DIFFERENTIAL EFFECTS OF ARSENIC, CHROMIUM, CADMIUM, NICKEL AND MITOMYCIN C ON GENE EXPRESSION AS MEASURED BY DNA MICROARRAY.

Angeline S. Andrew, A.J. Warren, K.A. Temple, and Joshua W. Hamilton. Dartmouth Medical School, Hanover, NH, U.S.A.

Examining global effects of toxins on gene expression can be useful for examining patterns of biological response, underlying mechanisms of toxicity, and identification of candidate metal-specific genetic markers of exposure and response. Using gene microarrays, we examined changes in gene expression following low dose, acute exposures of Cd, Cr, Ni, As or mitomycin C (MMC) in BEAS-2B human bronchial epithelial cells. Treatments were chosen that did not cause any overt signs of toxicity, or changes in cell survival or replication as measured by long term colony forming assays. Total RNA was isolated from cells exposed to 3 m M Cd(II) (as cadmium chloride), 10 m M Cr(VI) (as sodium dichromate), 3 m g/cm² Ni(II) (as nickel subsulfide), 5 m M or 50 m M As(III) (as sodium arsenite), or 1 m M MMC for 4 h. Gene expression profiles in response to these treatments were measured using a Clontech Atlas Human 1.2 cDNA microarray. A signal intensity threshold ratio of 1.5 was used for Cd, Cr, Ni and MMC, but a higher threshold value of 1.7 was used for As due to the much larger number of As-affected genes. Surprisingly, only a small subset of genes was differentially expressed in response to each agent: Cd, Cr, Ni, As (5 m M), As (50 m M), and MMC each differentially altered the expression of 25, 44, 31, 110, 65, and 16 individual genes, respectively. Interestingly, only a few genes were commonly expressed among the various treatments. Only one gene was altered in response to all four metals (hsp90), and no gene overlapped among all five treatments. Only three to seven genes overlapped among any two treatments, and similarly only a few genes were common to any three treatments. We also compared low (5 m M, non-cytotoxic) and high dose (50 m M, completely cytotoxic) arsenic treatments, which surprisingly affected expression of almost completely non-overlapping subsets of genes, suggesting a threshold switch from a survival response to an apoptotic response (NIEHS (SBRP) ES07373).



PB ISOTPPES, ARSENIC SOURCES AND ENRICHMENT PATHWAYS LINKING SULFTDES FROM MINES AND UNMINERALISED ROCKS TO SECONDARY IRON OXIDES, COASTAL NEW ENGLAND.

Robert A. Ayuso¹, Nora K. Foley¹, Joseph D. Ayotte ², Ann Lyon¹, John Burns¹, Robert G. Marvinney ³, Andrew S. Reeve⁴, and Gilpin R. Robinson, Jr.<sup>1

1</sup>U.S. Geological Survey, Reston, VA 20192; ²Pembroke, NH 03275; ³Maine Geological Survey, Augusta, ME 04333; ⁴ Department of Geological Sciences, University of Maine, Orono, ME 04401.

Sulfide and secondary iron oxy-hydroxides minerals (n = 56) collected from mines and sulfidic meta-shales were analyzed for their Pb isotopic compositions and for trace elements in an effort to test the link between arsenic-rich sulfide minerals and secondary oxy-hydroxides presently forming along rock surfaces, joints, crevices, fractures, etc. This study is part of a detailed mineralogical and geochemical analysis of iron-sulfide minerals in natural bedrock (e.g., the sulfide-rich Cambrian-Ordovician Penobscot Formation) and drill cores in coastal Maine and New Hampshire. Weathering of pyrite, pyrrhotite and other sulfide minerals generates acid and releases metals (e.g., Pb, Cu, As, Co, Ni) that are then sequestered in secondary minerals (e.g., ferrihydrite, goethite, scorodite, jarosite, and natrojarosite, rosenite, and melanterite)². The iron oxy-hydroxide minerals constitute the ideal substrates for sorption reactions involving As, Pb, and other metals in solution. Our study shows that Pb isotopic compositions of the sulfides and iron oxy-hydroxides overlap and establish a genetic link between the sulfides and secondary minerals. Pb isotopic compositions were determined by thermal ionization mass spectrometry using bulk samples and acid-leached samples of galena, arsenian-pyrite, pyrrhotite, lollingite, cobaltite, and arsenopyrite. Pb isotopic compositions range as follows: ²⁰⁶Pb/ ²⁰⁴Pb = 18.073-19.489; ²⁰⁷Pb/²⁰⁴Pb = 15.539-15.675; ²⁰⁸Pb/²⁰⁴Pb = 37.947-39.102 and plot as nearly vertical fields on standard uranogenic and thorogenic Pb diagrams. Pb isotopic compositions of secondary minerals, including goethite, jarosite, and natrojarosite had the following ranges: ²⁰⁶Pb/²⁰⁴Pb = 18.356-21.945; ²⁰⁷Pb/²⁰⁴Pb = 15.595-15.839; ²⁰⁸Pb/ ²⁰⁴Pb = 38.169-39.162. The isotopic compositions plot on uranogenic and thorogenic Pb diagrams as broad and steep fields that extensively overlap the field of the sulfides but extend to more radiogenic compositions. The similarity of isotopic compositions provides evidence that fluid-mineral reactions leading to the decomposition of the sulfides released metals and As and imprinted the Pb isotopic signatures of the sulfides on the secondary minerals. As and Pb contents in the sulfides vary widely. For example, in the Penobscot Formation, pyrites range from ~20 ppm to ~2000 ppm As and ~10 ppm to >400 ppm Pb, but pyrites from the mine areas are substantially higher in As and Pb and range to thousands of ppm. Bedrock occurrences represented by bulk rock samples also have a wide range in As and Pb contents. For example, in the Northport area, Maine, where As in groundwater is elevated³ , As contents as high as 730 ppm have been found in rocks of the Penobscot Formation. As and Pb contents of the secondary minerals are also highly variable but characteristically closely match the ranges found in the sulfides and bulk rock samples. Although the isotopic compositions of the sulfides can account for most of the variations in the secondary minerals as a result of weathering, a more radiogenic Pb component could be present in the secondary minerals but its identity has not been determined precisely. Possible contributions from anthropogenic sources cannot be disregarded.



ARSENIC PLUMES WHERE THE "SOURCE" CONTAINS NO ARSENIC: THREE CASE STUDIES OF APPARENT DESORPTION OF NATURALLY OCCURRING ARSENIC.

Richard S. Behr and John E. Beane.

Maine Department of Environmental Protection, 17 State House Station, Augusta, ME 04333.

