1288 Environmental Toxicology and Chemistry, Vol. 28, No. 6, pp. 1288–1303, 2009  2009 SETAC Printed in the USA 0730-7268/09 $12.00 + .00 TESTING AND APPLYING A FISH VITELLOGENESIS MODEL TO EVALUATE LABORATORY AND FIELD BIOMARKERS OF ENDOCRINE DISRUPTION IN ATLANTIC CROAKER (MICROPOGONIAS UNDULATUS) EXPOSED TO HYPOXIA CHERYL A. MURPHY,*† KENNETH A. ROSE,‡ M. SAYDUR RAHMAN,§ and PETER THOMAS§ †Department of Fisheries and Wildlife, Lyman Briggs College, Michigan State University, 13 Natural Resources Building, East Lansing, Michigan 48824, USA ‡Louisiana State University, Department of Oceanography and Coastal Sciences, Energy, Coast and Environment Building, Baton Rouge, Louisiana 70803, USA §The University of Texas at Austin, Marine Science Institute, 750 Channel View Drive, Port Aransas, Texas 78373, USA ( Received 1 July 2008; Accepted 22 December 2008) Abstract—Recently, hypoxia has been shown to act as an endocrine disruptor. We used a model of vitellogenesis in a female sciaenid fish to simulate the effects of hypoxia and to determine if reproductive impairment observed in field-caught fish could be attributed to dissolved oxygen conditions at the sampling sites. The model is a set of coupled, ordinary differential equations that simulate major biochemical reactions from the secretion of gonadotropin to production of vitellogenin. Various intermediatevariables in the model correspond to commonly measured biomarkers, and we assume a direct relationship between cumulative vitellogenin (VTG) and the gonadosomatic index (GSI). Model predictions were compared to results of laboratory studies that examined the effects of hypoxia on Atlantic croaker (Micropogonias undulatus) reproduction. When hypoxia was assumed to cause reduced gonadotropin and impaired aromatase activity, model predictions of VTG production were similar to laboratory-measured reductions in GSI. The model was then applied to reproductive biomarkers measured in fish from normoxic and hypoxic locations in Pensacola Bay (FL, USA). We simulated the relationship between reduced estradiol-17 and VTG production under hypoxia, and we compared these results with field data. Good agreement between field and simulation results suggested that croaker collected from hypoxic sites in October were exposed to hypoxic conditions for an extended period during gonadal recrudescence and that hypoxia was a dominant cause for the reduced GSIs. Monte Carlo uncertainty analyses suggested that the maximum rate of free testosterone production is the most sensitive parameter. Our simulations demonstrated that the model can be used identifying the mechanism underlying endocrine disruption and for interpreting field-measured biomarkers in situations of multiple stressors. Keywords—Hypoxia Atlantic croaker Biomarkers Simulation model Vitellogenesis INTRODUCTION Hypoxia, or low dissolved oxygen (DO), is becomingly increasingly widespread in estuarine and coastal waters. In North America, low-DO sites, or dead zones, occur on most coastlines; the most notable include Chesapeake Bay (USA) [1], the northern Gulf of Mexico [2], and the Oregon (USA) inner shelf [3]. Increased nutrient loading of nitrogen and phos- phorus from major rivers has contributed to the recent expan- sion of these dead zones [4]. The full economic and ecological consequences of in- creased hypoxia are not known for most ecosystems. Hypoxia clearly causes mortality in sessile organisms and displaces mobile organisms [3,5,6], but how these local effects translate into population-level responses is unknown except for a few well-studied systems [7]. Assessing the ecological effects of hypoxia on mobile organisms poses a special challenge, both because documenting their degree of exposure to spatially and temporally dynamic hypoxia is difficult [8] and because or- * To whom correspondence may be addressed (camurphy@msu.edu). Published on the Web 2/6/2009. Although the research described in this article has been funded wholly or in part by the U.S. Environmental Protection Agency, it has not been subjected to the Agency’s required peer and policy review and therefore does not necessarily reflect the views of the Agency, and no official endorsement should be inferred. This is publication 106 of the National Oceanic and Atmospheric Administration’s Center for Sponsored Coastal Ocean Research NGOMEX06 program. ganisms typically move to avoid low DO, with unknown con- sequences of these new locations on their growth and mortality [9]. In the Gulf of Mexico, the hypoxic area has been observed as early as February and as late as October [2,10], can extend horizontally from tens to thousands of kilometers, is usually restricted vertically to a few meters off the bottom but in shallower waters can occupy as much as 80% of the water column [11], and can last from days to weeks. Episodic hyp- oxia also occurs inshore in the estuaries [8]. Despite legislative mandate in the United States, quantifying the quality of the many alternative habitats potentially inhabited by coastal fish species remains an unresolved issue [12]. Most studies have focused on how low DO affects avoid- ance behavior, growth, and mortality [9,13,14], but several recent studies have documented that hypoxia acts as an en- docrine disruptor of fish reproduction, likely affecting the hy- pothalamus–pituitary–gonadal–liver (HPGL) system [15]. As with most endocrine disruptors, the observed hypoxic effects on adult fish are sublethal, resulting in reduced gonadotropin secretion, altered steroid levels, reduced vitellogenin, lowered fecundity, reduced gonadosomatic indices (GSIs), and skewed sex ratios [15–19]. Although the exact mechanisms by which hypoxia acts on the HPGL axis have not been ascertained, laboratory experi- ments have documented impaired gonadotropin secretion and gonadotropin-releasing hormone function [15] and also sug- gested reduced aromatase activity (caused by hypoxia) af-