Detection of airborne Stachybotrys chartarum macrocyclic trichothecene mycotoxins in the indoor environment

Abstract

The existence of airborne mycotoxins in mold-contaminated buildings has long been hypothesized to be a potential occupant health risk. However, little work has been done to demonstrate the presence of these compounds in such environments. The presence of airborne macrocyclic trichothecene mycotoxins in indoor environments with known Stachybotrys chartarum contamination was therefore investigated. In seven buildings, air was collected using a high-volume liquid impaction bioaerosol sampler (SpinCon PAS 450-10) under static or disturbed conditions. An additional building was sampled using an Andersen GPS-1 PUF sampler modified to separate and collect particulates smaller than conidia. Four control buildings (i.e., no detectable S. chartarum growth or history of water damage) and outdoor air were also tested. Samples were analyzed using a macrocyclic trichothecene-specific enzyme-linked immunosorbent assay (ELISA). ELISA specificity was tested using phosphate-buffered saline extracts of the fungal genera Aspergillus, Chaetomium, Cladosporium, Fusarium, Memnoniella, Penicillium, Rhizopus, and Trichoderma, five Stachybotrys strains, and the indoor air allergens Can f 1, Der p 1, and Fel d 1. For test buildings, the results showed that detectable toxin concentrations increased with the sampling time and short periods of air disturbance. Trichothecene values ranged from <10 to >1,300 pg/m3 of sampled air. The control environments demonstrated statistically significantly (P < 0.001) lower levels of airborne trichothecenes. ELISA specificity experiments demonstrated a high specificity for the trichothecene-producing strain of S. chartarum. Our data indicate that airborne macrocyclic trichothecenes can exist in Stachybotrys-contaminated buildings, and this should be taken into consideration in future indoor air quality investigations.

FIG. 1.

Andersen GPS-1 PUF high-volume air sampler setup. Panel A shows the collection module following a 24-hour sampling period. The top filter with a considerable amount of collected particulates is visible here. Panel B is a schematic of the collection module, with a top view on the left and a side view on the right. The module was modified to collect and separate particles using glass microfiber filters. Large particles, including most fungal conidia, were collected on 90-mm-diameter 2.7-μm-pore-size GF/D filters (1), while remaining particles able to pass through the first filter were collected on highly efficient EPM filters of the same diameter (2). A heavily mold-contaminated storage closet adjacent to the source of the water damage in test building 8 (shown in panel C) was chosen for sampling. Water-saturated air and ensuing fungal contamination were a result of major damage to the air-conditioning unit. The degree of the damage was evident by growth near the air exit grates throughout the building (D).

FIG. 2.

SpinCon PAS 450-10 bioaerosol sampler controlled setup. PVC pipe was filled with S. chartarum-contaminated ceiling tile and attached to the air inlet of the SpinCon sampler. Potential leaks surrounding the inlet were sealed with aluminum foil. Air passing over and through the ceiling tiles was directed into the collection chamber at a rate of 450 lpm. Aerosolized Stachybotrys chartarum conidia and other particulate matter were captured by a swirling column of PBS in the collection chamber. These trials were run in an outdoor environment. Sampling was performed for 10 and 30 min (n = 3 replicates for each time interval). For comparison purposes, collection was also performed using an equal area of sterile ceiling tile in the same manner.

FIG. 3.

Box plot data for average trichothecene equivalents per m³ of sampled air in Stachybotrys-contaminated and control indoor environments. Trichothecene equivalents (in picograms) were determined using a macrocyclic trichothecene standard curve. Graph A shows the data distribution from 120-min (m) control and test samples under static (S) and disturbed (D) conditions and from 30-minute samples from disturbed test environments. Medians (solid lines) and means (dotted lines) are shown. The 10th and 90th percentiles are designated by the bottom and top error bars, respectively. The 25th and 75th percentiles are indicated by the bottoms and tops of the boxes, respectively. Outliers are designated as filled circles above and/or below the plot. Test environments were compared to control environments (**) using a Kruskal-Wallis one-way ANOVA on ranks. Graph B shows the data distribution from control and test environments sampled for 10 min under disturbed conditions. Test environments were compared to controls by using the Mann-Whitney rank sum test. *, statistically significant differences (P < 0.05).