Multiplexing—assays with multiple personalities

October 2007
Feature Story

William Check, PhD

First there were assays for sodium, glucose, cholesterol, and other simple analytes that were measured one assay per tube. Next came the automated chemistry analyzer, which measures many samples at a time but still does only one assay per tube. More recently has arisen the need to detect many molecules to formulate one result, as embodied in genetic assays, which detect a gene by probing for its oligonucleotide fragments. In addition, it's now recognized that it would be desirable in some situations to measure many related pathogenic elements simultaneously—autoantibodies, respiratory viruses, bacteria that can cause sepsis. In the research laboratory now, but looming on the clinical horizon, are methods to evaluate changes in dozens or hundreds of proteins or nucleic acids—for instance, cytokines in inflammation or microRNAs (miRNAs) in cancer.

Microarrays based on solid supports provide one method for multiplexing—measuring many analytes simultaneously on the same sample, particularly for nucleic acids. Real-time PCR can also perform multiplexing, though the complexity of its reactions limits it to detecting only a few oligonucleotides per tube. Another approach to multiplexing that many find attractive is what is referred to as a "fluid microarray" or "liquid bead array" —microspheres suspended in a buffer to which capture probes are attached and to which target molecules can bind. Reactions on this type of microarray support have certain theoretical advantages. According to Luminex Corp., which markets the beads (called the xMAP system) and associated readers, the high ratio of surface area to volume of the microspheres and the three-dimensional exposure they provide yield superior kinetics relative to flat arrays.

Certainly, companies that have developed assays based on liquid bead array technology (xMAP is an open-architecture system) find practical advantages to the method. At Planet xMAP USA 2007, a conference sponsored by Luminex last spring, Harry Prince, PhD, scientific director of immunology at Focus Diagnostics, which has devised an assay that distinguishes herpes simplex virus types 1 and 2, said obtaining multiple results in the same well "saves labor, cost, and time and decreases turnaround time." Tm Bioscience, now a division of Luminex, developed a clinical assay, using the xMAP system, for a panel of respiratory viruses. "We are at least as sensitive as real-time PCR and in some cases more sensitive," Richard Janeczko, PhD, chief scientific officer of Tm Bioscience, said of the assay.

Scientists at Quest Diagnostics Nichols Institute have developed and validated two genetic assays using liquid bead array technology: an Ashkenazi Jewish panel and a thrombophilia panel. Of the problem of doing assays for different genes on different platforms, Weimin Sun, PhD, of Quest said, "Multiplexing is the solution." The advantages of the xMAP-based assays are fewer controls, better turnaround time, a saving of labor owing to fewer DNA extractions, and being able to run a full microtiter plate of samples much of the time, according to Dr. Sun, who is scientific director of the molecular genetics department at Quest.

Perhaps the highest endorsement of xMAP's potential was given by a group at the Broad Institute of the Massachusetts Institute of Technology and Harvard. They measured gene expression using a combination of ligation-mediated amplification with "an optically addressed microsphere and flow cytometric detection system" —xMAP. This combination, they wrote, has the potential to be a "transformative technology" in the gene expression field (Peck D, et al. Genome Biol. 2006;7:R61).

A variety of applications of xMAP have been devised or are under development, from the immediately clinical to the futuristic. They range from assays for a few analytes, such as a 10-autoantibody panel, to one that stretches the potential of the technology—measuring hundreds of miRNAs. But they all depend on the same fundamental reagents and procedures, which, along with underlying principles, are presented and illustrated on Luminex's Web site (www.luminexcorp.com/technology/index.html).

Basically, 5.6-micron polystyrene microspheres are impregnated with differing ratios of red and infrared fluorophores, giving rise to 100 distinct beads. Capture reagents, typically oligonucleotides or proteins, can be coupled to the beads, with a different capture reagent for each bead. A complete set of beads, up to 100, is incubated in each well of a 96-well microtiter plate with one test sample per well. Then a reporter molecule is added that attaches only to those beads to which a target molecule has bound. For detection, the beads from a single well are passed through two lasers in single file using cytometer technology. A red laser identifies each bead by its unique spectral signature, and a green laser tells whether an analyte is bound to each bead.