Arsenic is observed in groundwater at detectable concentrations throughout Maine. Natural arsenic concentrations sometimes exceed the 10 m /L state drinking water standard, but a variety of human activities can also create geochemical conditions capable of liberating arsenic from aquifer materials. We observed elevated arsenic in groundwater at three sites where groundwater chemistry was altered by the following three distinct activities: 1) a leaking underground gasoline storage tank; 2) a gravel pit reclaimed with manufactured topsoil; and 3) an attenuation landfill. The data from these sites demonstrate that non-arsenic contain-ing wastes can alter groundwater geochemistry and release naturally occurring arsenic from aquifer material. All three activities released biodegradable organic compounds that consumed dissolved oxygen, driving the groundwater anaerobic. Reduced forms of iron and arsenic are generally more soluble than the oxidized forms. Therefore, the development of anaerobic conditions could increase the solubility of both iron and arsenic and release both elements from the aquifer matrix. Groundwater samples were collected using standard low flow sampling procedures. Field analyses were conducted for dissolved oxygen (D.O.), specific conductance, and pH. Dissolved gasoline, arsenic and other inorganics were analyzed using standard laboratory methods. At the first site gasoline was released in a sandy aquifer. Gasoline, the primary contaminant, contains negligible arsenic. Samples were collected from an up-gradient monitoring well and three down-gradient recovery wells. The up-gradient well was strongly aerobic, with 7.1 mg/L D.O., no detectable dissolved gasoline and arsenic less than 3 m g/L. Biodegradation of hydrocarbons down-gradient of the source decreased the D.O. concentration, and increased dissolved iron and manganese. Iron and manganese decreased with distance from the source, but oxygen did not return. All samples within the dissolved gasoline plume contained measurable arsenic that ranged from 61 to 350 m g/L. Manufactured top-soil was used to reclaim a gravel pit. The topsoil consists of short paper fiber derived from a newsprint pulp and paper mill, and commercial fertilizer. Groundwater data from the downgradient monitoring wells indicated organic carbon that leached from the manufactured topsoil quickly depleted D.O. in groundwater beneath and downgradient of the reclaimed area. Iron and arsenic increased significantly with rapid deple-tion of D.O. The arsenic concentration in downgradient groundwater increased from less than 4.0 m g/L to 130 m g/L. Neither the short paper fiber nor the fertilizer contained much arsenic; therefore, it's likely the iron and arsenic are released from the aquifer matrix. An attenuation landfill containing caustic paper mill wastes contaminated a surficial aquifer. Contaminated groundwater contains very little D.O. As the pH increased to 10, arsenic concentrations increased to over 1000 m g/L. The increase in arsenic is not correlated with a similar increase in iron throughout the plume. This data suggests the increase in arsenic may result in part from the desorption of anionic arsenic at elevated pH. These data are consistent with the hypothesis that the decrease of redox potential due to biodegradation of dissolved organic compounds may liberate adsorbed arsenic with iron for transport in solution. An increase in pH can reduce iron oxides capacity to adsorb arsenic. At the landfill, the increase in arsenic as the pH increased indicates the higher pH may have substantially decreased arsenic adsorption.



ARSENIC IN BEDROCK WELLS IN CONNECTICUT.

Craig J. Brown¹ and Stewart K. Chute².

Water Resources Division ¹U.S. Geological Survey, 101 Pitkin St., East Hartford, CT 06108, ²Connecticut Department of Public Health, 410 Capitol Ave., Hartford, CT 06134.

Samples collected from private bedrock wells in two areas in Connecticut were analyzed to evaluate the relative importance of bedrock type and redox chemistry on the occurrence and mobility of arsenic in bedrock. Samples were collected from wells along transects in two areas where (1) bedrock is known to contain sulfide minerals, or (2) arsenic concentrations have been high (>10 m g/L) in water from bedrock wells. Samples also were collected from wells in adjacent bedrock types that were not expected to have a high arsenic content. Each area included 20 wells. Total and dissolved arsenic, redox-sensitive constituents (dissolved oxygen, iron, and sulfide), and major dissolved ions in the water from specific bedrock types were analyzed to help determine the sources and pathways of arsenic in bedrock wells. Dissolved arsenic concentrations were below the detection limit (< 0.18 m g/L or 0.9 m g/L) in 16 of the 20 wells in the towns of Colchester and East Hampton. Arsenic concentrations showed little or no difference in samples from wells in the Hebron gneiss and the Brimfield schist, but concentrations of dissolved iron, manganese, and sulfate were higher in the Brimfield schist. In northeastern Connecticut, where arsenic concentrations in public-supply wells in bedrock have historically been greater than 10 m g/L, eight wells had dissolved arsenic concentrations greater than the detection limit (0.18 m g/L or 3 m g/L). Six of these wells were in the Hebron gneiss near South Woodstock. Water samples from two of these wells had dissolved arsenic concentrations that exceeded the new U.S. Environmental Protection Agency arsenic standard (10 m g/L)—24 m g/L and 14 m g/ L, and total arsenic concentrations as high as 39 m g/L. Detected arsenic concentrations in ground water were not restricted to reducing conditions. Arsenic concentrations were highest in ground water with low concentrations of dissolved iron and manganese, and the well with the highest total arsenic concentration and the second highest dissolved arsenic concentration was oxic. Wells in the Hebron gneiss in Woodstock that had detected concentrations of dissolved sulfide, however, also had detected concentrations of dissolved arsenic and could indicate that arsenic was desorbed or reduced from ferric hydroxides under reducing conditions. Water from four of the wells in Woodstock with the highest arsenic concentrations had pH levels greater than 7.7 and somewhat higher bicarbonate; this indicates that arsenic occurrences could be related to the desorption of arsenate at high pH. The higher pH in wells in the Hebron gneiss in Woodstock compared to that in the Hebron gneiss in the Colchester-East Hampton area could result from higher concentrations of calcite, or greater extent of silicate weathering, in the Woodstock bedrock. The composition of the Hebron gneiss ranges from an interbedded quartz-biotite-plagioclase schist and calc-silicate gneiss in the Colchester area to a well-layered feldspathic biotite-quartz schist in the South Woodstock area; previous studies of the Hebron gneiss do not indicate the local presence of arsenic-bearing minerals, but mineralogical or whole-rock chemistry data are lacking. Small pegmatite intrusions and vein-filled fractures are com-mon throughout the Hebron gneiss and may be a source of arsenic-bearing minerals. Further study will be necessary to determine whether the high frequency of arsenic occurrences is from arsenic within the Hebron gneiss, from overlying glacial deposits, or from anthropogenic contamination, such as the application of pesticides that contain arsenic.



A CASE FOR BACKGROUND LEVELS OF ARSENIC IN GROUNDWATER AT THE MASSACHU-SETTS MILITARY RESERVATION.

Jay L. Clausen¹, Diane M. Curry¹, Joe Robb¹ , and William B. Gallagher².

¹AMEC Earth & Environmental Inc., 239 Littleton Road, Suite 1B, Westford, MA 01886, ²Impact Area Groundwater Study Program Office, PB 565/567 West Outer Road, Camp Edwards, MA 02542.

A comprehensive groundwater study is being conducted at Camp Edwards within the Massachusetts Military Reservation. A focus of the study is the Impact Area, where various types of ordinance, such as artillery and mortar rounds, were fired. Due to its presence in some ordnance items, arsenic is an analyte of interest. An objective of the groundwater and soil investigations was to determine if the presence of arsenic represented background conditions or an anthropogenic input. Groundwater samples were collected from monitoring wells and soil samples were collected from the surface and subsurface soil borings. These samples were analyzed for arsenic using Method IM40. The reporting limits for soil and groundwater are 1.1 mg/kg and 4.2 m g/L, respectively. The method detection limits are 1 mg/kg and 4 m g/L for soil and ground-water, respectively. Of the 540 unfiltered groundwater samples analyzed, arsenic was detected in 36 samples. One hundred thirteen filtered groundwater samples were analyzed and arsenic was detected in 16 of these samples. The mean arsenic concentration is 7.7 m g/L (both filtered and unfiltered results) with a maximum concentration of 53 m g/L reported. At the current arsenic maximum contaminant level (MCL) of 50 m g/L, 6.7 percent of the samples exceed this level and 8.1 percent exceed the recently promulgated MCL value of 10 m g/L. The majority of detections are associated with repeated elevated results from a single monitoring well. A relationship exists between pH, aluminum, and manganese levels and arsenic concentrations. Samples with elevated arsenic concentrations also exhibited elevated concentrations of aluminum and or manganese. There was a direct correlation between aluminum and manganese levels and the turbidity of the groundwater samples. Based on these observations, we hypothesize arsenic is adsorbed onto aluminum and manganese particulates. Acidification of highly turbid groundwater samples during sample preparation results in liberation of arsenic from the particulates. Analysis of the arsenic data using spatial and geostatistical techniques did not identify any trends in the distribution of arsenic suggestive of ground-water contamination. Spatial and geostatistical analysis of surface soil arsenic levels indicates no trends suggestive of anthropogenic inputs from training activities. A comparison of surface soil arsenic levels from the Impact Area to background levels reveals no statistical difference. Fate-and-transport modeling of the vadose zone indicates arsenic mobility is insufficient to explain elevated arsenic groundwater levels. Therefore, the levels of arsenic in groundwater beneath the Impact Area at Camp Edwards are reflective of natural background groundwater geochemistry. There is no evidence the presence of arsenic in groundwater is a result of training activities at Camp Edwards.



A PILOT STUDY OF ARSENIC SPECIATION IN DOMESTIC WELL-WATER SUPPLIES IN MAINE.

