| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Technical Briefs |
1
Laboratoire National de Santé, PO Box 1102, L-1011 Luxembourg, Luxembourg
2
pab productions, Bürgermeister-Herzog-Strasse 5, D-85241 Hebertshausen, Germany
aauthor for correspondence: fax 352-490686, e-mail claude.muller{at}santel.lu
Most immunoassays are based on microtiter plates with conventional polystyrene (PS) as a solid support. PS is produced from crude oil. The production of PS plates consumes several hundred thousands tons of valuable raw material of limited supply. Also, disposing of PS plates by incineration produces a complex mixture of pollutants including dioxins and polycyclic aromatic hydrocarbons (1), and it adds excess CO2 to the atmosphere. In contrast, biopolymers such as polylactic acids (PLAs) are produced from renewable resources; they are biodegradable and their CO2 balance is neutral. PLA-based polymers have successfully replaced synthetic polymers in several applications, including some recent applications in immunology (2)(3). For the development of a PLA formula suitable as a matrix for immunoassays, the physical properties and surface characteristics of the polymer had to be improved for the present study. Surface characteristics are affected by factors such as monomer content, the nature and concentration of various additives, and molding and releasing conditions. Moreover, lot-to-lot reproducibility is affected by industrial fermentation and extraction procedures.
For this study, the injection-molding process commonly used for PS was adapted to the PLA polymer to produce an experimental batch of 96-well microtiter plates from suitable PLA formulations. The performance of these plates in an ELISA was compared with conventional commercial PS plates.
We used recombinant measles virus hemagglutinin protein (H-ELISA) as antigen for the detection of measles-specific IgG (4)(5)(6). We analyzed a panel of 426 human sera using 96-well prototype microtiter plates made of PLA. The procedure was applied essentially as described (5). All study participants or their parents gave informed consent. Results were compared with those obtained with the same antigen and commercial PS Maxisorp microtiter plates (NUNC).
Statistical analysis of data was performed with SigmaStat software (Jandel Scientific). A z-test was applied to compare two proportions (of a population) with a 95% confidence interval. The Mann–Whitney U-test was used to evaluate the significance between median values. The correlation coefficient (r) was obtained by linear regression analysis. The 
correlation coefficient () and the proportion of agreement were calculated according to the method of Bland and Altman (7). The performance characteristics of the ELISAs were calculated following the definitions of Bland (8). Sera were collected from 221 vaccinated high school children (12.1–15.1 years; median age, 12.7 years; 37 males and 184 females) and 194 adults, including 82 males (25–91 years; median, 42 years) and 112 females (25–78 years; median, 45 years), who had measles during their childhood. The children were vaccinated at least once against measles at the age of 4.7–148.9 months (median, 19.5 months). All sera were tested by plaque neutralization assay (NT) as described by Huiss et al. (9); NT titers were expressed as -log2 of the dilution, with values of 1:24 being considered negative. This assay served as a gold standard (10)(11). All of the above sera (n = 415) were NT positive and further confirmed by a certified commercial ELISA (Enzygnost; Dade Behring). In addition, 11 NT-negative sera from seven unvaccinated 15-month-old children, two adolescents, and two adults were included.
The PLA formulation used for the production of the prototype plates had a slightly greenish coloration. The absorbance (mean ± SD) at 405 nm of the empty wells (n = 852) was 0.129 ± 0.005 with a CV of 3.7%, which was similar to the CV obtained for the commercial PS plates (n = 852; mean absorbance, 0.040 ± 0.002; CV, 3.8%). This demonstrated the highly uniform quality of the formulation of the PLA-molding mass. When the PLA plates were coated with increasing concentrations of recombinant H protein, the plates were saturated with 125 ng of total protein per well (data not shown). Similar values were found [e.g., with Maxisorp PS plates (125 ng/well) (5)]. Plates were coated in addition with an irrelevant control antigen derived from the same expression system, which was used as a negative control to assess the unspecific background of each serum (4). This background was subtracted from each measurement to obtain the corrected net absorbance value for each serum [absorbancespecific antigen - absorbancecontrol antigen (5)]. Raw absorbance of the serum reactivity with the H protein and the net absorbance values were evaluated. The correlation coefficient (r) between raw and net values was 0.991. The threshold for positivity was determined by two-graph ROC analysis (12)(13). The highest sensitivity and specificity corresponded to the mean absorbance + 3 SD of the negative sera for the PLA plates. Sera with absorbance >0.028 {[mean + 3 SD = 0.004 + 3 (0.008)] or 0.365 [0.245 + 3 (0.040)]} were considered positive when net or raw data were evaluated, respectively. H-ELISAs in PS plates were evaluated as described previously (4)(5). The threshold was determined to have an absorbance of 0.119.