Even when an assay does not have 100 elements, the system's high capacity can offer an advantage. For instance, the AtheNA Multi-Lyte autoantibody panels, distributed by Inverness Medical Professional Diagnostics, use the extra beads to obtain a standard curve in every well.

Information on Tm Bioscience's respiratory virus panel assay was provided in the March 2007 CAP TODAY article, "All in One and One for All: Multiplex Testing." At the Luminex conference last spring, James B. Mahony, PhD, director of McMaster University Regional Virology and Chlamydiology Laboratory at St. Joseph's Healthcare Hamilton, Ontario, Canada, presented further clinical data on the performance of this multiplex assay, which measures 19 respiratory viruses. Nearly 30 percent of cases of community-acquired pneumonias admitted to the hospital go undiagnosed, Dr. Mahony pointed out. One goal of developing the nucleic-acid-amplification-based respiratory virus panel, or RVP, was to have one test that would detect all known respiratory viruses, including SARS and influenza H5N1, and that would have a one-day turnaround time. The RVP assay effectively achieves same-day reporting, Dr. Mahony said.

Analytical sensitivity of the RVP assay is 0.1 to 10 TCID₅₀ (50 to 100 genome copies), more than adequate for clinical purposes. Six sites have completed their clinical evaluations, and data have been submitted to the Food and Drug Administration. Specificity was greater than 99 percent and sensitivity was 98.5 percent, compared with 63.2 percent for direct fluorescent antibody (DFA)/culture. Consistently, five percent to eight percent dual positive specimens were seen with the RVP. "We don't know the clinical significance of these dual positives," Dr. Mahony said.

At the McMaster clinical site, among 227 patients, the RVP detected all 125 specimens containing one of seven conventional viruses, while DFA detected 113 (90.4 percent). Considering all 19 viruses, the RVP detected 151/152 (99.3 percent), while DFA detected 110 (73.8 percent).

Other opportunities to apply multiplex technology, Dr. Mahony said, are gastroenteritis viruses, norovirus, and viruses that infect the central nervous system (herpes simplex virus, enterovirus, and varicella-zoster virus).

Dr. Prince of Focus Diagnostics presented data on the company's Plexus HerpeSelect 1 and 2 IgG assay, which detects type-specific herpes simplex virus antibody. This information can help guide therapy: Since HSV-1 has a very low recurrence rate, it can be treated with symptomatic, rather than suppressive, therapy. Type-specific serology is a useful supplement when lesions are negative for virus, when lesions cannot be sampled or are healed, or when the patient does not have classic symptoms.

Plexus HerpeSelect (HS) 1 and 2 has excellent sensitivity and specificity, Dr. Prince reported. And it turns positive earlier than the current gold standard, Western blot. For HSV-1, Western blot turns positive at 33 days, compared with 21 for Plexus HS; the corresponding figures for HSV-2 are 47 and 25. Plexus HS is done in a 96-well plate, like Focus' type-specific ELISA for HSV; turnaround time is 90 minutes for Plexus HS, compared with 120 minutes for ELISA. And, because of its multiplex nature, Plexus HS gives results for three analytes per plate, compared with one for ELISA.

Plexus HS demonstrated comparable performance to the HerpeSelect ELISA in clinical trials using 600 serum samples—300 from expectant mothers and 300 from sexually active adults. Overall sensitivity of the Plexus HS was 93.5 percent and 94.4 percent for HSV-1 in the two populations and 95.6 percent for HSV-2. Specificity was 94.4 percent for HSV-1 and 96.5 percent for HSV-2. Focus received FDA 510(k) clearance earlier this year to market Plexus HS.

(In all, 42 assays using xMAP technology are 510(k) cleared.)