Charles W. Culbertson¹, Deborah M. Moll², Lorraine C. Backer², Mary L. Gilbertson³ and Andrew E. Smith ³.

¹U.S. Geological Survey, Maine District, Augusta, ME 04330, ²Centers For Disease Control and Prevention, Atlanta, GA 30333 and ³Maine Bureau of Health, Augusta, ME 04330.

The U.S. Geological Survey, in cooperation with the Centers for Disease Control and Prevention (CDC) and the Maine Bureau of Health (MBH) participated in a pilot study to investigate geochemical controls on arsenic (As) occurrence and speciation in selected domestic water supplies in Maine. A second objective of the study was to investigate the effect of arsenic species on the efficiency of arsenic removal by household water treatment systems. This study was undertaken to augment a larger CDC/MBH investigation into epidemiology and arsenic exposure through household water in Maine. Houses were selected based on known high arsenic concentrations, whether or not treatment systems were installed, and known difficulty in effective arsenic removal by the treatment systems. Well-water samples were collected at point-of-entry, prior to pressure tanks and household water treatment systems. Temperature, specific conductance, pH, and dissolved oxygen (DO) were measured on site using water quality sensors in a flow- through chamber. Geochemical constituents, including total iron, ferrous iron, total manganese, reactive phosphorous and sulfide, were determined immediately upon sample collection using standard colorimetric methods. Samples for total and dissolved (0.45-mm filtrate) As, As(III) and As(V) were collected and prepared on site for later laboratory analysis by inductively coupled plasma-mass spectrometry (ICP-MS). Sample preparation included acidification to pH 2 for both total and filtered samples and filtration through 0.45-mm syringe filters for appropriate samples. An aliquot of the filtered/acidified sample was passed through an anion exchange resin column, which retains As(V) while allowing As(III) to pass through. As(V) was determined by difference between total dissolved As and As(III). Particulate As was determined by difference between total and dissolved As. Dissolved As concentrations ranged from 5 to 400 mg l^-1 with generally less than 10% occurring in the particulate phase. One exception (> 25% particulate As), co-occurred with iron concentrations that were 10 to 50 times higher than those measured at the other sites. No relationship was found between As concentrations and well depth (all wells, except one, were >200 ft). Variability in As concentrations within regional clusters was as large as that observed between the clusters. More than half of the sites were characterized by >60% As(III), the species most mobile and difficult to remove; one third were >90%. The percentage of As(III) decreased exponentially as DO increased, consistent with expected redox/speciation dependence. Arsenic concentration and speciation were not correlated with any of the measured geochemical parameters. Households investigated in this study employed several technologies for removing arsenic from their water supplies. As(III) clearly posed a problem for reverse-osmosis treatment alone, however a combination of technologies achieved >90% removal in these cases. One system, based on an adsorption technology, achieved >90% As(III) removal.



REDOX CONTROLS ON ARSENIC MOBILITY BENEATH WINTHROP LANDFILL, MAINE.

Saugata Dattalt², Alison R. Keimowitz¹, H. James Simpson ¹, Martin Stute^1,2, Steven Chillrud¹, Monique Tsang³, Yan Zheng^1,4, Alexander van Geen¹ and Greg M. Dobbs⁵.

¹Lamont Doherty Earth Observatory, Columbia University, 61 Route 9W, Palisades, NY 10964, ²Department of Environmental Science, Barnard College, 3009 Broadway, New York, NY 10027, ³Department of Geology, Bryn Mawr College, 10 IN Merion Ave., Bryn Mawr, PA 19010, ⁴School of Earth and Environmental Sciences, Queens College, C.U.N.Y, 65-30 Kissenna Blvd., Flushing, NY 11365, ⁵United Technologies Research Center, UTC, East Hartford, CT 06108.

The groundwater beneath a landfill near Winthrop, Maine shows elevated levels of dissolved arsenic (~300 m g/kg). Evidence suggests that the solid phase arsenic is natural and becomes mobilized by reducing conditions imposed by the decomposition of organic matter within the landfill. Based on the working hypothesis that oxidizing conditions might precipitate dissolved arsenic and iron from the groundwater, a pilot field experiment was conducted in which a commercial magnesium peroxide mixture was injected to assess the feasibility of oxidizing controlled segments of the aquifer. No systematic change in dissolved oxygen, arsenic or iron were observed in downstream monitoring wells. Bench-scale experiments, however, demonstrated that oxidation of groundwater does induce precipitation of dissolved iron and arsenic, given sufficient oxidizing inputs. The distribution and sources of chemical oxygen demand (COD) of sediments from this aquifer are being examined to better quantify cumulative oxygen demand imposed by the solid phase of the aquifer and further explore the feasibility of using commercial magnesium peroxide mixtures with respect to remediation of elevated [As] in groundwater. Samples from more than fifty core segments have been separated into five size fractions (A>2000 (m m, 2000 m m>B>410 m m, 410 m m>C>157 m m, 157 m m> D>60 m m and 60 m m>E). Among the five components, the two largest particle fractions A and B comprise most of the mass in most wells and also at the majority of depths in each well. The two finest grain-size fractions impose the highest oxygen demand per gram. For fraction D, the COD per gram has little variation, while much larger variation of COD per gram in fraction E is seen among the wells. The sources and magnitude of COD are clearly seen to be complicated and difficult to quantify in an aquifer with heterogeneous paniculate phases.



MINERALOGICAL PATHWAYS FOR ARSENIC IN WEATHERING META-SHALES: AN ANALYSIS OF REGIONAL AND SITE STUDIES IN THE NORTHERN APPALACHIANS.

Nora K. Foley¹, Robert A. Ayuso¹, Joseph D. Ayotte ², Robert G. Marvinney³, Andrew S. Reeve⁴, and Gilpin R. Robinson, Jr.<sup>1

1</sup>U.S. Geological Survey, Reston, VA 20192, ²Pembroke, NH 03275; ³Maine Geological Survey, Augusta, ME 04333; ⁴ Department of Geological Sciences, University of Maine, Orono, ME 04401.

Concern about arsenic-bearing groundwaters in New England has caused examination of possible sources in the local bedrock. Detailed mineralogical analyses of iron-sulfides from over 70 bedrock localities, including 22 within the regionally extensive and sulfide-mineral-rich Penobscot Formation and 10 associated with mineral deposits from coastal New Hampshire and Maine, coupled with data from drill core collected at several sites including areas where well waters contain anomalous As abundances (e.g., Northport, ME), establish a diversity of primary and secondary mineralogical hosts for arsenic in bedrock. Reactions involving As-minerals and either groundwaters at low pH or in bicarbonate fluids at near-neutral pH probably control arsenic contents in groundwater in the region. Bedrock mineralogy is critical to contributing arsenic to groundwater and suggests a number of possible mineralogical bounds on the pathways for arsenic that help define weathering processes. Primary arsenic-bearing minerals identified include pyrite (max. 4 wt.% As in FeS₂), pyrrhotite (max. 0.5 wt.% As in FeS_1-x), lollingite, realgar (?), cobaltite, arsenopyrite, cobaltian-arsenopyrite (max. 8.4 wt.% Co), and tennantite. Supergene minerals that constitute intermediate mineralogical sources include orpiment and arsenolite-like minerals, Co-Ni-arsenates (?), Ca-arsenates (rauenthalite, phaunouxite?), scorodite (FeAsO₄ 2H₂ 0) and secondary arsenopyrite, pyrite, and marcasite.² Pyrite, the most abundant iron-sulfide mineral in many of the rocks, is a primary host for As in low-grade mineral deposits (e.g., volcanic-associated massive sulfides, metamorphic-Au, and Carlin-Au deposit types). In meta-shales, coexisting pyrrhotite, cobaltite, and arsenopyrite constitute a probable source for high As contents (e.g., Penobscot Fm.). Weathering of pyrrhotite in the Penobscot Fm. results in (1) com-plex mixtures of pyrite + marcasite, and (2) iron oxy-hydroxides and secondary salts, such as ferrihydrite, rozenite and melanterite. Weathering of pyrite, lollingite, realgar (?), and arsenopyrite and other sulfide minerals in these settings causes the production of acid and release of trace metals, including As, Co, Ni, Pb, etc., which can sorb on iron oxy-hydroxide substrates. An alternate pathway to consider is the oxida-tion of arsenopyrite or other As-bearing minerals to produce iron oxides and release sulfur and arsenic, which under specific conditions may produce arsenolite or orpiment. Subsequent leaching of the goethite + arsenolite or orpiment assemblages by bicarbonate-bearing fluids could release As into the groundwater system. Lollingite occurring at mineralized localities oxidizes to scorodite and some iron oxy-hydroxides products. Ca-arsenates are thought to form at some sites, possibly by the reaction of acidic, As-bearing waters with calc-silicate substrates. The presence of Ca-arsenates also suggests a process whereby (1) As is liberated from bedrock by direct interaction between anaerobic HC0₃- groundwaters and As-minerals³ and (2) is subsequently re-precipitated at low pH. When solubility is controlled by Ca-arsenate, Ca in solution suppresses the solubility of arsenic, however, the long-term ability of these minerals to sequester As is untested.



