The sequencing of the human genome and the refining of molecular technologies have given doctors new opportunities to treat individual patients, not just their diseases. The narrowing gap between diagnosis and therapy has empowered both patients and health professionals, and has ushered in what is already hailed as a new era in healthcare.

IVD companies have made incredible strides in pharmacogenomics over the last few years; still, the release and approval of molecular tests have been less frequent than many had anticipated. Although some of the challenges for manufacturers are rooted in the enormous technological complexity of this new generation of tests, structural obstacles in the healthcare industry have also shared some of the blame.

To learn more about what manufacturers can do to encourage the spread of molecular testing, IVD Technology editor Richard Park spoke with David Persing, MD, PhD, executive vice president and chief medical and technology officer at Cepheid (Sunnyvale, CA). In this interview, Persing talks about what changes will need to be made for molecular testing to reach its potential. He also discusses trends in real-time polymerase chain reaction (PCR) and multiplexing, and how diagnostics manufacturers can produce more-laboratory-friendly tests.

IVD Technology: What have been the biggest technological developments and advances in molecular diagnostics during the last few years?

David Persing: I would have to point to real-time PCR as being the biggest advance. The instruments in which the reaction is carried out are self-contained systems for amplification and simultaneous detection.

Because these are closed systems, they don't require open manipulation of the amplification product. As a result, the technology is much less prone to contamination. Before real-time instrumentation, the problem of contamination had been one of the biggest barriers to the widespread introduction of PCR technology into clinical laboratories.

Labs no longer need to develop special facilities and procedures to address the problem. Being able to carry out PCR in the presence of the detection reagents in a closed system has had a huge impact on diagnostics and has created the potential for the technology to finally enter into the mainstream.

The other big area is multiplexing—the ability to carry out multiple analyses in a single reaction. This allows us to build controls into the reactions for quantitation purposes. In fact, we can build in multiple internal controls at various ranges of detection to allow accurate quantitation over a wide dynamic range. That's one benefit of multiplexing.

Another benefit is being able to carry out complex analyses of multiple targets, each of which may contribute to a diagnostic picture. To be properly interpreted, some tests require pieces of quantitative information from separate targets.

So, for instance, gene-expression analyses of breast cancer biopsies may require 15 or 20 targets to provide a complete picture of a particular pathological predilection. This is the next complexity frontier, and it shows where the diagnostics field is heading, especially on the oncology side.

What do you view as the latest trends?

I believe that the two areas I mentioned—real-time PCR and multiplexing, either by direct amplification or by use of array technologies—are emerging as trends in the field.

Having said that, it's surprising that, given how long real-time PCR technology has been available, there are only three FDA-approved tests that make use of it. Even after close to 10 years, there aren't FDA-approved tests offered by the major diagnostics manufacturers.

They're coming instead from smaller companies, including Cepheid and Genome IDI, which was recently acquired by Becton Dickinson. Even though the larger companies had early access to the technology and had plenty of opportunities to exploit it, they chose alternative strategies that most in the field now consider to be inferior to real-time PCR. They've gotten entrenched with these older methods and it's taken them time to reinvent the quantitative aspect of the real-time PCR technology.

There are some inherent challenges in multiplex technologies. Are these compounded when developing molecular diagnostics?

There are advantages and disadvantages to the technologies. One of the advantages of working off the same sample is that there is less variability. There's a built-in control in the system that's not available when, for example, researchers obtain multiple biopsy specimens and try to correlate the information. In the case of a tumor, different specimens may contain varying amounts of stroma, connective tissue, and other materials that could alter test results.

Of course, analyzing gene-expression profiles from the same specimen also poses a considerable sensitivity challenge. Often, a researcher will have to divide a sensitive diagnostic technique among what amounts to very small quantities of extracted nucleic acid. For a needle biopsy specimen, there is often only a very small amount of tissue that's actually available. So, the sensitivity requirements for molecular diagnostics need to err on the side of high sensitivity. This provides the greatest flexibility for laboratorians to be able to provide access to diagnostic technologies.

In molecular design, there are also issues surrounding multiplexing. Those of us who've developed diagnostic tests for a while all know that a multiplexed assay is a bit like a recipe for a soufflé. Both can be very finicky, and if you throw a new ingredient into the assay—maybe a new primer design or a new probe—it can have disruptive effects.

As a result, the development time for multiplexed assays tends to be longer. If you're asking for each one of those targets to be quantitative as well, the process can become even more complicated. The need for quantitation imposes an additional requirement for controls, perhaps at the level of each target within the multiplex assay.

Despite the issues and complications related to developing multiplex assays, I do think that's where things are headed. In general, high-level multiplexing is going to require access to more colors—that is to say, different channels of fluorescence detection in which amplification products can be set during real-time PCR.

We're going to need dyes that allow more channels to be detected. Cepheid is going from a four-color system to a six-color system by the end of this year. And we are developing our own dyes and systems to go with six-channel detection. There are plenty of examples right now in which the assays that we're developing are going to exploit all six channels of detection.

Expanding the IVD Market

With these emerging trends, what are the primary obstacles in developing and selling molecular diagnostics?

If you look at the universe of laboratories that are performing molecular diagnostics now, it's a fairly restricted number, especially those that can be considered full-service molecular diagnostics labs. You're dealing with only 400 or 500 laboratories in the entire country that fit this category.

Then there is a second tier of laboratories—about 2200—that are still high-complexity rated per the Clinical Laboratory Improvement Amendments (CLIA) standards. Beyond that, there are about 5000 laboratories that have the CLIA rating of high complexity but which do not run molecular diagnostic tests of any kind.

And yet, the 7000 or so laboratories that fall into the high-complexity CLIA rating represent only a fraction of the potential laboratory universe. There are approximately 27,000 laboratories that are able to run moderate-complexity tests but don't have the specialized facilities and staff to run highly complex molecular diagnostic assays.

So, in terms of what the major challenges and opportunities are, I think the IVD industry has to do a better job of democratizing the technology. We've got an ideal platform in real-time PCR, but we've not done a good job of commercializing it and reducing the complexity of the testing to make it more available.

I think the next trends are going to be to broaden the marketplace, to allow laboratories that fall outside of the high-complexity categories to run tests locally. These technologies are better for the patient because they return results quicker. They're better for the hospital because running more tests locally reduces the costs of sending out assays for analysis. In fact, hospitals can turn what has long been a cost center into a revenue center. And I think these technologies are better for the doctors running the tests as well, because they don't have to wait for the results. The time it takes to transport tests from the hospital to a centralized reference lab can be a major barrier to getting quick test results.

The trends point toward moving sophisticated testing closer to the patient. The key is going to be doing so with- out sacrificing test quality. Another challenge will be using these tests in a moderate-complexity environment, where they can be analyzed by a histopathologist or a surgical pathology lab tech with no formal training in molecular diagnostics and without a specialized facility dedicated to molecular diagnostics. That's the key in my mind to making this technology a commercial success.

At the same time, I really do think there is increasing justification for bringing the technology closer to the patient. For instance, there may be a real demand for drug-resistant organism testing within intensive-care units (ICUs) and other hospital environments, to know within an hour whether a patient coming into the unit is colonized or infected with methicillin-resistant Staphylococcus aureus (MRSA). The ideal scenario is one where a nasal swab from a patient coming into the ICU can be tested on the spot.

To receive these test results quickly is to be able to generate actionable intelligence regarding management of patients colonized or infected with MRSA. Hospitals can isolate the patient appropriately and minimize exposure of hospital staff and other patients. What often happens now is that patients are admitted without this information, and several days pass before it's discovered that they're MRSA colonized or infected. By this point, much of the harm has already been done.