Multiplex assays for bacterial pathogens that cause bloodstream infections are being evaluated by James Versalovic, MD, PhD, director of the Division of Molecular Pathology and of the microbiology laboratories, Texas Children's Hospital, and associate professor of pathology, molecular and human genetics and molecular virology and microbiology, Baylor College of Medicine. According to a 2003 report, the incidence of sepsis in the U.S. increased between 1979 and 2000 from 82.7 per 100,000 population to 240.4 per 100,000, for an annualized rate of increase of 8.7 percent (Martin GS, et al. N Engl J Med. 2003;348:1546-1554). Gram-positive organisms and fungi are now more common causes of sepsis. The 10 most frequent pathogens cause 75 percent of bloodstream infections.

Dr. Versalovic has explored whether molecular diagnostics can reduce the burden of undiagnosed infections and improve diagnostic accuracy. He posed the challenge: "Can we detect organisms directly in positive blood cultures?" Currently, after a positive culture, organisms are subcultured onto plates for biochemical testing for identification. Dr. Versalovic wants to know if molecular testing can improve this process.

Since PCR is "focused and limited," he said, in his laboratory he is working with two global methods: DNA microarrays and liquid bead arrays. In December 2003 his laboratory began doing DNA pyrosequencing of the amplified V1 and V3 regions of 16S rRNA genes. By May 2006 the diagnostic microbiology laboratory at Texas Children's Hospital had tested 363 samples by this method, 41 percent from blood and 38 percent from respiratory specimens, with a focus on isolates that can't be identified from standard cultures. Positive growth in blood culture was sent for Gram stain and biochemical identification. If isolates could not be identified by biochemical testing, organisms were submitted for PCR and DNA pyrosequencing. Dr. Versalovic found that DNA sequencing, coupled with conventional microbiologic methods, provided a genus- or species-level identification for about 90 percent of isolates that lacked a definitive identification by biochemical testing (Luna RA, et al. J Clin Microbiol. 2007;45:2985-2992). The investigators concluded, "Coupled with isolation by bacteriologic culture and biochemical testing, DNA pyrosequencing-based bacterial identification was a valuable tool that markedly improved bacterial pathogen identification in a pediatric hospital setting."

Dr. Versalovic said "very preliminary" data showed that liquid bead arrays targeting the V1 or V3 region, or both regions, can identify many gram-positive organisms, including S. aureus, S. epidermidis, S. pneumoniae, P. aeruginosa, and E. coli, and can distinguish enterococci from streptococci and staphylococci. Additional pathogens, such as E. faecalis and E. faecium, can be distinguished using the V6 region of the 16S rRNA gene. "So we can make a preliminary sepsis panel using only three primer sets," he said. For improving the diagnosis of sepsis, Dr. Versalovic said, multiplex testing with arrays is a "promising new strategy." Mary McBride, PhD, formerly deputy program leader for science and technology, Chemical and Biological National Security Program, Lawrence Livermore National Laboratory (now manager of the Molecular Test Program at Agilent Laboratories), described a point-of-care instrument for respiratory viruses based on xMAP technology. Called FluIDx, the 16-plex POC instrument was designed to detect influenza A and B and other high-priority respiratory pathogens—respiratory syncytial virus, parainfluenza 1 and 3, and adenovirus.

A patient presenting to the emergency department spends on average four hours waiting for a diagnosis; those with respiratory diseases wait even longer, Dr. McBride said. So a test that gives a result in a fashion that is timely enough to have an impact on treatment "constitutes a POC diagnostic." In the new POC instrument, no nucleic acid extraction is necessary. Material from a nasal swab is directly amplified and tested. In preclinical evaluation the lower limit of detection was found to be 0.1 to 10 TCID/mL, or 50 to 1,000 copies per mL. Early work shows that it performs well except for parainfluenza virus. A prototype POC instrument is being evaluated in the emergency department at the University of California Davis Medical Center where it is being used by "minimally trained" staff.