ARSENIC REMEDIATION OF DRINKING WATER IN NEW ENGLAND: POINT-OF-USE (POU) and POINT-OF-ENTRY (POE) OPTIONS USING ADSORPTION TECHNOLOGY.

Gregory C. Gilles.

Apyron Technologies, Inc. 4030-F Pleasantdale Road, Atlanta Georgia 30340.

In November of this past year, the U.S Environmental Protection Agency (USEPA) lowered the existing drinking water standard for arsenic from 50 parts per billion (ppb) to 10 ppb. It is estimated that as many as 56 million people in the U.S. currently drink water containing unsafe levels of arsenic. Citizens living in New England are particularly affected. Over 4,000 small and large public water systems, and thousands of other small systems serving day care centers, trailer parks, schools, hotels, and restaurants will be impacted. It is estimated that over 500,000 private residences (which are largely unregulated) are also exhibiting high arsenic levels. Data from selected states in New England estimate that as high as 30% of the private wells may exhibit arsenic over 10 ppb. As more extensive testing is conducted, this figure is expected to rise. These entities are seeking solutions for remediation. New advances in arsenic adsorption-based technology are emerging as one of the best available technologies to meet this growing need. This presentation discusses some of the conventional and new adsorption technologies to remediate arsenic from drinking water. Over the past three years, Apyron Technologies, Inc. has developed, produced, and successfully implemented several metal oxide-based inorganic medias to address both Arsenic (III) and Arsenic (V) in drinking water. Patented oxides of iron and or manganese have been utilized in small systems, within POU cartridges for under-counter use, and within POE systems for whole household treatment. Over 30 such systems have been deployed throughout New Hampshire, Maine, and Massachusetts successfully reducing arsenic from as high as 970 ppb to below 10 ppb under a wide variety of water quality. Unlike conventional adsorbents such as activated alumina, some advanced technologies are capable of reducing both ionic arsenic (V) and uncharged arsenic (III) without pre-oxidation. In addition to arsenic, these specialty medias have demonstrated affinity for reducing other heavy metals such as lead, copper, antimony, chromium, and selenium. This presentation will also include the latest results from an ongoing arsenic treatment demonstration project in Standish, Maine where the technology has been recently deployed to remediate high levels of arsenic in residential drinking water.



THE NATURAL OCCURRENCE OF ARSENIC IN GROUNDWATER AT THE COMBUSTION ENGI-NEERING SITE IN WINDSOR, CONNECTICUT.

Nadia S. Glucksberg¹, Nelson M. Breton¹, Hank Andolsek ², and Elaine M. Hammick³.

¹Harding ESE, Inc., A MACTEC Company, Portland, ME 04042, ²Maine Department of Environmental Protection, Augusta, ME 04333, ³Combustion Engineering Windsor, CT 06095.

Arsenic has been identified in groundwater at concentrations exceeding the Connecticut Department of Environmental Protection (CTDEP) Remedial Standard Regulations (RSR) Criteria for Surface Water Protection of 4 micrograms per liter (m g/L) at the Combustion Engineering, Inc. (CE) Site in Windsor, Connecticut. Over 130 groundwater monitoring wells have been installed over 600 acres to investigate 27 areas of concern (AOCs) under the RCRA Voluntary Corrective Action Program. Four rounds of groundwater samples have been collected using low flow sampling techniques. During each sampling round, arsenic has been detected at different locations at concentrations that range from non-detect to exceeding the RSR Criteria for Surface Water Protection. Although arsenic has been detected above the RSR Industrial/Commercial Direct Exposure Criteria in soils at two AOCs, the pattern of detection of arsenic in groundwater does not suggest a potential Site-related source, but rather that the spatial and temporal variations are most likely due to the natural conditions within the aquifer. To understand the distribution of arsenic in groundwater, statistical comparisons were made to turbidity, Eh, total volatile organic concentrations (VOCs). None of the comparisons provided a significant correlation to the distribution of arsenic in groundwater. Additional evaluations were then conducted to compare the distribution of arsenic in groundwater to the distribution of arsenic in soils. The mean and maximum concentration of arsenic in soils were compared to the mean concentrations in all groundwater samples for each area of concern (AOC), where arsenic data are available for both media. The soil data set consisted of samples from 0 to 15 feet below ground surface. There was no correlation between elevated arsenic concentrations in soils and arsenic concentrations in groundwater. The lack of a clear link between arsenic concentrations in soils and groundwater suggests that there is not a significant source of arsenic present in shallow soils. Arsenic is not particularly mobile and all potential sources at the Site are present in surface soils. Furthermore, if a shallow arsenic source were to impact underlying groundwater, the shallow concentrations would have to be significantly higher than background concentrations to impact the underlying groundwater. None of the data indicate a relationship between a potential source area and concentrations above the CTDEP RSR Criteria for Surface Water Protection. Given the distribution pattern, concentration range of arsenic in groundwater, and the poor correlations between groundwater conditions and soil concentrations, a reasonable conclusion is that the arsenic is not Site-related but naturally occurring.



COMPARISON OF TWO ARSENIC EXPOSURE ASSESSMENT PROTOCOLS IN A CHRONICALLY EXPOSED POPULATION.

Edward E. Hudgens¹, Dina M. Schreinemachers¹, David J. Thomas², X. Chris Le³, and Rebecca L. Calderon ¹.

¹Epidemiology and Biomarkers Branch, and ²Pharmacokinetics Branch, National Health and Environmental Effects Research Laboratory / USEPA, Research Triangle Park, NC 27711, and the ³University of Alberta, Edmonton, Alberta, Canada T6G 2G3.

Consistent with the USEPA 1997 Arsenic Research Plan's emphasis on studies in US populations to obtain data to support a revised Maximum Contaminant Level (MCL) for arsenic, two studies were conducted in Millard County, Utah; the first in 1997 and a second in 1999. This location was chosen because the arsenic levels measured in drinking water supplies ranged from 2 Fg/l to 650 Fg/l. A total of 96 nonsmoking, non-drinking individuals from 28 families participated in the 1997 study. A questionnaire requesting demo-graphic information, exposure to other environmental pollutants, and medical history was completed for each study subject. First morning void urine samples were collected for five consecutive days and all voids were collected separately over one 24-hour period to assess the intra- and inter-individual variability in arsenic output. The 24-hour time course samples confirmed that the first morning void was a representative sample when corrected by creatinine. A positive relationship was found between total urinary arsenic and the arsenic concentration in the water source. Measurement of arsenite, arsenate, monomethylarsonic acid (MMA), and dimethylarsinic acid (DMA) in urine provided direct measures of an individuals exposure and metabolism. These individuals showed stable metabolite profiles consistent with chronic exposure as described in other studies in the U.S.: As(III) =11.1" 4.3%, As(V) = 3.9" 3.6%, MMA = 14.0" 4.1%, DMA = 71.0" 7.9%. The results of a mixed model regression analysis of the 1997 data showed that it was possible to reduce the number of urine samples required to give a good estimate of the extent of exposure to arsenic. To validate this result, 16 subjects from the 1997 study were asked to give only two urine samples on consecutive days in April 1999. The metabolite profile for the four arsenic species above was consistent with the profile found in the earlier study, even though the arsenic concentration of some drinking water sources had dropped to 50 percent of their 1997 levels. A similar relationship between the total urinary arsenic and the concentration of inorganic arsenic in the drinking water source was found. Reducing the number of required samples and removing the restrictions on the timing of sample collections did not affect the exposure assessment for this group of chronically exposed individuals. Larger, potentially more powerful, population based studies can be done for a reasonable cost through the use of this shorter protocol.