This democratization of the technology is something that needs to happen in molecular diagnostics; it just hasn't happened yet on the part of many laboratorians because of the turf issues surrounding molecular diagnostics. There's going to be some resistance because of the restricted nature of the technology. Some may resent the democratization of the technology and view it as competition for their own existing operation. Others may welcome it as a means of expanding their testing capabilities more and bringing sophisticated tests in-house.

So, it's not going to be an easily solved problem. It's not going to be an easily solved marketing challenge. But it's something that's going to happen eventually.

What can IVD manufacturers do to address this issue?

I think they have to improve their tests from the beginning of the development process. New products need to be designed with moderate complexity in mind. IVD manufacturers need to ask themselves how they can design a diagnostic system that meets the moderate-complexity requirements of CLIA.

That's the only way to truly democratize the technology. That's the only way that you'll be able to see a lab technician run a real-time PCR test for viral meningitis on a sick 3-month-old baby in the middle of the night.

Evening laboratory staff in, for example, the stat laboratory are not necessarily going to be well suited to run a high-complexity test, even in a laboratory environment. The movement toward moderate complexity is going to provide benefits both at the laboratory level and at the point of care.

Manufacturers need to focus on developing tests that can meet these requirements. The requirements for moderate complexity are well defined. Doing so will help open up new markets and benefit patients and doctors alike.

Finding New Targets

Are developments in molecular diagnostics more likely to be driven by traditional targets like infectious or sexually transmitted diseases, or is the field going to start to look more toward genetic mutations as a basis for diseases?

I think there are going to be new opportunities. In the past decade or so, much of the market has been driven by high-throughput, high-volume infectious-disease testing, such as for HIV and hepatitis C.

In the next few years, we may begin to see second-tier diagnostic products being developed for quantitative cytomegalovirus determination, for example. Testing for this type of target is not at the volume of HIV and hepatitis C testing, but it's an emerging market and a potentially high-margin opportunity. The transplant population, for one, requires sophisticated testing, and with this comes a higher price and potentially higher margins.

On the mutation side, we're going to see a lot of interest in pharmacogenetic testing. FDA recently recommended that the patients receiving the anticoagulant warfarin be tested for one of the cytochrome P450 enzymes involved in the metabolism of the drug. I've heard that recipients of the breast cancer drug tamoxifen may end up getting a recommendation from FDA regarding P450 testing for polymorphisms in the 2D6 enzyme. There's also news that there are some genetic polymorphisms associated with the enhanced efficacy of Xigris, a sepsis drug from Eli Lilly and Co.

These kinds of diagnostic/therapeutic connections are going to become stronger over the years. There is considerable interest in the connection between the administration and metabolism of drugs, or the association of certain genetic markers with a higher likelihood of side effects from drugs. This area is a huge opportunity for diagnostics that will bridge the therapeutic and diagnostic sides of patient treatment in unprecedented ways. We're seeing just a few examples of this now, but I think that activity in this area is going to increase.

Going back to my original point about the democratization of the technology, sepsis patients who might be getting Xigris don't come to the ICU in batches. They come in as single patients, often in the middle of the night, and you can't wait for a genotyping test to be sent off to an outside laboratory and to come back days to a week later. Treatment decisions regarding dosing or a particular drug selection need to be made much quicker than that.

Drug companies don't want to be hamstrung by a diagnostic test because that's going to reduce the likelihood that their drug will be administered in the proper way. So, they're very anxious to see the democratization of this technology as well. If a test for a 2C9 polymorphism is required to determine the right warfarin dose, and a patient comes in with a pulmonary embolism or deep vein thrombosis, you can't wait for the test result to be performed as part of a larger batch which is run in the molecular diagnostics lab three times a week. It needs to be performed on a stat basis with timing that's commensurate with the administration of the drug.

Personalizing Personalized Medicine

How will the continuing emergence of theranostics, personalized medicine, and pharmacogenomics affect the development of molecular tests?

As I indicated earlier, I strongly believe that making the technology for pharmacogenomics more widely available is a critical-path item for widespread patient adoption. If there are pharmacogenetic tests that need to be performed prior to the prescription of a drug—to determine a proper dose or to help identify patients at risk of side effects that preclude them from therapy—these tests need to be available in a manner that's commensurate with the management of the disease itself.

For a patient with disease A, there may be plenty of time to make these types of decisions, plenty of time to send a test swab to a central laboratory facility across the country and receive a result a week later. But for a patient with disease B, for which a decision needs to be made quickly, waiting this long is just not practical.

My point, then, is that diagnostics manufacturers should focus not just on the target but also on the way the technology is going to be used to prescribe a particular drug.

With pharmacogenomics again gaining attention, do you think that molecular diagnostics will finally get a push to reach its growth potential?

Within transcriptional profiling of cancer, which is a large growth segment, there are lots of opportunities for tumor detection, detection of minimal residual disease, and profiling cancers for the purpose of determining their potential for metastatic spread or susceptibility to therapy. This area has been a direct spin-off of the Human Genome Project. I'd put it on an equal footing with pharmacogenetics in terms of high-growth opportunities for diagnostics companies.

Have money issues such as reimbursement also proven a major stumbling block in the wider adoption of molecular technologies?

It's the chicken and egg dilemma. We have reimbursement guidelines for quantitative HIV viral-load assays that are pretty reasonable, and manufacturers need to design their technologies to be able to fit within these reimbursement levels. On the other hand, the agencies involved in reimbursement need to recognize the value of diagnostics-driven outcomes and ultimately provide appropriate funding for those tests. They need to realize that in providing better overall patient care, they're going to save more individualized money by avoiding unnecessary complications.

The Road Ahead

What trends and challenges can we expect to see next year, and further into the future?

I think future trends will include a greater dependence on real-time PCR. It is the most powerful technology out there for quantitation and detection, and it's the one that has the greatest opportunity for democratization.

And in that vein, I think the trend is going to be toward increased movement of assays out of the high-complexity laboratory environment. We'll see more movement away from batch-mode testing. Right now, most molecular diagnostics assays are run in batches. However, this often requires patients to wait for results until batch sizes are adequate to justify running a test. This can also restrict the throughput of the assay. Batch-mode tests are typically run by specialized techs in specialized laboratories. This restricts where these tests can be run.

As a result, I think we'll see a movement away from batch mode to smaller batch sizes that provide faster turnaround times. This could even reach the level of point-of-care-type tests, which are run with built-in controls on a one-off basis.

Jan S. Krouwer, PhD, is president of Krouwer Consulting (Sherborn, MA), which provides statistical and reliability consulting to the IVD industry. He can be reached at jan.krouwer@comcast.net.

FDA's proposed CLIA waiver guidance is different from previous waiver guidelines.¹ This article describes the differences to help IVD manufacturers understand the proposed guideline.

Waiver assays are “simple laboratory examinations and procedures that are cleared by FDA for home use; employ methodologies that are so simple and accurate as to render the likelihood of erroneous results negligible; or pose no reasonable risk of harm to the patient if the test is performed incorrectly.”² The advantage of achieving a waiver classification for assays is that labs with a waiver certificate can perform such assays. Since 63% of labs have a waiver certificate, an assay with waiver status can be sold to many more labs.

The key difference between the proposed and previous waiver guidances is the assessment of performance. Historically, most assay evaluations obtain performance information by estimating one of the following:

• Average bias and imprecision.
• 95% of the differences between the assay results and a comparison method.³
• 95% uncertainty interval for the results.⁴

While FDA has traditionally used the first method, the second method is appearing more often in the literature. However, for the second and third methods, performance goals are often lacking. Although much has been written about goals for medically acceptable differences between assay results and reference methods, there are few standards that address this issue.⁵ One standard that has medical acceptability limits is ISO 15197 for home-use glucose assays. This standard states that in order to achieve medically acceptable performance, 95% of the results should be within stated limits.⁶ But if this goal were met, then for every 1 million results, up to 50,000 medically unacceptable results would be allowed.

