Effects of different extraction protocols on endotoxin analyses of airborne grain dusts.

J'C. Effects of different extraction protocols on endotoxin ana lyses of airborne grain dusts. Scand J Work Environ Health 1989;15:430-435. The detection of gram negative bacterial endotoxins in occupational dusts, specifically those from agricultural environments, is of increasing importance in research on occupational lung disease. In this study, the quanti tative chro mogenic Limulus amebocyte lysate test for the detection of endotoxins in airborne dusts from spring wheat and oats was examined. Different extraction fluids were tested, as were the effects of time on extraction and of repeated freeze-thaw cycles on the extracts. The data suggest that the chromogenic method can be used effectively in the analysis of environmental dusts or their frozen extracts for endotoxin quantita tion. Water appears to be the preferable extraction medium, and the length of extraction time may affect the results.

Gram-negative bacteria and their endotoxins appear to be ubiquitous contaminants o f materials and dus ts related to agricultural occupations. Endotoxi ns are known to stimulate a wide range of profound biological ef fect s (I, 2), and the lung is a target or gan for endotoxin-induced damage (3,4). Reports of endotoxin levels in animal confinement bu ildin gs (5)(6)(7)(8) , cotton textile mills (9,10), port gr ain terminals (11), poultry-processing plants (12), and tower silos (13) attest to the potential for respiratory impairment in member s of the diverse work for ce that inhales endotoxinladen du sts. Controlled exposures o f humans to endotoxin-laden cotton du sts have led to the definition of an association bet ween de cre ases in acute pulmon ary function and the level o f airbo rne endotoxins (14)(15)(16) . Cotton mill studies in The People's Republic of China ha ve also shown a do se-response relationship between endotoxin levels in cotton dust and chronic lung impairment such as chronic bronchitis among cotton workers (17). In the acute studies, the thr esholds for no pulmonary function change were defined as 80 ng/rn' for cotton workers who smoke (16), but the results from a lar ger study of a mixed population indicated a calculated zero change threshold o f as little as 9 ng/rn' (IS) .
The two con trolled-exposure st udies that obtained zero pulmonary fun ction change th resholds for endotoxin levels used two markedly different techniques to quantify the endotoxins . Haglind (18), while in another study the endotoxic respirable dust pot ent ial of baled cotton was determined by a kineti c Limulus test (19). Endotoxins in compost samples have been detected by gas chromatography (20), and other investigations have assessed endotoxins in swine confinement buildings (5) and different agricultural enviro nme nts (21) with a qu antitative chromogenic mod ificat ion of the Lim ulus test.
In an attempt to develop scientific sta nda rd ization o f the tests used 10 determine end otoxins in enviro nment al dusts, the Work Group on Agents in Organi c Dusts in the Fa rm En vironment recommended the use of th e qu antitative chromogenic mod ification of the Litnulus amebocyte lysate test (22). Th is same technique was recommended again at an intern ational symposium on work-related respiratory disorders among farme rs becau se of its accuracy and reproducibility (23) .
The purpose of this paper is to evalua te th e efficacy of the qu antitative chromogenic Limulus amebocyte lysate test in determining end oto xin co ntamination in enviro nmenta l dust s. Specifically, the detection of endo tox ins in airborne spring wheat and oa t du sts was follo wed when vario us extrac tio n prot ocols were used.

Endotoxin analyses
Throu ghou t th is study , ster ile, pyrogen-fre e plasticware was used. Sterile, non pyro genic water and saline were obtai ned co mmercially (T ravenol Lab o ratories, Inc, Deerfield, Illinois, Unit ed Sta tes) a nd used throu ghou t this study un less otherwise not ed. The en-dotoxin analyses were performed in duplicate according to a quantitative chromogenic modification of the Limulus amebocyte lysate test (QCL-1000; Whittaker Bioproducts, Walkersville, Maryland, United States) .