ESTIMATING RESIDENTIAL EXPOSURE TO DRINKING WATER ARSENIC IN INNER MONGO-LIA, CHINA FOR EPIDEMIOLOGIC STUDIES.

Richard Kwok¹, Pauline Mendola¹, Zhixiong Ning ², Zhiyi Liu², and Judy Mumford¹.

¹Epidemiology and Biomarkers Branch, Human Studies Division, NHEERL, USEPA, RTP, NC 27711, ²Institute of Endemic Disease for Prevention and Treatment, Inner Mongolia, China.

In the Ba Men region of Inner Mongolia, China, a high prevalence of chronic arsenism has been reported in earlier studies. A survey of the water supply system was conducted between 1991-1998 to better character-ize the arsenic (As) concentrations in the drinking water supply of local villages. A total of 14,866 wells were analyzed for their As content. Colorimetry based on silver diethyldithiocarbamate, an adaptation of the mercury bromide stain technique, and atomic absorption spectroscopy were used to determine the As content of the water supply. As concentrations ranged from below the limit of detection to 1.2 mg/1. Elevated concentrations were related to well depth (maximum at the 15 to 25 meter (m) category), well type (most high concentrations associated with the small household pump wells) and the date the well was built (peaks from 1980-1990). Over 43,600 persons consumed water with As concentrations above 0.01 mg/1 (14,500 above 0.05 mg/1, 480 above 0.5 mg/1). There were significant differences between different counties and villages within each county. Methods used to assign individual level exposure information based on aggregate exposure data, their advantages and disadvantages will be discussed. The presented database of As in wells of the Ba Men region provides a useful tool for planning future water explorations when combined with geological information. It also helps in the design of upcoming epidemiological studies on the effects of arsenic in drinking water.



ARABIDOPSIS MUTANTS EXHIBITING INCREASED TOLERANCE TO ARSENATE.

David A. Lee, Alice Chen, and Julian I. Schroeder.

Division of Biology, University of California, San Diego, La Jolla, CA 92093-0116.

One proposed approach towards the remediation of arsenic is phytoremediation, the use of plants to remove and detoxify arsenic from contaminated sites. While native plants have been identified in contaminated regions with increased tolerance to toxic metals, the genetic and molecular mechanisms which confer arsenic tolerance remain largely unknown. To elucidate some of the mechanisms involved in arsenic detox-ification, we developed a genetic screen using the model plant Arabidopsis thaliana. From this screen we identified a number of mutants which exhibit a significantly increased ability to grow in the presence of toxic arsenate concentrations. The strongest of these mutants, ars1, can grow on levels of arsenate which completely inhibit growth of wild type seeds, ars1 accumulates as much arsenic at the whole plant level as compared to wild type plants, suggesting that arsi plants have an increased ability to detoxify arsenate. Phytochelatins, small metal binding peptides, are currently believed to be the primary mechanism of arsenic detoxification in plants. However, ars1 produce phytochelatin levels similar to wild type plants, and the mutation does not map to the known phytochelatin synthase genes. Furthermore, ars1 plants do not show resistance to arsenite or other toxic metals such as cadmium and chromium. These data suggest that Ars1 functions upstream of arsenite chelation by phytochelatins. Progress at the genetic, physiological and bio-chemical characterization of ars1 will be presented, along with models suggesting altered arsenate biotransformation could be responsible for the ars1 phenotype.



ARSENATE REDUCTION BY ANAEROBIC SEDIMENT ISOLATES.

Anbo Liu, Elizabeth Garcia-Dominguez, E. Danielle Rhine, and Lily Y. Young.

Biotechnology Center for Agriculture and the Environment, Cook College, Rutgers University, New Brunswick NJ 08901.

The ability to grow at the expense of arsenate (As(V)) reduction to arsenite (As(III)) has recently been reported in several strains (Oremland et. al., in press). In order to expand the diversity of environmental isolates able to reduce As(V) a range of substrates was tested; these included acetate, succinate, lactate, and aromatic compounds such as syringic acid, ferulic acid, phenol and benzoate. Enrichment cultures were established using Onondaga Lake (OL) (NY) and Arthur Kill (AK) (NY/NJ harbor) sediment. In OL samples all the substrates supported the microbial reduction of arsenate to arsenite in the enrichment cultures. The rate, extent and stoichiometry of As(V) reduction were determined for the aromatic compounds. The relative rates of arsenate reduction for the different substrate are in the order of acetate/succinate > lactate > syringic acid >ferulic acid > phenol/benzoate. In AK samples lactate, syringic acid, and phenol/benzoate supported arsenate reduction;. Several pure cultures were isolated and are being phylogenetically characterized. Preliminary results show that one of the arsenate reducers isolated on lactate is also able to reduce nitrate, but cannot reduce sulfate. Activity of microorganisms in anoxic habitats appears to be widely distributed and may be important in arsenic biogeochemistry.



MICROBIAL ARSENATE REDUCTION IN ANAEROBIC GROUNDWATER.

Kevin A. McCaffery.

Department of Civil and Environmental Engineering, University of Maine, Orono, 102 Boardman Hall, Orono, ME 04469.

Arsenic is the only carcinogen for which a causal link between exposure through drinking water and human cancer has been established (USEPA, 1998). The microbial mechanisms that affect arsenic mobility are not well understood. It is known, however, that speciation affects adsorption and mobility characteristics, where As (III), the reduced form, is much more mobile and bioavailable than the oxidized form, As (V). In groundwater environments, where reducing conditions are normally present, microorganisms may catalyze the reduction of As (V) to As (III), and through this process harness energy. Microorganisms may also con-tribute to the release of arsenic through reduction of Fe (III) and Mn (IV), which bind As. These processes would result in increased soluble As concentrations and could lead to the contamination of groundwater and drinking water supplies. The purpose of this investigation is to determine if there are microorganisms in the groundwater environment that take part in arsenic transformations, or indirectly contribute to the release of arsenic from the parent material through the reduction of iron or manganese. Contaminated wells were sampled in June of 2001. Experiments are underway to establish rates of As (V) reduction, and subsequent mobilization of As from the geologic matrix. Microorganisms that can effectively reduce arsenic will be isolated for further characterization of these processes. To simulate the effect of increased organic loading on water quality, the effects of organic carbon enrichment (lactate addition) on the reduction and dissolution reactions related to arsenic are also being investigated.



ARSENIC IN GROUNDWATER IN MICHIGAN: STANDARDIZED MORTALITY RATIO ANALYSIS AND DEVELOPMENT OF A SPACE-TIME INFORMATION SYSTEM.

Jaymie R. Meliker¹, Jerome O. Nriagu¹, Robert Wahl ², Pierre Goovaerts³, and Geoffrey M. Jacquez⁴ .

¹University of Michigan School of Public Health, ² Michigan Department of Community Health, ³University of Michigan School of Engineering, ⁴Biomedware Inc., Ann Arbor, MI 48109-2029.