Figure 1. Total error and outliers.

What is missing in ISO 15197 is presented in Figure 1, which adds limits for outliers. Outlier is a generic term for a large error. Outliers can be classified as either a statistical outlier, which fulfills the criteria of a specific test such as being greater than 3 standard deviations from the mean, or an erroneous result, which exceeds a stated limit and may or may not be a statistical outlier. The proposed waiver guidance document focuses on erroneous results.

Based on ISO 15197, 95% of test results should be within region A in Figure 1. Values that are just outside of region A would probably not cause problems, which is shown as region B. However, results in region C will likely cause patient harm. In addition to requiring that most (95%) of the values fall in region A, very few (ideally zero) results should be in region C.

IVD manufacturers are required by the waiver guidance to implement risk management tools (e.g., hazard analysis, failures mode and effects analysis (FMEA), fault trees) to ensure that severe test errors do not occur.⁷ For example, hazard analysis looks for ways in which severe errors are identified and mitigations are established to prevent such errors from occurring. While risk management is not really new, considering the design control requirements for hazard analysis, the proposed waiver guidance includes an additional emphasis that reinforces the importance of hazard analysis to ensure that dangerous test results are not released.

What is new in the guideline is a different analysis for a method comparison study. Risk management differs from method comparison studies in several ways. For example, a hazard analysis is a model that attempts to predict how a system can fail. The problem with any model is that there is no guarantee it is correct. A method comparison is an empirical study that simply collects results. A big difference from the reference in a method comparison study may be observed regardless of whether there is an explanation for it or not.

Method Comparison Study

Protocol. The protocol for a method comparison study requires 360 samples to be tested by both the candidate assay, or waiver method, and the reference or comparison method. The proposed CLIA waiver guidance permits the following three types of comparison methods:

• Standard reference methods.
• Comparison methods that are traceable to a reference method. Traceability means that “the results of measurement can be related to a stated reference method, usually a national or international standard, through an unbroken chain of calibrations of a measuring system or comparisons where measurement uncertainties have been documented at every step in the procedure.”^1,8 Traceability can partly be achieved mathematically from regression studies, which generate a slope of one and intercept of zero.
• Comparison methods that are similar to those stated above, except that the slope of one and intercept of zero are not achieved. If such comparison methods are used, it is suggested that IVD manufacturers should contact FDA.

The samples must be patient samples instead of controls. Unfortunately, the waiver guidance allows only 60 samples to be spiked, while at the same time requiring the samples to span the measurement range and represent equally low, medium, and high concentrations. If this number of spiked samples is not adequate, the guidance suggests contacting FDA. In addition, the intended lab operators of the tests must conduct the method comparison study. This change is an improvement over past guidances, which required operators with no previous laboratory experience.

Analysis. This review focuses on applying quantitative methods for the cases in which the comparison method is either a standard reference method or a comparison method traceable to a reference method.

Regression. Regression analysis is relegated to coming after calculating descriptive statistics. Rather than being the primary analysis method, regression is part of the background.

Total Error and Erroneous-Results Analysis. This is the main change between the proposed and previous waiver guidance, and follows recommendations and the standard EP21A by the Clinical Laboratory and Standards Institute (CLSI; Wayne, PA).^9,10 The proposed guidance is requiring that regions A and C in Figure 1 be quantified. In the guidance, FDA defined region A as the allowable total error (ATE) zone and regions C as the limits of erroneous results (LER) zones.

Figure 2. Total error and outliers.

FDA goes one step further than Figure 1 by requiring a Parkes glucose error grid for the assay that is being evaluated (see Figure 2).¹¹ This grid makes the LER zones specific to those cases that are likely to cause patient harm.

In Figure 2, at least 95% of test results should fall in region A, and no results should fall in region C. This implies that up to 5% of results could be in region B. The proposed waiver guidance could be clearer by defining region B, which is unnamed in the document. The guidance is also confusing since it discusses results within the ATE or LER zones; but from the context, the guidance actually means results within the ATE zone and up to, but not in, the LER zones.

What Does This Mean for IVD Manufacturers?

Table I. Key differences between the previous and proposed CLIA waiver guidances.

Since most IVD manufacturers rigorously pursue the investigation of outliers during product development, the goals of having 95% of test results in region A and 0% in region C should be met. Table I lists the main differences between the previous and proposed CLIA waiver guidances.

For many assays, IVD manufacturers will have to construct an error grid in order to calculate ATE and LER. There is a clear benefit to constructing such an error grid. For example, if a glucose test result is 160 mg/dl and the reference value is 40 mg/dl, this erroneous difference of 120 mg/dl falls in the most severe error region in an error grid.¹² However, this same 120 mg/dl error for another test (glucose result is 320 mg/dl, reference value is 200 mg/dl) falls in region B in an error grid. This result means that large errors by themselves can be tolerated as long as they are not in the LER zones.

From a chemistry standpoint one may wonder, did the second case happen by chance, meaning that on other occasions, could it turn out like the first case? Although there is more leeway for reference and comparison methods, they should provide unbiased and precise results. The error analysis measures differences, and the interpretation is that any differences are due to errors in the waiver method.

Such studies do not guarantee that an assay is completely free from erroneous results. For example, if there are no values beyond the LER zones in 360 samples, the estimated erroneous-result rate is 0%. However, the upper 95% confidence bound for this rate is 1%. (The guidance asks for confidence intervals to be calculated.) Therefore, IVD manufacturers can guarantee no more than 10,000 erroneous results (i.e., results that can cause patient harm) per 1 million reported results. This is not necessarily a problem with the waiver guidance, but rather a reflection of the fact that it is difficult to prove that rare events do not happen.

There is a problem with the waiver guidance with respect to the ATE zone. For assays that have CLIA limits, the guidance requires 95% of the test results to meet such limits. This differs from current CLIA goals that require only 80% of results to meet CLIA limits.¹³ To be consistent with the CLIA goals, the waiver guidance should expand the CLIA limits when requiring 95% of results to meet such limits.

Process Capability

While process capability statistics (e.g., Cpm) are largely unknown for IVD assays, the process capability concept is relevant to the proposed CLIA waiver guidance.¹⁴ For example, in Figure 1, if a distribution of test differences did not go beyond region B, this would be considered a capable process and would meet the waiver guidance requirements. However, if the distribution extended into region C, then the process would not be considered capable. A process that is not capable may have no statistical outliers and could also have no failed quality control results.

Traditional Evaluations

In traditional evaluations, the method comparison results are analyzed by regression analysis using a method such as the CLSI standard EP9A2. An imprecision study is also carried out by using the CLSI standard EP5A2. In both evaluations, the results would be tested for statistical outliers, and if outliers were found, they would be discarded. The results are expressed in terms of parameter estimates such as regression slopes and intercepts, average bias, and within-run, and longer-term imprecision. One of the problems with this type of evaluation is that severe erroneous results can be missed, which are precisely the results that can harm patients.¹⁵

The method comparison analysis in the proposed CLIA waiver guidance is easier. This analysis counts the number of results falling into the ATE and LER zones. The percentages based on the results, particularly the LER percentage, provide an estimate of the risk of the waiver assay.

Risk Management

The burden of having no results in the LER zones raises the importance of risk management. Risk management involves several tools, including FMEA, which can prevent potential errors, and failure review and corrective action system (FRACAS), which can prevent the recurrence of observed errors. Both tools can be aided by fault trees and process flowcharts. Perhaps the most important tool will be FRACAS, in which the assay is repeatedly tested under as close to actual use conditions as possible to expose as many problems as possible.¹⁶ As such problems are exposed, fail-safe systems will be put in place to prevent them.

