 3 gene products related to growth are indeed upregulated by CMX-8933. The third part of this thesis investigated the SOD stimulatory effects of full-length EPN versus its cleavage product CMX-8933. Previous studies showed that full-length NTFs BDNF and NGF upregulate SOD in neuronal cells. Because CMX-8933 upregulates SOD, maybe full-length EPN does too. We hypothesized that if CMX-8933 represents the receptor-binding domain of full-length EPN, that full-length EPN may show the same stimulatory effects as CMX-8933, and may upregulate SOD. Extracellular fluid (ECF) was prepared from goldfish brains, the traditional source for isolating EPN. Analysis of the ECF on protein gels demonstrated the presence of a complex protein pattern dominated by two bands at 37,000 and 31,000 daltons, the known sizes of EPN-â (glycosylated) and EPN-ã (non-glycosylated), respectively. Immunoblots performed with EPN antibody “Sheila” (directed against the C-terminal end of EPN, Shashoua and Moore, 1978) confirmed the identity of these two ECF bands as EPN. Cultured mouse Nb2a neuronal cells were treated in six independent experiments with 12 ìg/ml ECF protein for 5 hrs, and whole cell lysates were tested for levels of SOD by immunoblots. This ECF treatment of neuronal cells produced a mean 4-fold increase in SOD levels (p = 0.007), supporting our hypothesis. However, since ECF is a complex mixture, this data did not show which ECF component was resonsible for the SOD signal. Since ECF is known to contain CMX-8933 EPN cleavage product, which by itself can upregulate SOD, it is possible CMX-8933 was responsible for the signal, not full-length EPN. To address this issue, microdialysis was performed using an 8,000 dalton MWCO membrane to remove low MW components from the ECF (including CMX-8933, MW = 1149), leaving EPN-â (MW 37,000) and EPN-ã (MW 31,000) present in the dialyzed ECF. Four independent experiments indicated no significant difference (p = 0.116) between dialyzed versus non-dialyzed ECF for activating SOD. Thus, CMX-8933 does not appear to be responsible for ECF’s ability to increase SOD, but instead, high MW molecules (including full-length EPN) appear to be the active components. Altogether, the data from this thesis extends our knowledge of the mechanism of action of both full-length EPN and its cleavage product CMX-8933."


Worcester Polytechnic Institute

Degree Name



Biology & Biotechnology

Project Type


Date Accepted





ependymin, SOD, Neurotrophic functions, Ependymin