Reported arsenic concentrations in well waters of 11 contiguous Michigan counties range from 1 to 1310 m g/L, with most common levels being 5-50 m g/L. To investigate the health outcomes of this arsenic exposure, a standardized mortality ratio (SMR) analysis was performed and an enhanced arsenic exposure model is being developed. To perform the SMR analysis, Michigan Resident Death Files data were com-piled from 33 underlying causes potentially associated with arsenic exposure from 1979 through 1997. The Michigan Department of Environmental Quality supplied data from 4317 water tests conducted 1993 through 1996. SMRs were calculated using observed and expected numbers of deaths for each underlying cause for each county. Only Genesee, Huron, and Lapeer Counties had average arsenic concentrations above 10 ppb. Using Bernoulli confidence intervals to account for multiple testing, all three counties had statistically significantly (<0.00015) positive SMRs for cerebro vascular disease, Huron and Genesee Coun-ties had significantly positive SMRs for ischemic heart disease, and Lapeer and Genesee Counties has significantly positive SMRs for diabetes mellitus, and kidney disease. To accurately characterize exposure to low-to-moderate levels of naturally-occurring arsenic in drinking water in Michigan, a geostatistical groundwater model and spatio-temporal analyses are being incorporated into the construction of arsenic exposure scenarios. The spatio-temporal analyses will address the spatial and temporal variation in both arsenic concentration and daily activity patterns. To account for these different types of spatial and temporal variability, the project consists of three main components: personal interview, measurement of arsenic in drinking water, and the construction of exposure scenarios. Subjects will be long-term residents of eleven counties in Michigan with highest levels of arsenic in their groundwater and part of a case-control study, designed to evaluate the association between arsenic exposure and bladder cancer. Structured personal interviews will be administered to obtain information on exposure and health outcomes. A geostatis-tical groundwater model is being developed to predict water concentrations at past workplaces and past residences. Exposure scenarios will be generated using information provided in the interview, the measured arsenic concentration, the geostatistical groundwater model, and exposure factors. Current efforts by the USEPA to reduce the maximum contaminant level for arsenic in drinking water have been bedeviled by contradictory and un validated predictions of the risks of chronic exposure to low levels (< 100 m g/L) of arsenic in water. The SMR study suggests that future epidemiologic studies (with less confounding than an SMR study) should be conducted to investigate the association between low levels of arsenic ingestion and cerebrovascular disease, ischemic heart diseases, diabetes mellitus, and kidney disease. The development of an arsenic exposure model is designed to shed some light on the dose-response relations for exposure of the U.S. population to arsenic concentrations in the 5-100 m g/L range where little information currently exists.



EFFECT OF ARSENICALS ON CELL CYCLE DISTRIBUTION AND EXPRESSION OF CELL CYCLE PROTEINS IN HUMAN PRIMARY KERATINOCYTES.

Anuradha Mudipalli, R. Julian Preston, and James C. Fuscoe¹.

Environmental Carcinogenesis Division, NHEERL, ORD, USEPA, Research Triangle Park, NC 27711; ¹Division of Genetic and Reproductive Toxicology, National Center for Toxicological Research, USFDA, Jefferson, AR 72079.

Environmental exposure to arsenic is a major public health concern. Epidemiological studies have demonstrated a strong correlation between levels of arsenic in drinking water and incidence of cancers of skin, lung, bladder and peripheral and cerebro vascular diseases. Despite enormous efforts to understand the biological effects of arsenic, the specific mechanism(s) of action for cancer development are very poorly understood. The role of confounding factors such as UV adds another dimension to the study of arsenic carcinogenesis and risk assessment. Here we hypothesize that arsenicals may cause an override of cell cycle arrest caused by a DNA damaging agent such as UV, and promote the proliferation of unrepaired cells. An in vitro human primary keratinocyte model was developed to study the initial cell cycle events pertaining to the initiation and propagation of cells leading to cancer. Cells were treated with a single dose of UVB (100mJ/cm²) and then exposed to various concentrations of arsenite (iAs; 0-12uM) and two of the trivalent methyl derivatives, methyl oxoarsenie (MAsIII; 0-2uM), and iododimethyl arsine (DMAsIII;0-3uM), for 48h. Cell proliferation indices and cell cycle distributions were determined by MTT (3- [4,5-dim-ethyl thiazol -2-yl] 2, 5- diphenyl- tetrazolium bromide) and flow cytometric analyses, respectively. Additional controls include treatment with arsenicals but not UV, UV alone, and untreated cells. At 48h arsenicals were found to significantly increase the cell proliferation in a concentration dependent manner in UV exposed cells (DMAsIII> MasIII> iAs) as compared to UV exposed controls. Flow cytometric analyses revealed differential effects on cell cycle distribution. These data suggest that arsenicals are capable of overriding the cell cycle arrest caused by UV-induced DNA damage. Preliminary studies on the expression of cell cycle specific proteins by western blot analysis indicated significant differences in the expression of Cyclin Dl, CdK5 and PCNA, that were dependent on UV exposure and/or methylation status of arsenic (with several fold increase in PCNA and decrease in Cdk5 expression in UV-exposed cells treated with methylated arsenicals). These results indicate that arsenicals are capable of overriding the cell cycle arrest caused by UV damage by the up or down regulation of specific cell cycle proteins. Replication on a damaged template may result in mutations that initiate carcinogenesis associated with As exposure. Further studies utilizing this in vitro cell culture model may aid in elucidating the molecular mechanisms involved in arsenic carcinogenic.



EXPOSURE TO ARSENIC VIA BATHING AND OTHER CONTACT IN HOUSEHOLDS THAT USE BOTTLED WATER OR POINT-OF-USE TREATMENT DEVICES FOR DRINKING WATER.

Chris A. Paulu¹, Deborah M. Moll², Lorraine C. Backer ², Raquel I. Sabogal², Robert L. Jones³, Mary L. Gilbertson¹, and Andrew E. Smith¹.

¹Environmental Toxicology Program, Bureau of Health, Department of Human Services, State of Maine, Augusta, ME 04333; ²Division of Environmental Hazards and Health Effects; ³Division of Laboratory Sciences, National Center for Environmental Health, Centers for Disease Control, Atlanta, GA 30333.

For users of private well water high in arsenic, the usual remedy has been to use bottled water or point-of-use treatment devices for drinking water. While this effectively removes the primary means of exposure to arsenic in water, it does not necessarily mitigate secondary exposures via bathing, cooking, teeth brushing, and occasional drinking from unprotected taps. The assumption that these secondary exposures result in no or minimal arsenic intake needs to be tested, especially for the sub-population of young children, who are likely to spend more time in baths and engage in play-related ingestion of untreated water. To measure the extent of arsenic intake among adults and children who live in homes with high-arsenic well water, but use low-arsenic alternatives for drinking. Information and reply cards were mailed to prospective subjects from two sources: private homes with arsenic water tests _50 m ig/1 in the Maine Health and Environmental Testing Laboratory's database, and citizens who contacted the State Toxicologist regarding arsenic in their well water. Those respondents willing to participate were screened by telephone to determine that they used bottled water or a point-of-use treatment device (usually reverse osmosis) for drinking water. Consenting participants were asked to refrain from eating seafood (high in organic forms of arsenic) for three days, and to complete a diary of diet and bathing during that time. They were asked to collect toenail clippings on the third day, and a first-morning void urine specimen on the fourth day. We then collected the urine and toenail specimens, diary, and samples of drinking and bathing water, and administered a survey covering water use habits, water system configuration, and potential arsenic exposures from non-water sources. We analyzed urine for total arsenic and creatinine concentrations. We analyzed drinking and bath water samples for inorganic arsenic concentration. This study is ongoing and in the middle of its data collection phase. In the pilot phase of the study, we analyzed urinary arsenic, and arsenic in bath and drinking water, for 19 children under 6 and 22 subjects 6 or older. After excluding subjects whose drinking water contained arsenic >5 m g/l or who had eaten seafood within three days prior to urine collection, we divided subjects into two groups: those with bath water arsenic _100 m g/1 (n=16) and those with bath water arsenic <100 m g/l (n=15). We found the mean urinary arsenic level was 17.1 m g/1 for children in the higher arsenic group (n=7) compared to 8.4 m g/l for children in the lower arsenic group (n=8) (p=0.02 by Mann-Whitney U test). The difference was less pronounced when urinary arsenic was adjusted for density: 19.7 m g arsenic per g creatinine for the higher arsenic bath water group compared to 15.4 m g/g for the lower arsenic bath water group (p=0.15). The same comparisons of urinary arsenic for subjects 6 years of age were within 2 m g/1 and 2 m g/g creatinine. Further analysis of dietary factors and of more subjects is necessary to determine if there is an association between bath water arsenic and urinary arsenic among young children.