Multiplex testing can add value in medical genetics with its expanding repertoires, said Dr. Sun of Quest/Nichols. For instance, Tm Bioscience's cystic fibrosis panel has expanded from 23 to 70 mutations, and the Ashkenazi Jewish panel, which started with Tay-Sachs, is now up to 12 diseases. For some assays, the different mutations or alleles are tested on different platforms. For thrombophilia, for instance, Quest used four separate assays. Multiplexing offers an opportunity to combine these assays. Dr. Sun said Quest chose xMAP technology because it now has the best multiplexing capacity.

A 30-mutation Ashkenazi Jewish panel based on Tm Bioscience xMAP reagents was evaluated in 820 samples (Strom CJ, et al. Genet Med. 2005;7:633-639). Patient or control material for 24 of the 30 mutations was available. There was 100 percent concordance between the multiplex method and Quest's existing technique in 19,680 genotyping calls with 112 positive cases, Dr. Sun said. No errors occurred with either method. Dr. Sun and her colleagues concluded, "The Tm Bioscience Ashkenazi Jewish analyte-specific reagent ... can be used with confidence in the clinical molecular genetics laboratory."

A multiplex thrombophilia assay that detects five markers in three genes was devised using Celera Diagnostics reagents. In almost 1,000 samples with more than 4,000 genotypes, there was one discrepant call and a repeat rate of less than two percent.

Quest/Nichols is now using these two multiplex assays routinely.

One problem Dr. Sun noted with multiplex assays is how to handle results for tests not ordered. Their solution is to mask the raw data from tests not ordered, both bead counts and interpretations. Another problem is how to bill only for tests and procedures (DNA extractions) performed. To address this, each test has been constructed to have two components, an assay-specific procedure and a process common to that assay group.

Dr. Sun added two further considerations: Storage conditions are critical for Luminex bead stability, and the multiplex assay reduced setups by 21 percent for the Ashkenazi Jewish panel and by 30 percent for the thrombophilia panel.

For an example of a multiplex assay that pushes the limits of the xMAP technology, one need look no further than the assay for miRNA being developed under Keld Sorensen, PhD, director of research and development in the Luminex Bioscience Group. Discovered recently, miRNAs appear to have fundamental functions in many biological processes, from normal development to oncogenesis. Called Flexmir, the assay combines liquid bead multiplexing with locked nucleic acid, or LNA, technology. LNA is simply a chemical modification of the sugar ring in an RNA nucleotide that locks the sugar in the 3'-endo conformation. Practically, this modification increases the sensitivity and specificity of hybridization, increasing the melting temperature for DNA by two degrees. According to Dr. Sorensen, LNA technology greatly increases the affinity of probes for their complementary, mature miRNA targets and allows single-base discrimination of closely related miRNAs. In Flexmir capture probes, some bases are locked and others are not.

"Combining LNA technology with xMAP liquid bead arrays allows the design of miRNA assays that are both flexible and very specific," Dr. Sorensen said. The assay is performed on a total RNA sample. Assay time is about four hours, and throughput can reach thousands of data points per day. Currently, Flexmir contains capture probes for 320 miRNAs in five bead pools. Sensitivity is as low as 0.5 fmol. Dr. Sorensen's group is now using Flexmir to analyze normal tissues and cancers.

For gene expression work generally, the xMAP system seems to be effective. When the group at the Broad Institute of MIT and Harvard suggested that liquid bead arrays could be part of a "transformative technology," they were basing their conclusion on specific qualities of this method. Gene expression signatures consisting of small numbers of genes—10 to 20—are being found to be prognostic in medical settings. "However," they noted, "the lack of a practical and cost-effective technology for detection of these gene expression 'signatures' in large numbers of samples has severely limited their exploitation in important medical and pharmaceutical discovery applications." Their evaluation showed that the combination of ligation-mediated amplification and the Luminex xMAP system provides "a simple, flexible, cost-effective, and high-throughput gene expression signature analysis solution tailored for the measurement of up to 100 transcripts in many thousands of samples," they wrote. No doubt this approach will be used increasingly.