Commercial lipopolysaccharides
Lipopolysaccharides (LPS) from different bacteria were obtained commercially, reconstituted with water or saline, and diluted (with the same water or saline) to a final concentration of 5000 EU /ml, as defined by the manufacturer, and assayed.
In the assessment of the effect of possible differences in the extraction media on the resulting endotoxin determinations , two separate vials (100 rng contents) from the same manufacturer's lot number of LPS from Eschericheria coli 0127:88 (DIFCO Laboratories, Detroit, Michigan, United States) were reconstituted with 20 ml of water or saline, and the endotoxin activities were determined.
In evaluating the differences in endotoxin activities, commercial LPS samples from various bacterial sources and different preparations [Wesphal (W), Boivin (8)] were reconstituted to a concentration of I mg/ml or 5 mg/ml (depending upon the initial content of the vial) with water and assayed. LPS from E coli (W) 0127:88, Salmonella typhosa (B) 0901, S typhimurium (B), S typhimurium (W), and S enteritidis (B) were obtained from mFCO; samples of LPS from E coli KI2 D31m4 (Re), E coli K12, mm294 (complete core), and S minnesota Re 594 were obtained from List Biological Laboratories (Campbell, California, United States).

Grain dusts
Airborne dusts from spring wheat and oats were collected as described previously (24) from active port grain terminals in the northern United States. The standard extraction protocol for bulk dusts in our laboratory is rocking in water (in this case 500 mg of dust in 10 ml of water) for 60 min at room temperature. The mixture is then centrifuged for 10 min at 1000 g, and the supernatant fluid is decanted and assayed for endotoxin content.

Extraction solutions
The effect of different extra ction solutions on the efficacy of the quantitation of endotoxin levels in grain dust samples was studied. One percent (volume /volume) solutions were made in water from Tween 20 (polyoxyethylene sorbitan monolaurate) (Technicon Corporation, Tarrytown, New York, United States), Tween 80 (po!yoxyethylene sorbitanmonooleate) (Fisher Scientific Co, Fairlawn, New Jersey, United States), Triton X-IOO (polyethylene glycol p-isooctylphenyl ether) (Fisher Scientific Co), and Saponin (sapogenin glycosides) (I % weight/volume; Sigma Chemical Co , St Louis , Missouri , United States). In addition, I % and 0 .01 % (weight /volume) solutions were made from human, bovine, and chicken egg albumin (Sigma Chemical Co).
Each solution was used in place of water to extract dusts from both spring wheat and oats by our standard protocol. Dilutions were made with the extraction solutions.
In the standard protocols for our laboratory, the commercial chromogenic Limulus amebocyte lysate assay requires a standard that is reconstituted with water before the standard curve is measured. In order to compare the effects of different extraction solutions on the efficacy of endotoxin removal from the grain dust samples, we made separate standard curves by reconstituting the endotoxin standard with two different extraction solutions, Tween 20 or Saponin. All the standards were from the same manufacturer's lot number, and the dilutions for the standard values were made with the same solutions that were used for the reconstitution.

Timed extractions
We made standard extractions of bulk dusts by rocking the dusts for 60 min. The effect of time on the extraction efficacy was measured. Aliquots (10 g) of spring wheat dust and oat dust were placed in separate sterile, nonpyrogenic, 75-cm' tissue culture flasks (Corning Glass Works, Corning, New York, United States) and mixed on a tube rotator (Scientific Equipment Products, Baltimore, Maryland, United States) for 2 h at room temperature to insure that the dusts were homogeneous . Nine samples (500 mg each) were taken from both dusts , and each was extracted by rocking in sterile, nonpyrogenic water for different lengths of time at room temperature. The endotoxin measurements were made after extractions of 15 and 30 min and after extractions of 1,2,4,8, 12,24, and 36 h.