REMOVAL OF ARSENIC SPECIES BY FOAM FLOTATION.

Zhonghua Pan, Lei Zhang, and P. Somasundaran.

NSF-Industry/University Cooperative Research Center for Advanced Studies in Novel Surfactants, Langmuir Center for Colloids and Interfaces, Columbia University, New York, NY 10027, U.S.A.

Foam flotation, a relatively inexpensive technique, has been employed for enhanced removal of arsenic species from the solution. The method involves interactions between arsenic species or arsenic-hydrolys-able multi-cation complexes and surfactants, and removal of the hydrophobic complexes formed by air bubbles under quiescent conditions. The system in our experiments consisted of Arsenic (V) oxide hydrate, aluminum chloride, sodium dodecyl sulfate (CH₃(CH₂)₁₁OSO₃Na), dodecylamine (CH₃(CH₂)₁₁NH₂), and frothers. It was found that as much as 99% removal of arsenic species can be obtained by adsorbing colloid flotation, the removal being depended upon the solution pH, the ratio of arsenic to aluminum chloride and the interactions between the arsenic-multication species and the surfactants used as collectors. The low removal (less than 10%) of arsenic species by ion flotation in our experiments suggests that the interactions between arsenic species and oppositely charged surfactant species used as a collector may be weak, and therefore an activator is needed to transform arsenic species into more ionic forms for the ionic flotation of arsenic species. These results show that flotation based on the interactions between arsenic species and surfactants is a new effective method either by itself or in combination with other techniques for the removal of arsenic from water.



ARSENIC TARGETS THE DEVELOPING RAT LUNG: GENE EXPRESSION ALTERATIONS FOL-LOWING CHRONIC LOW-DOSE EXPOSURE.

Jay S. Petrick¹, Francoise M. Blachere², Kevin A. Greer³, Mark A. Schwartz³, Matthew J. Scholz⁴ , Omella Selmin⁵, Raymond B. Runyan^4,6, James B. Hoying ³, and R. Clark Lantz^4,6.

¹Department of Pharmacology and Toxicology, ²Department of Pediatrics, ³Department of Biomedical Engineering, ⁴ Department of Cell Biology and Anatomy, ⁵Department of Nutritional Sciences, and ⁶The Center For Toxicology, The University of Arizona, Tucson, AZ 85724.

The effects of chronic, low-level arsenic exposure in the drinking water remain controversial. Arsenic exposure has been correlated with increased lung cancer incidence, thus implicating the lung as a target organ for arsenic toxicity. The molecular effects of maternal arsenic consumption on the developing fetal lung remain poorly understood. We hypothesize that in utero exposure to inorganic arsenic causes altered gene expression in the lung, leading to molecular and functional changes. We exposed pregnant Sprague-Dawley rats to 500 ppb arsenic in the drinking water, in the form of sodium arsenite or sodium arsenate, from conception to embryonic day eighteen. Subtractive hybridizations of embryonic lung cDNA from control and treated day 18 embryos yielded numerous differentially expressed cDNA clones. We have sequenced 352 cDNA clones from the subtractive hybridizations and thus far have confirmed differential expression of 93 of these cloned genes by reverse dot blot hybridization. Using rat alveolar type II cells exposed to arsenic as a model for lung gene expression effects, we have carried out a dose and time-response study of chronic, low-dose arsenic exposure (10, 50, 100, and 500 ppb). Using these samples, we are profiling gene expression changes of clones from the subtractive hybridization procedure using cDNA microarray. The functional effects of altered gene expression during lung development will be assessed using specific probes to show alterations in protein expression and localization (Supported in part by the NIEHS Superfund Basic Research Program Grant P42 ES04940).



LANDFILL INDUCED REDUCTIVE DISSOLUTION OF ARSENIC AT A MASSACHUSETTS LAND-FILL.

Stanley W. Reed, RE.¹ and David I. Margolis, RE.<sup>2

1</sup>Harding ESE, Inc., P.O. Box 7050, Portland, ME 04112-7050, and ²U.S. Army Corps of Engineers, New England District, 696 Virginia Road, Concord, MA 01742-2751.

Groundwater monitoring at the downgradient edge of a municipal waste landfill in north central Massachu-setts shows the consistent presence of arsenic (up to 5,100 micrograms per liter [m g/L]), iron (up to 90,000 m g/L), and manganese (up to 13,000 m g/L). These high concentrations are coincident with low dissolved oxygen and oxidation-reduction potentials (ORPs) that typically range between -60 and -190 millivolts. Groundwater samples collected from within the landfill footprint show even higher concentrations and similar ORP values. Concentration differences between filtered (0.45 micrometer) and unfiltered samples show little difference, particularly at higher concentrations indicating that the arsenic, iron, and manganese are predominantly dissolved. This has lead to the conclusion that reducing conditions associated with waste degradation are causing the dissolution of iron and manganese hydroxides that coated soil particles in the aquifer. These hydroxide coatings likely provided sorption sites for arsenic under aerobic conditions. Under anaerobic conditions, the coatings dissolve, releasing arsenic, iron, and manganese. The reducing conditions and high dissolved arsenic, iron, and manganese persist for at least 800 feet downgradient of the landfill with no sign of attenuation. An association with pH is not apparent at the edge of the landfill where pH is typically high (i.e., 6.0 to 7.0 standard units [s.u.]) even in low concentration samples; however, the downgradient plume coincides with a region of relatively high pH (i.e., 5.5 to 7.0 s.u.) within a background of 3.3 to 5.5 s.u. The 84 acre landfill is located within the Hinkley-Merrimack-Windsor soil association and above an apparent fault zone that transitions between the metasiltstone of the Berwick Formation and the Devens-Long Pond facies of Ayer Granite. Analysis of soil samples from within the landfill footprint shows arsenic concentrations as great as 81 parts per million (ppm), and analysis of bedrock chips shows arsenic concentrations as great as 43 ppm.



COST COMPARISONS FOR ARESENIC CONTAMINATION AVOIDANCE ALTERNATIVES FOR MAINE HOUSEHOLDS ON PRIVATE WELLS.

Jessica M. Sargent-Michaud and Kevin J. Boyle.

Department of Resource Economics and Policy, University of Maine, Orono, ME 04469.

Arsenic in drinking water in Maine has become a public health concern. There may be as many as 30,000 private wells in Maine with arsenic levels in excess of 0.01 mg/L. This study was undertaken to help health officials and homeowners assess the relative costs associated with treatment alternatives for private well water with elevated levels of arsenic. Annual costs of reverse osmosis (RO), activated alumina (AA), bottled water (BW), rented (RWC) and purchased (PWC) water coolers were compared. Costs were calculated based households with one, two, three and four residents (the average Maine household has 2.39 residents). The least expensive treatment option for a single-person household is to purchase one-gallon jugs of bottled water (Table 1). For households larger than one person the least expensive treatment option consistently is to install a RO point of use (POU) system. The second-best option for a single person household is to pur-chase 2.5-gallon jugs of bottled water. For households larger than one person the second-best option is to install a POU AA system. Point of entry (POE) systems and water coolers were not cost effective. Although RO POU was found to be least expensive for households larger than one person it must be noted that mitigation costs can vary for any household with unique aspects of installing an RO or AA system in the home, other services provided by the installer and changes in market prices for the treatment systems or for bottled water. Before taking specific actions to mitigate exposure households should carefully investigate specific features of the systems they are considering and the exact cost to their household. Households installing RO or AA systems must also test their water on an annual basis to ensure their systems are effectively removing arsenic. No consideration was given to any differences between technologies effectiveness of removing arsenic. We assume all technologies are capable of reducing arsenic levels in well water to below 0.01 mg/L. Bottled water concentrations to this date have been below detection limit. Bottled water costs, in contrast to RO and AA systems, increase as the number of people in the household increases. Bottled water must also be stored.



INTERACTIONS IN ARSENIC BINDING SURFACES; A STUDY USING SURFACE PLASMON RES-ONANCE SPECTROSCOPY.