The absorbances obtained with the H-ELISA with the PLA plates correlated closely (r = 0.91) with those obtained with the commercial PS plate, as shown in Fig. 1
. The maximum absorbance values were higher in the PS assay, which translates into a somewhat wider dynamic range of the PS assay.
|
In Table 1
, the two assays using either the PLA or the Maxisorp PS plates are compared. Although the number of negative serum samples was limited, no false-positive results were detected by the H-ELISA on the basis of the above thresholds, irrespective of whether PLA or PS plates were used and whether net or raw data were considered. There was also no significant difference between the two types of plates with respect to the number of false-negative results (PLA net = 2 of 415 vs PS net = 7 of 415; P = 0.18) and whether raw or net data of PLA plates were compared (PLA raw = 2 of 415 vs PLA net = 3 of 415; P = 0.99).
|
The specificity and sensitivity of the PLA plates were as good as those obtained with the commercial PS plates, even when the raw data of the PLA plates were evaluated. According to these results, both negative and positive samples have a >99% chance of being tested correctly with the PLA-based H-ELISA. The absorbances of both types of plates were highly significantly correlated whether net ( = 0.83; proportion of agreement, 98%) or raw values ( = 0.87; proportion of agreement, 99%) were considered for the PLA plates.
The negative predictive value was significantly better when PLA was used as a matrix independently of whether raw or net data were analyzed (PLA raw, 77.7%; PLA net, 83.9%; PS net, 59.9%; PLA raw or net vs PS net, P <0.001). Thus, the prediction for an individual who tests negative will be more reliable when PLA plates are used, avoiding unnecessary revaccination. Generally, vaccinees tend to have considerably lower anti-measles antibody titers than late convalescents (5). In this study, whole measles antibodies as measured in a certified ELISA (Enzygnost) were 
2.5-fold lower in the vaccinees in comparison with the late convalescent sera (median A450 nm values, 0.597 vs 1.525, respectively; P <0.001). The performance of both the PLA- or the PS-based ELISAs did not significantly change when only the vaccinees were analyzed (data not shown).
We demonstrate here for the first time that PLA can be formulated to feature molding properties and adhesive surface properties that are compatible with the development of an immunoassay. Despite a somewhat narrower dynamic range of the PLA plate, substituting the conventional commercial PS-based ELISA plates by the PLA prototype plates did not compromise the specificity or the sensitivity of the assay. The PLA plates were also shown to bind a variety of antigens, including antibodies, peptides, and others (manuscript in preparation).
A new generation of PLA plates is currently being molded with an improved PLA formulation and a reduced absorbance to minimize plate background and to enhance the dynamic range. Plates that use 8-well strips are also now available. The incorporation of specific functional groups into the PLA matrices is also possible. These groups are available on the PLA surface and can be used to couple antigens of low molecular weight to the plates by conventional coupling procedures. Thus, PLA has properties suitable for the development of immunoassays that rival those that use PS. Moreover, in contrast to PS, PLA is produced from sustainable resources and is of minimal burden to the environment.
Acknowledgments
We are grateful to NUNC (Denmark) for performing the initial molding of the prototype PLA plates. We thank Dr. V. Reddy-Bemtgen and S. Willieme for technical assistance. This work was supported as a Demonstration Project under the contract BIO4-CT98-0273 of the 4th Framework Program of the European Union.
References
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
represent NT-positive and -negative sera, respectively. The r value was calculated on the basis of the best polynomial trendline (order 2). For both assays, the thresholds for positivity are represented as dashed lines.