Freeze-thaw studies
The effect of repeated freeze-thaw cycles on the level of detectable endotoxins in water extracts of the airborne grain dusts was studied. The standard extraction protocol was followed. Endotoxin measurements of the same extracts were made on day 0, never frozen, and at 2-to 3-d intervals for the next 30 dafter freezing ( -85°C). The same extracts were thawed and frozen again on each test day.

Commercial Iipopolysaccharides
Commercial LPS from E coli were reconstituted in sterile, non pyrogenic water or saline and diluted to a concentration of 5000 EU /ml, on the basis of the manufacturer's stated concentration. Analyses of the samples for endotoxin activity resulted in endotoxin levels of 3934.72 EU /ml for LPS dissolved in water

Effect of extraction solutions
Before the effect of different extraction solutions on the detectable endotoxin levels could be determined, the endotoxin contamination of the solutions was quantified. Table I lists the concentrations of endotoxin in the extraction solutions. The data show that human and chicken albumin contained modest endotoxin contamination. The other solutions were basically free of endotoxins.
The endotoxin levels detected in the airborne dusts from spring wheat and oats after extraction with different solutions are listed in table 2. In addition, the percentage of change from the " standard " water extraction for each solution is presented. While the majority of solutions resulted in the extraction of lower levels of detectable endotoxin than when water was used, extraction with Tween 20 and Saponin resulted in increases of detectable endotoxins for both dusts. In both cases, however, the increase over water extraction alone was less for oat dust than for spring wheat dust.
Although the data suggest that Tween 20 or Saponin resulted in potential improvements when used as extraction solutions instead of water, it must be remembered that table 2 reflects data that were obtained with the use of a standard curve that was made with water , as required by the chromogenic Limulus amebocyte lysate test protocol. To determine if these solutions affected the chromogenic assay, we made standard curves from endotoxin standards of the same lot number with the Tween 20 and Saponin solutions (figure 2). The coefficient of correlation for both curves was >0.99. However, the standard curve made with Saponin was not usable because valid absorbance readings could not be achieved after repeated attempts. Therefore, only the curves for Tween 20 and water are shown in figure 2.
The endotoxin levels in spring wheat dust and oat dust when a Tween 20 extract was used with a water standard curve were 2687.38 EU/mg and 2833.08 EU/mg, respectively (table 2). When the Tween-20 extract was used with the Tween-20 standard curve, however, the detectable endotoxin levelswere only 37.16EU/mg and 87.57 EU/mg for spring wheat dust and oat dust, respectively, because of the marked change in the slope of the standard curves.  Figure 3 illustrates the effect of extraction time on the level of detectable endotoxins in homogeneous samples of both spring wheat dust and oat dust. For both dusts, the peak extractable endotoxins were determined after extraction for 2 h. The levels dropped after that time and remained low through 36 h of extraction. The standard l-h extraction yielded 68.8 and 77.4 % of the peak levels for spring wheat dust and oat dust, respectively.

Effect of repeated freezing and thawing
Single water extracts of spring wheat dust and oat dust were used in the study of the effect of freezing and thawing on the detectable endotoxin levels. The results o f the 30-d experiment are shown in figure 4. Both extracts were analyzed at the same times, and a new endot oxin kit was used each day. The kits were from a single lot number until day 23, when a new lot number was used. Day-to-da y variation from the original, unfro zen, extract could be ob served.  ment and processing facilities, cotton textile mills, port grain terminals, and tower silos pro vides new insight into the pot ential respira tor y hazard s of workers who breathe the dusts. Recent report s indicate that accurate measurement of airborne endotoxin levels can be used to estimate the threshold of acute pulmonary function response to inhaled cotton dusts (14)(15)(16). In addition, the acute pulmonary response to cotton dust correlated with the endotoxin levels and not the gravimetric dust levels. Thu s there is need for an accurate and reproducible method for detecting airborne endotoxin levels. While most laboratories that study airborne endo toxins rely on the Limu/us amebo cyte lysate gelation test for their studies (25), the chromogen ic modificat ion of the Limu/us test has been endorsed by two groups of scientists who study agriculturally related