In addition, flex studies are required, which are experiments to show what happens when a range of values are tested, including those exceeding the minimum and maximum allowed. This type of risk management relies less on a model but rather on exercising the system to expose problems, which are then corrected. The actual hazard analysis submitted would document the hazards and mitigations. As with any FDA study, IVD manufacturers should practice the intended protocol to ensure that goals are achieved, which is another benefit of conducting FRACAS during product development.

Regulator's Dilemma

Regardless of the results of the CLIA waiver protocol, regulators have to balance their decision to approve or reject an application based on not only the possible errors that might occur with the waiver assay (i.e., risk) but also its benefits. IVD assays provide valuable information, which if unavailable could increase morbidity and mortality. A waiver assay could increase the availability of an assay by having a lower selling price and by being performed in more labs. This availability would translate into more people being tested and a potentially lower morbidity and mortality.

Conclusion

The proposed CLIA waiver guidance has parts that are easier (e.g., operator and reference method requirements) and others that are more difficult (e.g., patient samples instead of controls) for IVD manufacturers. The analysis is also both simpler and more relevant. It remains to be seen whether this guidance will be the basis of a revision to the 510(k) or premarket approval (PMA) guidance.

Maxfield L. Williams is the director of policy and external affairs at COLA (Columbia, MD). He can be reached at mwilliams@cola.org.

Physician office laboratories are complex operations in part because of the state and federal regulations that govern them. The prospect of complying with such regulations can cause doctors to think twice before purchasing new IVD equipment. Since 60–70% of healthcare decisions are made using IVD tests, physician office laboratory directors must carefully decide if the new equipment will perform better than what they already have.

Deciding whether to purchase medical diagnostic equipment can be an arduous decision-making process because of the significant expense and the perceived burden of accreditation and compliance requirements. Physicians may view the equipment as being too far removed from what they need to treat patients effectively. They may also question the value of going through the regulatory process, as it can appear to be too bureaucratic.

Today's physician offices are better educated about running profitable labs. Such offices no longer consider labs to be merely another cost center. On the contrary, labs can be profitable entities when they are run effectively and efficiently. Running a lab effectively begins by selecting the correct test menu for the patient population. Choosing a test system encompassing tests that are appropriate for the practice and can be run cost-effectively is most desirable.

However, having an IVD system may not make sense if costly analytes are infrequently used. Savvy laboratory directors at physician offices will examine patterns associated with ordering tests by performing a patient test count for each analyte. For example, if a test is performed less than 20 times per month, a practice should consider outsourcing the test to a reference laboratory. Physician office laboratories may also build their in-house test menus around tests that require fast turnaround times to maximize patient care and benefit.

The CLIA Effect

Changes in IVD reimbursement policies and concerns over higher costs and more-stringent regulations have primarily contributed to the hesitancy by physicians to purchase complex testing equipment. By passing the Clinical Laboratory Improvement Amendments (CLIA) in 1988, the federal government established quality standards for all lab testing. Under CLIA, labs are required to extensively document their activities. The chilling effect of such heightened regulation on physician office laboratories was immediately apparent. For example, fewer labs than expected completed applications for a provisional certificate in May 1992, indicating that many had closed. When inspections were conducted for the first time to enforce the new regulations, 80% of the labs surveyed were found to be deficient in some way. By this time, the number of regulated labs had further declined by 8% since 1988.

The revised 2003 CLIA regulations made the requirements for operating a laboratory even more rigorous, which is significant for physician office labs. Under the new CLIA regulations, laboratories performing tests of moderate complexity must comply with many of the requirements of high-complexity labs. For example, laboratories must verify that a test's performance in the lab is similar to the IVD manufacturer's claims for accuracy, precision, and reportable range. This change, among others, requires laboratory directors and manufacturers to work together to ensure that such requirements are met.

The 2003 CLIA regulations also placed greater emphasis on effective quality control systems in the laboratory. At first glance, lab directors and staff may view such process control requirements as onerous and costly; for low-volume testing, the direct costs may sometimes outweigh the benefits. Running negative and positive controls each day when a lab performs few tests increases point-of-care testing costs. In some cases, continuing to test may not be cost-effective. However, in labs where the testing volume justifies expenditures for quality control, the reality is far different. The emphasis on quality control is intended to complement a laboratory's ongoing quality assessment of policies, processes, and procedures, and contributes to the design of internal lab systems to detect errors. The end goal is continuous improvement, not punishment. If laboratory personnel can embrace quality control as an effective tool rather than another compliance requirement, they may be less hesitant about moving forward.

In addition, the revised CLIA regulations sought to better acknowledge those new IVD technologies with more advanced internal quality control systems. If a testing system has such internal automatic monitors, the regulations allow labs to conduct less-frequent external quality control. However, laboratory directors must establish the lab's quality control program and make such critical decisions. This is yet another daunting challenge as laboratories are responsible for performing the due diligence for determining whether or not the IVD systems are an asset to their practices. Nonetheless, it is an opportunity for IVD manufacturers to partner more closely with the laboratories to establish a culture of quality and high-quality testing programs.

Despite such regulatory challenges since the advent of CLIA, physician practices have developed a greater understanding of the role of labs in providing customer service and quality patient care, and have consequently expanded services. By implementing a cost-effectiveness analysis of each test on a lab's menu, physicians are seeing greater revenue generation. This has lead to an evolution in doctors' perceptions of the lab from being a cost center to a profit center that can improve their businesses.

At the same time, the negative perceptions of IVD equipment purchases and the regulatory process are harmful to the IVD industry and laboratory medicine. Improving the quality of patient care through appropriate diagnostics should not be viewed as being more trouble than it is worth. This fear and misunderstanding may lead doctors to avoid the right IVD equipment for their labs. In many cases, doctors will select waived test systems, simply because they are free from routine oversight, and not because it is best for their practices. With such decisions, the labs may not be as effective and convenient for patients and clinicians as they could be. An IVD manufacturer's salespeople can serve as ambassadors for the continual improvement of patient care by understanding the regulatory process and compliance requirements.

Salespeople Can Help

An IVD manufacturer's sales force should educate their laboratory clients to establish a proper understanding of this highly regulated industry. This is the first step toward overcoming anxieties based on false assumptions. The salespeople themselves may sometimes feel overwhelmed by the large amount of information associated with the accreditation and regulatory processes. However, if they want to gain the trust of potential clients, they must become familiar enough with such processes in order to be an effective adviser and information resource.

Without sufficient knowledge of the accreditation processes, the salespeople cannot provide the information necessary to address their laboratory clients' concerns. Even with sufficient knowledge at hand, other challenges may emerge. For example, although laboratory leadership knows how to use test results, they may not be well versed in obtaining them. The details of running a lab may be new to lab directors, and they may look to the salesperson for help. While the salespeople do not have to memorize every regulation and compliance requirement, they should be familiar with the issues that are most likely to be of concern.

A salesperson should be able to explain how soon a laboratory can start testing with new IVD equipment, and what effect or benefits the equipment will have on the quality of patient care. Payment issues will also need to be addressed. If a question cannot be answered immediately, a salesperson should conduct research and provide the information as soon as possible. The salespeople should walk their laboratory clients through the process of verifying and establishing performance specifications for newly introduced systems. In addition to providing counsel on the design of control procedures, they should prepare a “how to” binder to help the physician labs. By doing so, the salespeople can forge a relationship with their clients and evolve beyond being merely order takers.

Maintaining the Relationship

The high cost and complexity of IVD equipment creates the need for intensive after-purchase care, an aspect of customer service that should be considered a part of doing business with physician office laboratories. Newly established laboratories and those with new directors may need more help. Knowledge of equipment and compliance standards remains a salesperson's most effective tool at this stage of the client relationship. However, after-purchase attention should be kept in mind from the beginning. The salespeople should share as much information with the laboratory client as early as possible. This can result in fewer concerns and less confusion after the sale. They should educate their clients before the purchase, which can also serve to prevent second-guessing later on.