Diptabhas Sarkar and P. Somasundaran.

NSF IUCR Center for Studies in Novel Surfactants, Langmuir Center for Colloids and Interfaces, Columbia University, 911 Mudd Building, 500 W 120th St.. New York, NY 10027. U.S.A.

Many treatment technologies are capable of removing arsenic from potable water. But the socio-economic situation of the affected region complicates the choice of treatment technologies. So far, adsorption of Arsenic (III and IV) by oxides of Aluminum, Iron and their mixtures have shown encouraging results. Atomic force microscopy was employed here to understand the basic interaction mechanism of arsenic species with Fe³⁺ activated silica surfaces. Activation was achieved by exposing glass coverslips to 1 mM ferric chloride solution the pH of which was controlled at 3.00 ± 0.10. After activation the coverslips were exposed to arsenic solutions for 24 hrs. The appearance of precipitates confirmed the deposition of arsenic. Experiments performed using Fe₂O₃ particles as the activating species failed to provide any evidence of arsenic deposition, which we believe is due to the lack of z-resolution.



AN AUTOMATED ION CHROMATOGRAPHY-INDUCTIVELY COUPLED PLASMA MASS SPECTROMETRY METHOD FOR SPECIATION OF ARSENIC IN GROUND WATERS.

Jonathan L. Talbott, John W. Scott, and Marvin D. Piwoni.

Illinois Waste Management and Research Center, One E. Hazelwood Drive, Champaign, IL 61820.

Arsenic concentrations in well water from several locations in the Mahomet Aquifer exceed the 10 ppb drinking water standard being implemented by USEPA. The aquifer is the water supply for over 500,000 people in Central Illinois and western Indiana. While some arsenic data exists from both public and private wells, there are still many uncertainties regarding distribution of arsenic in the aquifer with depth. Further, little is known about the chemical species of arsenic in the aquifer and how the chemical speciation impacts arsenic removal at public water works. The Illinois Waste Management and Research Center is participating in a study to examine, in systematic fashion, arsenic distributions and chemical characteristics in the Mahomet Aquifer. Such information might assist decisions on treatment methodologies and on drilling depths for private wells. To support this effort, WMRC staff has developed an automated ion chromatography-inductively coupled plasma mass spectrometry (IC-ICPMS) method for the determination of arsenic species in ground waters. Speciation for arsenite (As³⁺) and arsenate (As⁵⁺) is achieved by coupling a Dionex AS 11 IonPac anion-exchange column to the sample introduction system of a Thermo Elemental PQ ExCell ICP-MS operated in a time domain mode. Automation of the entire IC-ICPMS system was achieved by modifying the sample introduction system of the PQ ExCell with a high-pressure sample valve. The high-pressure sampling valve serves as the sample loop for the 1C system, enabling control of the entire system from the ICPMS software. The 1C column is operated in an isocratic mode with 0.005M phthalic acid, pH ~3, as the eluent. Although chloride elutes at the same retention time as arsenite, concentrations up to 3000 ppm yield only minimal interference equivalent to ~1 ppb of As. The sum of arsenic species measured in ground-water samples compares favorably (p < 0.05) with total arsenic determined by ICPMS. Sample throughput is approximately 20 samples per hour and detection limits for arsenic species are less than 0.5 parts per billion (ng/mL). The technique is robust and reproducible for ground-water samples.



INSTRUMENTAL VARIABLE ANALYSIS FOR ARSENIC AND CANCER.

Tor D. Tosteson¹, Raymond J. Carroll², David Ruppert ³, and Margaret R. Karagas¹.

¹Department Community and Family Medicine, Dartmouth Medical School, Hanover, NH 03755; ²Department of Statistics and Department of Epidemiology and Biostatistics, Texas A&M University, College Station TX 77843—3143; ³School of Operations Research & Industrial Engineering, Comell University, Ithaca, NY 14853-3801.

An ongoing population based study in New Hampshire (Karagas et. al., 1998, Karagas et. al. 2001) is examining the effects of arsenic on the incidence skin and bladder cancer in response to low to moderate exposures, primarily due to natural sources of arsenic contamination in well water. Because of intense regulatory interest in the effects of possible abatement strategies, the shape of the exposure response relation-ship at lower exposures is important and strategies for nonlinear modeling are being developed. Exposure assessment is accomplished through the measurement of arsenic concentrations in both tap water from home water supplies and toenail samples for individuals newly diagnosed with skin or bladder cancer (cases) and individuals belonging to an age and gender matched sample of other state residents (controls). We have proposed combining these two measures in a flexible nonlinear regression model for case control data using measurement error methods for "instrumental variables." We report on the statistical properties of these methods and apply them to our data on non-melanoma skin cancer and arsenic. A new result shows that both the nonparametric regression function and all instrumental variable parameters are identified under relatively weak conditions. Using simulation studies, the proposed methods are found to increase the precision of estimated dose response curves by reducing the bias associated with measurement error. For our example, we consider data for 215 controls and 233 basal cell skin cancer cases having both water and toenail samples. Because we are interested in characterizing changes in cancer incidence due to changes in arsenic water contamination, we specify the water measurement as the unbiased exposure. The toenail arsenic measurements are interpreted as the instrumental variable. For the purposes of this analysis, the results were not adjusted for possible confounding factors such as age and gender. A modest increase in the odds of incidence is seen over the range shown for tap water concentrations. The results of correction for measurement error are not dramatic, but seem to be somewhat more pronounced for lower concentrations. In summary, new methods of analysis have been devised to help study the incidence cancer as a function of arsenic exposure using both individual toenail and household water concentrations. The methods are shown to have improved statistical properties in the presence of measurement error.



CYCLIC VOLTAMMETRIC STUDY OF REDOX REACTIONS OF ARSENIC.

Zhenqiang Wei, Paul F. Duby, and Ponisseril Somasundaran.

NSF IUCR Center for Studies in Novel Surfactants, Langmuir Center for Colloids and Interfaces, Columbia University, 911 Mudd Building, 500 W120th St., New York, NY 10027, U.S.A.

Adsorption of arsenic species by oxides of aluminum, iron and their mixtures has shown to be feasible for removing arsenic from potable water. It was determined that the removal of arsenate (As(V)) is more facile than that of arsenite (As(III)). The present study is aimed towards developing a knowledge base for the oxidation or reduction between arsenic species in water for designing treatment techniques such as electrore-mediation. The redox reactions between As(III) and As(V) in acidic solutions were investigated using cyclic voltammetric technique with a platinum rotating disk electrode (RDE). Only one cathodic and one anodic peak were observed in the potential region corresponding to the reduction or oxidation of arsenic species, indicating that the two electron redox reactions between As(III) and As(V) are controlled by one slow reaction. Judging from the shift of the peak potential as a function of the scan rate and the potential difference between anodic and cathodic peaks, the redox reactions between As(III) and As(V) can be considered to be irreversible slow reactions.



OCCURRENCE OF ARSENIC IN RESIDENTIAL DRINKING WATER WELLS PROXIMATE TO OLD MINE SITES IN NEW HAMPSHIRE: A SURVEY OF FOUR SELECTED MINE SITES.

Michael J. Wimsatt¹ and Thomas P. Ballestero².

¹New Hampshire Department of Environmental Services, Concord, NH and ²Environmental Research Group, University of New Hampshire, Durham, NH.

Arsenic concentrations were measured in 31 drinking water samples collected from households located near four New Hampshire mine sites where rock containing arsenic-rich minerals is known to be located, and which may have been mined for arsenic. Results for the 25 bedrock wells sampled showed a higher incidence of arsenic detections above the proposed 5 m g/1 drinking water standard than was found in a previous study of 218 randomly selected New Hampshire domestic bedrock wells. Anomalously high results were obtained in the five wells located near one of the mine sites. These data do not allow the statistical conclusion that wells located near mine sites are at an increased risk of having elevated arsenic concentrations. However, the data encourage further study of two of the mine sites to explore the potential connection between rock that may contain arsenic-rich minerals and elevated groundwater arsenic.