Ferroxidase activity has been reported to be altered in various biological fluids in neurodegenerative disease, but the sources contributing to the altered activity are uncertain. Here we assay fractions of serum and cerebrospinal fluid with a newlyvalidated triplex ferroxidase assay. Our data indicate that while ceruloplasmin, a multicopper ferroxidase, is the predominant source of serum activity, activity in CSF predominantly derives from a <10kDa component, specifically from polyanions such as citrate and phosphate. We confirm that in human biological samples, ceruloplasmin activity in serum is decreased in AlzheimerÕs disease, but in CSF a reduction of activity in AlzheimerÕs disease originates from the polyanion component. Multicopper ferroxidase inhibition with sodium azide 12 verified that CP was the major component of retentate activity (Fig. 1B). However, despite all activity being ablated when using the Ferric Gain assay (Fig. 1Bi), a small proportion of retentate activity
Keywordswas not inhibited by azide when measured by TF loading (1.44 ± 0.13 µM Fe 3+ /min) ( Fig. 1Bii). This is consistent with previous TF loading reports identifying the presence of a non-CP ferroxidase activity in serum 11,17 . End point quantitation of Fe 2+ loss corroborated the production of Fe 3+ during the assays in both the presence and absence of TF (Fig. 1C).Triplex assay parameters for CSF were found to be as previously determined for purified CP 1 and human serum ( Fig. 1), at an optimal CSF volume of 20 µl (Fig. S2B). In contrast to serum fractions ( Fig. 1 In assaying the same volume (5µl) of serum and CSF in both the ÔFerric GainÕ and ÔTF loadingÕ components of the triplex assay, activity was markedly greater in serum (Fig. S3A). Upon fractionation of either biological fluid, retentate had markedly greater activity from serum ( Fig. S3B) and negligible activity from CSF ( Fig. 2A).Despite this greater percentage of filtrate activity in total CSF, when compared to the same serum volume this fractionÕs activity only had a minor elevation in Ferric Gain Azide (2.5mM) did not inhibit CSF activity (Fig. 2B), excluding the source of the activity being a multicopper ferroxidase e.g. CP. The absence of CP activity in human CSF is in alignment with normal levels of CP (~1.5 µg/ml) 18 being <1% of those ! 6 found in serum (200-500µg/ml) 19 . Consistent with the previously reported level of CP in CSF, recombinant CP at this concentration (~2.5 µg/ml) is outside the active range of our assay (<20nM; Fig. S4). Our findings are similar to the activity that has been reported to arise from <10kDa species in CSF from healthy and sporadic CreutzfeldtJakob disease (sCJD) patients 14 , but at variance with a report that ferroxidase activity in CSF originates from CP 13 .To identify whether the ferroxidase activity from the filtrate was of protein enzyme origin, fractionated serum and CSF were treated with proteinase K. In serum, the protease abolished the retentate activity, but had no effect on the filtrate activity ( Fig. 3A...