Report

A B Cell-Based Sensor for Rapid Identification of Pathogens

See allHide authors and affiliations

Science  11 Jul 2003:
Vol. 301, Issue 5630, pp. 213-215
DOI: 10.1126/science.1084920

Abstract

We report the use of genetically engineered cells in a pathogen identification sensor. This sensor uses B lymphocytes that have been engineered to emit light within seconds of exposure to specific bacteria and viruses. We demonstrated rapid screening of relevant samples and identification of a variety of pathogens at very low levels. Because of its speed, sensitivity, and specificity, this pathogen identification technology could prove useful for medical diagnostics, biowarfare defense, food- and water-quality monitoring, and other applications.

The diagnosis of infectious diseases such as severe acute respiratory syndrome (SARS) and detection of potential bioterrorism agents such as Bacillus anthracis (anthrax) and variola major (smallpox) would benefit greatly from a pathogen identification method with better combined speed and sensitivity than existing methods such as immunoassays (1) and polymerase chain reaction (PCR) (2). This report describes a pathogen sensor that achieves an optimal combination of speed and sensitivity through the use of B lymphocytes: members of the adaptive immune system that have evolved to identify pathogens very efficiently. B cell lines were engineered to express cytosolic aequorin, a calcium-sensitive bioluminescent protein from the Aequoria victoria jellyfish (3, 4), as well as membrane-bound antibodies specific for pathogens of interest. Cross-linking of the antibodies by even low levels of the appropriate pathogen elevated intracellular calcium concentrations within seconds (5), causing the aequorin to emit light (6, 7). We named the sensor CANARY (cellular analysis and notification of antigen risks and yields).

We first developed a system for efficient production of pathogen-specific B cell lines. A parental cell line with stable expression of cytosolic aequorin (8) was generated from the M12g3R (IgM+) B cell line (9), and the clone with maximum light emission upon cross-linking of the surface immunoglobulin M (IgM) was selected (10). The M12g3R-aequorin cells were subsequently transfected with plasmids containing antibody light- and heavy-chain constant-region genes, into which variable regions specific for a particular pathogen were inserted. Clones from the second transfection were selected for optimal response to that pathogen (10).

The resulting cells responded to pathogens with excellent speed, sensitivity, and specificity. Cells specific for Yersinia pestis, the bacterium that causes plague, could detect as few as 50 colony-forming units (CFU) in a total assay time of less than 3 min, which included a concentration step (Fig. 1A). The probability of detection for Y. pestis ranged from 62% for 20 CFU to 99% for 200 CFU (fig. S1), whereas the false-positive rate for the CANARY assay was 0.4%. These cells did not respond to large numbers of unrelated bacteria (Francisella tularensis), nor did excess F. tularensis block the response to very low levels of Y. pestis. Similar performance was observed with other cell lines, including one specific for orthopoxviruses (Fig. 1B). The sensitivity of a B cell line specific for Venezuelan equine encephalitis (VEE) virus, a virus too small to be concentrated in a microcentrifuge (10), is currently 5 × 105 plaque-forming units (PFU) (Fig. 1C).

Fig. 1.

Dose response and limit of detection for pathogen-specific B cells. (A) Formalin-inactivated Y. pestis and/or F. tularensis (50 μl) were concentrated for 60 s at 10,000g, then Y. pestis–specific cells (20 μl) were added and spun for an additional 5 s (10). Photon emission was measured in a luminometer. Identical results were obtained with a live avirulent strain of Y. pestis. (B) Vaccinia virus was concentrated and tested with orthopox-specific cells as described in Fig 1A. (C) Irradiated TC-83 VEE was assayed (without concentration) with VEE-specific cells (16).

CANARY can also identify pathogens in complex relevant samples. B cells specific for Escherichia coli strain O157:H7, an important pathogen found in vegetable, fruit, and meat products (11), detected as little as 500 CFU/g in lettuce in less than 5 min, which included the initial sample preparation time (Fig. 2A). These results compare favorably with reports describing the detection of bacteria in food with PCR protocols that take 30 to 60 min and achieve a limit of detection of 10 to 10,000 CFU per gram or per milliliter (12, 13). Rapid pathogen identification methods are necessary to ensure timely accurate diagnosis of disease in patients with infections requiring immediate treatment (14). CANARY can detect as few as 1000 CFU of B. anthracis spores extracted from seeded nasal swabs (Fig. 2B), demonstrating the potential to rapidly screen patients for inhalation anthrax exposure.

Fig. 2.

Detection of pathogens in samples. (A) Lettuce (25 g) was contaminated with the indicated amounts of inactivated E. coli O157: H7 and shaken in sterile bags with extraction medium (assay medium without fetal bovine serum). The supernatant was passed through a 5-μm filter to remove large particulates, the eluate was centrifuged, and the liquid was replaced with assay medium. The samples were tested as described in Fig. 1 with cells that respond to E. coli O157:H7. (B) Nasal passages were wiped with sterile cotton swabs, which were subsequently seeded with 10,000 or 1000 CFU of formalin-inactivated B. anthracis Sterne-strain spores. Nasal swabs were extracted in 1 ml of extraction medium, and the samples were treated as described for the lettuce supernatant in (A). The treated samples were tested as described in Fig. 1 with a B cell line engineered to respond to B. anthracis spores.

The cell preparation protocol that yields optimal performance requires less than 1 hour of labor over 2 days (10) and can be performed on a scale that cost-effectively prepares enough cells for millions of assays. Prepared cells can be stored at room temperature for at least 2 days; refrigerated for at least 2 weeks; or frozen indefinitely, while retaining full activity.

The pathogen sensor described here provides an optimal combination of speed and sensitivity that is currently unmatched by any other identifier. Currently available immunoassay methods require at least 15 min and have a much higher limit of detection (1). Although an ultrafast PCR with detection of 5 CFU in only 9 min has been reported (2), the total assay requires at least 20 to 30 min to complete when coupled with the fastest sample-preparation technology (15). Another feature of CANARY is that the antibody expressed determines the cell specificity and can be tailored to a desired application. We have produced B cell lines that respond to just a single strain of foot-and-mouth disease virus (fig. S2) or to many strains of VEE virus (fig. S3), as well as cells that distinguish pathogenic O157:H7 E. coli from nonpathogenic E. coli strains (fig. S4). The speed, sensitivity, and specificity of CANARY are valuable attributes for applications including medical and agricultural diagnostics, biowarfare defense, and food- and water-quality monitoring.

Supporting Online Material

www.sciencemag.org/cgi/content/full/301/5630/213/DC1

Materials and Methods

Figs. S1 to S4

References

References and Notes

View Abstract

Navigate This Article