Due diligence in education at every phase of the purchase process and beyond is the best strategy to achieve complete customer satisfaction and successful implementation. Beyond closing sales, educating the healthcare industry about the undeniable value of diagnostic technologies improves patient care and saves lives.

A lab technician evaluates PCR results using an iCycler by BioRad and a GenAmp 5700 sequence detector by Roche Molecular Systems. (Photo Courtesy Iris international inc.)

Diagnostic immunoassays can detect targets as dilute as a few million molecules per milliliter. However, many plasma proteins, including potential diagnostic markers and drug targets, may be present below this threshold.

The polymerase chain reaction (PCR) and other amplification methods can detect specific nucleic acid sequences at the single-copy level. A combination of immunoassay and PCR methodologies can, therefore, theoretically detect proteins and intact organisms at a sensitivity approaching that of its detection of nucleic acids.

The underpinnings of this strategy, which is known as immuno-PCR (IPCR), were first described in 1992.¹ This amplified DNA–immunoassay approach is similar to that of an enzyme immunoassay, which makes use of antibody binding reactions and intermediate washing steps. In the updated method, the enzyme label is replaced with a strand of DNA and detected by PCR amplification.

The early IPCR method described the detection of bovine serum albumin adsorbed onto microtiter plate wells through contact with a series of substances—antibody, followed by protein A-avidin chimera, and then biotinylated plasmid DNA. Detection by ethidium bromide–stained gels was semiquantitative. The extensive intermediate washing steps resulted in a cumbersome and lengthy process. In addition, the nonspecific binding of template DNA created background signal, which limited sensitivity. The resulting amplification products also produced significant contamination and a false-positive hazard, common to all DNA amplification– based assays. Even so, the IPCR inventors clearly demonstrated the detection of several hundred copies of protein, far eclipsing the sensitivity of the best radioimmunoassay or enzyme-linked immunosorbent assay (ELISA). Still, despite its analytical promise, IPCR has since been largely ignored.

The Evolution of IPCR

In 1993, a five-order improvement in sensitivity was reported over a control ELISA for the IPCR detection of human protooncogene ETS1 absorbed onto Immulon-4 strips.² A biotinylated 2.5-Kb Klenow fragment was used as the detection molecule. The DNA label in this case was coupled to streptavidin and attached to the primary antibody through a biotinylated secondary antibody.

Several investigators have also achieved highly sensitive IPCR in a sandwich immunoassay format by using reporter antibody labeled with a DNA molecule, and a capture antibody coating the surface of a microtiter well, bead, or particle.³ Human thyroid-stimulating hormone or chorionic gonadotropin is immobilized by specific binding to microtiter plate wells. The bound antigen is further complexed with reporter antibody labeled directly with a molecule of DNA. Washing between steps with detergent solution is required to remove nonspecifically bound conjugate. The amount of DNA label remaining associated with the microtiter well is assessed by PCR amplification, followed by gel electrophoresis and ethidium bromide staining. Detection limits for the sandwich IPCR assay format exceed those of conventional enzyme immunoassays by two to three orders of magnitude. In addition, researchers have explored two-antibody sandwich IPCR for the detection of several antigens with similar improvement over conventional sandwich ELISAs.⁴

A sandwich format of IPCR for the detection of multiple analytes has also been described.⁵ In this research, DNA labels of varying length were attached to multiple reporter antibodies to distinguish the presence of more than one antigen in a sample. The multiple DNA labels were detected simultaneously in the assay by observing the size of the PCR amplification product revealed by gel electrophoresis. As with all heterogeneous IPCR assays, stringent washing of the solid support was required to limit background signal from nonspecifically bound label DNA.

More recently, researchers have taken advantage of the presence of multiple copies of p24 antigen over RNA and have developed a real-time IPCR assay for the detection of HIV-1.⁶ In this instance, a commercial ELISA test employing coated TopYield strips from Nalge Nunc International (Rochester, NY), biotinylated reporter antibody, and streptavidin-HRP label was enhanced by the addition of 500 base pairs (bp) of biotinylated DNA and then subjected to real-time PCR. Direct nucleic acid amplification could detect about 50 copies/ml. The IPCR assay of diluted samples was able to detect the equivalent of 1.7 viral RNA/ml, with a range of three orders of magnitude.

The ability to use any sequence as the label DNA for IPCR has led researchers to investigate a unique approach to IPCR.⁷ Sandwich antibodies are labeled with one of a matched pair of oligonucleotides. The antibodies bind to antigen in proximity to one another. The oligonucleotide strands are complementary for only several bases on the free 3′ ends and they overlap, or hybridize, under appropriate conditions. Overlapping strands are self-priming for PCR at both 3′ sites and will each extend in the presence of DNA polymerase and nucleotide triphosphate. The DNA strands can only extend when overlapped, and overlapping occurs when the labeled antibodies are bound in proximity on an antigen.

The completed strands from the extension of each overlapping end represent new sequences that were not initially present on the antibody. The new sequences may then be exponentially amplified by the inclusion of downstream primers. Since the true DNA template is created only in response to the first chain extension in a sealed well, each individual Ab-DNA conjugate reagent cannot contribute to contamination. In addition, Ab-DNA conjugates composed of overlapping sequences exhibit much-reduced background signal after washing. Although these conjugates can bind nonspecifically to a solid support, this process is random. Background signal in this format requires the proximal nonspecific binding of two molecules, not just one.

In a similar approach called proximity probing, scientists have employed DNA aptamers such that proximal binding to platelet-derived growth factor B-chain (PDGF-BB) enhances enzymatic ligation.^8,9 The ends of the DNA hybridize to a common splint template, effectively creating a 10-bp-overlap oligonucleotide structure. This structure is treated with T4 DNA ligase, then amplified by PCR in the presence of primers and detected in real time with TaqMan probes from Applied Biosystems (Foster City, CA). As few as 24,000 molecules of PDGF-BB were detected, an improvement of 3 log orders over a standard sandwich ELISA.

The Nucleic Acid Detection Immunoassay

Overlapping oligonucleotide labels have been synthesized and used to produce conjugates of monoclonal sandwich antibodies for detecting a solution-phase antigen in a homogeneous format. This helps avoid the use of a capturing support and eliminates washing steps. The paired antibodies bind to separate epitopes on the antigen molecule at a distance that allows some fraction of the oligonucleotide labels to hybridize and extend in the presence of polymerase. The newly formed sequences are detected by real-time PCR. An affinity-purified polyclonal antibody has also been conjugated to demonstrate the detection of intact Escherichia coli 0157 cells by proximity binding to an antigen-embedded surface.

Methods

Prostate-specific antigen (PSA) and sandwich-paired monoclonal antibodies were obtained from BiosPacific Inc. (Emeryville, CA). Polyclonal antibody to Escherichia coli 0157 was obtained from KPL Inc. (Gaithersburg, MD). Oligonucleotides of 60 bases were synthesized to contain a functional amine attached to the 5′ end through a 12- carbon spacer arm from Glen Research Corp. (Sterling, VA) and purified by preparative polyacrylamide gel electrophoresis. The 5′ amino function was activated with a 100-fold excess of disuccinimidyl suberate to minimize cross-linking. The intermediate was rapidly purified by gel-filtration fast protein liquid chromatography (FPLC) in 5 mmol sodium citrate (pH 5.4) in order to maintain the second succimidyl function. The DNA was concentrated by centrifugal ultrafiltration at 4°C and combined immediately at room temperature with 10 mg/ml antibody in 0.3 mol phosphate buffer (pH 8) and 0.45 mol NaCl for 1 hour. Unreacted antibody was removed by size-exclusion FPLC using a Superose S-200 column from GE Healthcare (Piscataway, NJ) that had been equilibrated in Tris-buffered saline (pH 7.4). Unreacted oligonucleotide was removed by anion-exchange FPLC using a Mono Q column from GE Healthcare and 5%/min salt-gradient elution to 1 mol in 20 mmol Tris (pH 7.4). Typically, 50% of the protein was recovered as conjugate.