Freeze-Thaw Study
'" Spri ng 'rmot I1Jst lung diseases (22,23). Our stud y is a first att empt to examin e th e chromogenic mod ificat ion of the Limu-Ius amebocyte lysat e test to determ ine its usefulness for environmentally derived du st stud ies. Specifically, different extraction protocols were exam ined with two different airborne grain du sts. Data from recon stitution studies with commercial LPS suggest that water is the medium of choice, althou gh commercial LPS samples from different organisms may vary in their endo toxin activities when reconstituted with water . The problem of whether these differences are a result of th e rela tive toxicities of the different LPS samples or of incons istency in the product preparation was not addressed fully in these experiments. Howe ver , it is interesting to note that no difference in endotoxin activity was noted when two prepa rations of S typhimurium , Boivin and Wesphal, were examined. This result may imply that the differences in the endotoxin activities of the LPS samples co uld be due to the relative toxicities of the LPS from different source microorganisms.
Water was used , therefore, to extra ct two samples of airborne grain dust s. Th ese samples were chosen to be representati ve of a variety of agricultur al dusts, and we could obtain a sufficient quantity of well-mixed du st so that con sistent repeated sampling could be made. It can be seen from the data tha t the du sts from spri ng wheat and oats pro vided different values throughout these tests. Agricultural dusts would be expected to be unique and different from each other in endotoxin contamination becau se of th e diversity of th e growth area , the varieties of plants, and the storage and transport of the bulk grai ns.
Because extraction with water alone may harbor its own disad vantages, especially when the extractio n of whole bacterial cells is the research objective (26), several agents were added to the water to test the efficacy of extraction. Of th e solutions that we examined , only I 070 Tween 20 and I 010 Saponin increased the detectable endotoxin concentration for both dusts. However, examination of the characteristics of the endotoxin standard curve when the wetting agents were used revealed that the data were art ificially elevated. As shown for the calculations of end oto xins when the Tween 20 standard curve was used with Tween 20 extra ction , th e true values were mark edly lower. These results imply that these agents affected the chro mogenic assay .
The length of extraction time can affect the dete ctable endotoxin co ncentratio n in extracts, as demonstrated by the timed extractions o f spring wheat and oat du sts ( figure 3). Extraction o f both du sts for 2 h pro vided maximal levels of detectable endo toxin. Th e levels declined rapidl y, and , by 12 h of extrac tion , the y stabilized at approximately 30-40 070 of the ma ximal level. One can onl y specula te at th is point in the resear ch as to the cause of th e lowered detectable endotoxin s with longer extraction times. It is possible that the endotoxins adsorbed onto the tube s, or ont o par-434 ticles, and were no longer available in the superna ta nt fluid . It is possible a lso th at the molecules joined into la rger structures and hid the active moieties. Another possibility lies in the environment al du sts themselves. Perhap s interactions wit h an agent( s) in the du st resulted in an inactivation or altera tio n of the Limulusreactive moiety, either through oxidati on or anot her chemical interaction . Whatever the cause, the data do suggest that th e length of extrac tio n sho uld be held constant within a given study. An examination of other environmental du sts should determ ine if the 2-h extraction maximum is consistent o r if it varies with di fferent dusts.
Finally, the freeze-thaw exper iments demon str ated tha t endotoxi n analyses of spring wheat an d oat du sts can be performed on extracts that have been frozen. Ou r experiments illustrated th e worst case situation wherein the extracts underwent repeated freeze-thaw cycles for 30 d .
In conclusion, the recommended chro mogenic modi fication of the Limulus amebocyte lysate test can be used effectively in the an alysis of enviro nmental dusts for endotoxin quantitation . Of the extraction media that we tested, water is a preferabl e extractio n med ium , and the length of extrac tio n sho uld be exami ned for each du st if sufficient qu ant ities of a th orou ghly homo genized bulk du st are ava ilable.