Figure 1. Using threshold cycle to measure MAb labeled with DNA. The label does not interfere with antigen binding.

Both native and sodium dodecyl sulfate (SDS) gel electrophoresis revealed the presence of antibody containing predominantly one or two strands of 60-mer. An overall MAb-DNA ratio of 1:1.6 was confirmed by absorbance ratio at 260 and 280 nm. Figure 1 shows the results of real-time PCR of a dilution series of MAb directly labeled with an oligonucleotide template strand. The presence of covalent antibody does not interfere with PCR signal. Likewise, the DNA label does not obstruct binding to antigen, as determined by HRP-labeled second-antibody detection of solid-phase antigen.

Oligonucleotides of 60 bases having the following sequences were used to form three DNA-antibody conjugates:

(a) 5′ NH3-C12-GCTACGGCTA GATCGTGTCCATGCGCTTAC GACTTCGATGCTCGGCTCGC TAGCTAGATG-3′

(b) 5′ NH3-C12-TCTCCAACTCTT CAACGCCATGTTCTTATGATAC GAGAGATTCAGCGGAGGCATC TAGCT-3′

(c) 5′ NH3-C12-TCTCCAACTCTT CAACGCCATGTTCTTATGATAC GAGAGATTCATCATCTAGCTAGC GAG-3′

Oligonucleotide sequence (a) is complementary to the other sequences, (b) and (c), for the last 9 and 15 bases, respectively, at the 3′ ends. The overlapping duplex has a specific Gibbs free energy (DG), which is controlled by the nearest neighbor base sequence.¹⁰ The overlaps of 9 and 15 base pairs have basic melt temperatures of 26 and 42°C in 50 mmol NaCl. The overlapping strands hybridize (>50% duplex) when present in solution at concentrations greater than 50 nmol at 25°C equilibrium. The strands exist predominately as monomers (<50% hybrid) in concentrations of less than 100 pmol at equilibrium.

Figure 2. A depiction of the first-chain extension. The new base pair (bp) DNA duplex includes new sequences.

According to the principles of thermodynamics, such DNA strands may transition between single- and double-stranded form as a function of temperature, concentration, and salt effects. When hybridized, each strand 3′ end can serve as the starting point for replicating the other strand. Each strand will extend in the presence of DNA polymerase and nucleotide triphosphates, resulting in a DNA duplex of 111 or 105 base pairs. The newly formed duplex contains sequences that were not present initially in the partially overlapped structure (see Figure 2). These new sequences can be replicated exponentially by PCR in the presence of two downstream primers—5′ GCTACGGCTAGATCGTGTCCA 3′ and 5′ TCTCCAACTCTTCAACG CATGTTC 3′. The initial oligonucleotide label strands cannot replicate in the presence of these primers without forming the first chain-extension product.

The first chain extension was performed at room temperature in the presence of Taq polymerase from Invitrogen Corp. (Carlsbad, CA) and dNTPs for 3 minutes. Real-time PCR was then performed using an iCycler iQ from Bio-Rad Laboratories Inc. (Hercules, CA) in the presence of 200 nmol downstream primers, 1:30,000 SYBR Green from Invitrogen, and 10 nmol fluorescein. Thermocycling was performed for 45 iterations of 1-minute extension at 62°C and 15 seconds denaturation at 95°C. The reaction volume was 50 µl.

Figure 3. Real-time PCR amplification of 9 and 15 base pair overlapping double-stranded DNA (dsDNA).

Figure 3 shows the real-time PCR amplification of the overlapping oli- gonucleotide strands of 9 and 15 base pairs. It can be seen that amplification does not occur until a sufficient concentration of strands are present in the solution to ensure the probability that one strand will be close enough to a complementary strand for hybridization and first-chain extension. Increasing the concentration of overlapping strands results in an exponential increase in template generation, as measured by real-time PCR threshold cycle. Conversely, diluting the concentration decreases the signal exponentially. Amplification of a normal 60-mer template is shown for comparison.

Procedures and Results

Figure 4. A graphic depiction of the homogeneous nucleic acid detection immunoassay (NADIA) process.

Homogeneous Format. Anti-PSA MAb1 has been labeled with oligonucleotide sequence (a), and MAb2 has been conjugated to sequences (b) and (c), using the methods previously described. A schematic representation of a homogeneous nucleic acid–detection immunoassay (NADIA) is given in Figure 4. The conjugate pair was diluted to 10–100 pmol in 10 mmol Tris (pH 8.0) containing 0.1% bovine serum albumin (BSA) and combined in the presence of PSA for 2 hours. The solution was then diluted with Tris/BSA to reduce the bulk conjugate concentration to below 1 pmol and was held at 52°C for 1 minute to fully melt unbound conjugate. PCR reagent mixture, containing Taq polymerase and downstream primers, was added, and the reaction was sealed. The temperature was lowered to 23°C to fully hybridize the DNA strands associated with the immune complex and to initiate the first chain extension. Free MAb-DNA cannot hybridize to the same degree in the time frame of the first extension in dilute solution, and cannot participate in subsequent exponential amplification. The overlapping DNA labels that were associated with the PSA immune complex were extended for 5 minutes, and completed by ramping the temperature to 85°C over 3 minutes. Real-time PCR amplification of the formed template was begun immediately, destroying the immune complex, which is no longer needed.

Figure 5. Results of an assay of prostate-specific antigen (PSA) by homogeneous NADIA. Sensitivity is approximately 100 fg/ml.

Figure 5 shows the results of the assay of PSA in a homogeneous NADIA format employing 15 base pair–overlapping MAb-DNA conjugates. The sensitivity was determined to be about 100 fg/ml. This represents about 500 molecules of PSA in the PCR reaction after the dilution step. The dilution step reduces the signal in a linear fashion, but reduces background exponentially, thus increasing the ratio of signal to noise (see Figure 3).

Results using the nine base pair–overlapping conjugates were similar when temperature conditions were adjusted for a lower melting temperature.

Heterogeneous Format. Affinity-purified polyclonal antibodies to Escherichia coli 0157 have also been conjugated to the overlapping oligonucleotide strands in order to demonstrate a heterogeneous NADIA format for the detection of intact microorganisms. The cell surface is estimated to exhibit several thousand copies of the specific antigen. Calculations show that the distance between randomly distributed sites should fall within the spanning distance of overlapping oligonucleotides conjugated to anti-Escherichia coli 0157. In this case, proximity binding is due to individual antigen spacing, as opposed to separate epitopes on a single protein antigen.

Figure 6. Results of a NADIA assay of live Escherichia coli with 9 and 15 base pair overlapping Ab-DNA.

Polyclonal anti-Escherichia coli 0157 was labeled in three separate reactions with the three oligonucleotide sequences and maintained as individual conjugates. Escherichia coli 0157:H7 was obtained from ATCC 700728 and grown at 37°C. Cells from liquid culture were diluted in fresh media and incubated with 10 nmol of overlapping oligonucleotide-Ab conjugates for 1 hour at room temperature. The resulting cell-antibody complex was washed free of excess, unbound conjugate by centrifugation and was resuspended in cold Tris-buffered saline, effectively diluting the overlapping antibody reagent to below 1.0 pmol. The washed cells were simultaneously streaked onto plates and assayed by incubation in PCR reagent mixture for 10 minutes at 33°C, followed by 40 cycles of real-time PCR in the presence of downstream primers. The presence of 10–50 cells, as determined by colony formation in overnight culture, was sufficient to elicit signal above background in a 2–3 hour assay (see Figure 6).

Streptavidin Model System

The first chain extension of overlapping DNA labels in the NADIA assay appears to be highly inefficient. Well under 1% of the input antigen molecules result in the formation of a template sequence. In theory, every antigen molecule should result in the formation of a sandwich immune complex and should produce an amplifiable sequence if the label strands can overlap and extend to form the new primer binding sites. A three-dimensional spatial analysis of a PSA sandwich complex reveals that such a structure may not be able to be spanned by every label.

Insufficient oligonucleotide length was investigated as a possible limitation by employing a smaller immune complex model consisting of streptavidin labeled with one each of overlapping 60-mer sequences biotinylated on the 3′ ends. Incomplete first-chain extension due to steric hindrance was also tested as a possible limitation by employing 3′–5′ overlap labels. A single first-chain extension occurs away from the center of mass to create a new primer-binding site and a novel template in this orientation.

Such a construct, when subjected to first-chain extension and subsequent PCR, yielded thresholds slightly less than those observed for the equivalent amount of actual double-stranded template. There was little difference between the 3′–3′ and 3′–5′ orientations. The efficiency of first-chain extension was improved to around 50%, allowing for the detection of a few hundred copies of complex in a homogeneous assay. These observations indicate that if DNA label chain length is sufficient to easily span a sandwich immune complex, overlap and first-chain extension will occur efficiently. The assay sensitivity of a homogeneous format then becomes limited predominantly by nonspecific (i.e., thermodynamic) overlap occurring in the bulk antibody reagent. Reduction of nonspecific signal generation by the use of low conjugate concentration (1–10 pmol) must be balanced against kinetics and equilibrium of sandwich complex formation in solution.

Efforts are under way to use oligonucleotide labels containing multiple spacer phosphoramidite 18 molecules from Glen Research Corp. on the 5′ ends to effectively increase chain length, and to employ Fab and F(ab′)2 fragments to diminish immune complex size.

Conclusion

Ed Jablonski is vice president, research and development, and Tom Adams is chief science officer at Iris International Inc. (Chatsworth, CA). The authors can be reached at ejablonski@leucadiatechnologies.com and tadams@leucadiatechnologies.com, respectively.

In the detection of protein antigen in any format, the sensitivity of IPCR can never match the absolute detection of a nucleic acid sequence on a molecule-per-molecule basis. This is because an immunoassay detects a binding reaction, which has specific and nonspecific pathways. The presence of a nonspecific pathway sets the limit of detection. The enhanced sensitivity of IPCR is manifested in its ability to detect nonspecific interactions that would otherwise be unobserved with traditional label molecules. The advantage of IPCR over the direct detection of a nucleic acid sequence for an infectious organism or disease state would be the overwhelming abundance of specific protein molecules over a specific nucleic acid sequence.

NADIA may represent a major advancement in diagnostic technology. The addition of a proximity-binding condition for either a solid- or solution-phase assay introduces another level of discrimination unavailable in other assay formats. The advantages include sensitivity, range of response, and high throughput in microtiter plate arrangement. NADIA has the potential to detect a few hundred copies of protein molecules or a few intact cells using relatively simple homogeneous protocols. The use of real-time PCR instrumentation has become routine, and assays can be formatted to work on systems such as those developed by Cepheid (Sunnyvale, CA) or Roche Diagnostics (Indianapolis). Other amplification systems such as transcription-mediated amplifica- tion (TMA), strand-displacement amplification (SDA), and nucleic acid sequence–based amplification (NASBA) isothermal assays could be used in handheld devices. The immune components are available. Both monoclonal sandwich pairs and affinity-purified polyclonal antibodies to many protein antigens, toxins, and cells have been developed.

Molecular diagnostics is an emerging segment of the IVD market that is growing rapidly and garnering attention due to its potential to revolutionize disease management. Personalized medicine relies on analyzing genetic factors that guide the identification of disease and the management of each individual patient's disease. Considering such potential, numerous platform technologies are being evaluated as vehicles for conducting molecular diagnostic tests.

One promising platform is the DNA microarray, which facilitates the multiplex analysis of multiple genetic factors simultaneously. Since many diseases and potential treatment outcomes are determined by the function of multiple genes, DNA microarrays will be an important molecular diagnostic technology. However, there are some major concerns with DNA microarrays for diagnostic applications. Such issues include cost, complexity, reproducible performance, and regulatory concerns. This article will discuss the use of DNA microarrays in diagnostic applications, specifically highlighting one platform that addresses many of the concerns and provides a rapid approach to developing and delivering microarray-based diagnostics.

Microarrays in Disease Management

Since the completion of the Human Genome Project, there has been a movement toward taking a molecular approach to the diagnosis and management of disease. However, clinical reference laboratories are grounded in a singleanalyte paradigm: the sequential analysis of single genes or markers, and slide-based morphological analysis. While clinically relevant, such approaches ignore the fact that most diseases are multivariate, and are the sum result of complex genetic, social, and environmental factors.

DNA microarrays have recently been used to obtain global perspectives of human disease and to identify genetic markers that are important for diagnosis and therapy.^1–7 Such applications for microarrays have focused on analyzing genetic alterations and gene expression patterns underlying the biological properties and clinical behavior of pathogenic diseases. This genomewide approach is required to understand the fundamental complexities of disease. For example, cancer displays amazing heterogeneity, including gene amplifications, deletions, and genetic instability, even among tumors considered to be within the same class. Microarray technologies have already contributed to the discovery of new biomarkers for diseases and have spurred the rapid growth of a multianalyte paradigm for disease diagnosis and treatment. The resulting gene signatures, which are groups of gene expression patterns associated with a disease or trait, offer the promise of improved disease staging, risk stratification, and treatment decisions.

By focusing on genetic determinants of drug responses at the human genome level, pharmacogenomics plays an important role in prescribing safer and more-effective individually tailored drugs. DNA microarrays have shown great promise in clinical medicine by paving the way toward such effective individualized drug regimens. Microarray technologies and proteomics are instrumental in predicting drug sensitivity and potential side effects by studying the cellular pathways through gene expression profiles. Such information is changing the understanding of how genetics influences disease development and drug response, and contributing to the discovery of new treatments. Prospective genotyping of patients for various genes to determine drug targets, drug metabolism, and disease pathways is the first step toward individualized therapy, by matching the patient's unique genetic makeup with an optimally effective drug.⁸

This article will review some of the key emerging microarray technologies, discuss the impact of these technologies on the IVD industry, and explore some key challenges, opportunities, and applications.

Microarrays: An Overview

DNA microarrays are a combination of technologies that are grouped by their ability to measure global changes in gene expression. The human genome consists of thousands of genes, each with a unique nucleotide pattern of G, A, T, and C molecules. Within any group of cells, a pattern of expression for these genes is observed, generating an mRNA copy with a nucleotide pattern that is specific to each gene. Microarrays are a hybridization-based platform that is composed of thousands of probes; such probes are positioned at defined locations, and are capable of binding to a specific gene-associated mRNA. Cellular gene expression is assessed by labeling an RNA sample, applying the labeled RNA to a DNA microarray, enabling hybridization between the probes and mRNA, and measuring the signal intensity at the probe positions. By measuring the amount of label present on the probes, a microarray can analyze the expression levels of thousands of genes in a single assay reaction.

a)	b)
Figure 1. The CustomArray 12K (a) and CustomArray Synthesizer (b) by CombiMatrix (Mukilteo, WA). The CustomArray 12K is an active semiconductor-based array on which up to 12,500 DNA probes can be synthesized in situ utilizing an electrochemical approach. The synthesis is performed on the CustomArray Synthesizer, a benchtop instrument that has a capacity of up to 8 customized microarrays per run.

A number of companies have been producing DNA microarray-based technologies, including Affymetrix (Santa Clara, CA), Agilent Technologies Inc. (Santa Clara, CA), Applied Biosystems (Foster City, CA), CombiMatrix (Mukilteo, WA), and GE Healthcare (Chalfont St. Giles, UK). These companies have microarray platforms that offer solutions to the problem of multianalyte analysis. For example, the CodeLink Bioarray by GE Healthcare is a slide-based array in which each synthesized oligonucleotide probe is embedded in a proprietary three-dimensional polyacrylamide aqueous-gel matrix. In contrast, the Affymetrix platform is an in situ synthesized approach that utilizes light-activated deprotection chemistry for in situ synthesis onto a silica-based surface. Similarly, Agilent's arrays are constructed in situ with synthesis directed by ink-jet printing, thereby precisely building oligonucleotides from a silicon surface. CombiMatrix's core technology is based on a modified electronic silicon chip. Using an electrochemical reaction, oligonucleotide probes are synthesized off the surface of a live active semiconductor. The flexibility of this platform enables rapid customization of probe design and synthesis of the DNA microarray (see Figure 1a).

During the past decade, DNA microarrays have been used in every facet of biological research, from basic science to the study of clinically relevant diseases. Microarrays have also been utilized in various other manners, including gene expression profiling, singlenucleotide polymorphism (SNP) analysis, and comparative genomic hybridization (CGH) analysis.

Microarray Applications

Gene Expression Profiling. A vast majority of the published clinical studies involving DNA microarray technologies have focused on identifying gene expression profiles of pathogenic and genetic diseases. Such studies have demonstrated the ability to identify gene expression patterns or classifiers that can distinguish between multiple common malignancies.^1–7 Microarrays offer the ability to diagnose tumors based on molecular rather than morphologic characteristics, which allows tumor classification based on pathways rather than on appearance. Microarray-based gene expression analysis of breast cancer and leukemia have revealed that similar tumor types have distinctly different molecular characteristics.^2–4 For example, leukemia includes more than 20 morphologic, genetic, and molecular subclasses. This characterization has led to the development of cancer databases, and the movement from a subjective phenotypic classification to a uniform molecular-based diagnosis of cancer.

SNP Analysis. SNPs are DNA sequence variations that occur when a single nucleotide (A, T, C, or G) in the genome sequence is altered. They are useful polymorphic markers that are often associated with susceptibility to diseases or are related to drug responsiveness. A small subset of SNPs may directly influence the quality or quantity of the gene product, resulting in an increased risk of certain diseases or severe drug side effects.⁹

CGH Analysis. Gross genetic aberrations are responsible for many diseases (e.g., Down, Prader-Willi, Angelman, and Cri du Chat syndromes), and for the development of neoplastic potential in a number of malignancies. CGH was developed for genomewide analysis of DNA sequence copy number in a single experiment. CGH analysis has been applied to study the chromosomal changes that occur in cancer cells, including the loss or duplication of regions of chromosomal DNA. Microarray-based CGH analysis has enabled researchers to explore rapidly such chromosomal changes on the same platform used to study gene expression changes. Combining such powerful tools on a single platform has greatly affected the discovery of oncogenes, tumor suppressor genes, drug targets, and biomarkers.¹⁰

Bringing Microarrays into the Lab

In clinical diagnostics, microarrays have been historically viewed as powerful research tools to identify markers for traditional IVD applications. Once markers were identified from a survey of tens of thousands of initial candidates, more-affordable, sensitive, and rapid assays have been created by using other platforms, such as quantitative polymerase chain reaction (PCR). However, with the recent advent of commercially available, low-cost, reproducible, technically simple arrays, and easy-to-use analytical software, the direct use of microarrays as diagnostic platforms in reference laboratories is beginning to be explored.

While the opportunities to improve patient care are clear, the path to clinical implementation of DNA microarrays is less defined. However, with increased awareness of the diagnostic potential of DNA microarrays, reference laboratories are being challenged to implement assays based on this technology. Implementing any new technology in a clinical setting creates a series of challenges, including staff licensure, technical training, and equipment purchases.

Staff licensing requirements for microarray applications are similar to most other molecular technologies that must comply with either national credentialing agency or state certification in the general or molecular categories. The technical component of a microarray assay is similar to quantitative PCR or hybrid capture assays, which are conducted daily in thousands of laboratories. In contrast, the professional interpretative aspects of microarray analysis, including quality control, data normalization, and results analysis, are unique to microarray technologies.

One area of concern that the licensing bodies currently overlook is the field of bioinformatics and microarray data analysis. While most diagnostic applications provide automated computer analysis algorithms, bioinformatics will clearly become an integral component of esoteric microarray assay analysis.¹¹

The essential equipment used for DNA microarray applications is not significantly more expensive than that for other molecular-based diagnostic assays. Besides adding a specialized imaging system, many microarray-based assays can be performed with the same basic equipment found in a molecular diagnostics lab.

Developing Clinical Microarrays

There are two routes to developing and commercializing a clinical diagnostic assay: submitting a test for FDA approval, or conducting an internal validation to create a laboratory-developed test which is commonly known as a home-brew test.

Through the first route, there are two options: either the premarket approval (PMA) process, or 510(k) premarket notification that is based on a comparison to a predicate device. While these options allow a test to be sold to licensed reference laboratories in the United States, the process can be slow, expensive, and time-consuming. This route is required for high-volume products which are intended for manufacture and sale to third-party laboratories.

The second route is routinely undertaken by clinical laboratories and involves in-house internal assay validation. IVD tests developed and validated internally are indeed considered medical devices, and are created utilizing general-purpose reagents and analyte specific reagents (ASRs). In 1996, FDA introduced regulations that outline how ASRs and general-purpose reagents should be used to develop home-brew assays.¹²

In this regulation, FDA defined ASRs as “antibodies, both polyclonal and monoclonal, specific receptor proteins, ligands, nucleic acid sequences, and similar reagents which, through specific binding or chemical reaction with substances in a specimen, are intended for use in a diagnostic application for identification and quantification of an individual chemical substance or ligand in biological specimens.” According to this regulation, all ASRs purchased from third-party vendors must be registered with FDA, be labeled as an ASR, and comply with certain quality control guidelines.

In essence, FDA recognizes a legitimate home-brew test as an assay comprised of ASRs and general-purpose reagents (e.g., buffers or reactive materials without specific intended uses). In addition, all laboratories conducting in-house testing are required to meet CLIA high-complexity certification requirements, establish the performance of the home-brew tests per CLIA regulations, and label such tests that are developed using ASRs with the following statement: “This test was developed and its performance characteristics determined by (laboratory name). It has not been cleared or approved by FDA.”¹³

As of January 2006, FDA has not authorized any microarrays to be sold as ASRs. Consequently, the only available route for clinical implementation of non-FDA-approved microarrays is through the second route, by producing the entire microarray in-house from general-purpose reagents. The CustomArray and Microarray Synthesizer by CombiMatrix is the only commercially available platform that offers in-house synthesis capability (see Figure 1b). The benchtop Microarray Synthesizer allows clinical research groups to rapidly develop clinical diagnostic assays in-house. By using either the 12k (with up to 12,500 DNA probes) or the 95k (with up to 95,000 DNA probes) microarrays, molecular gene signatures can be rapidly created and validated.

Figure 2. Tumor classification and drug response database. A bioinformatic database comprised of gene expression signatures of tumor biopsy samples is built. The expression signature of each biopsy sample is correlated with clinical history that includes information on the response of the patient to therapy. This database is comprised of signatures from multiple cancers and patients treated with multiple drugs. When a new patient is diagnosed with cancer, the corresponding biopsy or tumor is analyzed, and comparison of that expression signature with those in the database provides information to the physician that aids in the management of the patient.

Another feature of the CombiMatrix platform is the ability to segment arrays. Through segmentation, a 12k array can be subdivided into four separate arrays of 2000 features. While many research applications require genomewide analyses, many clinical diagnostics will only necessitate the measurement of a smaller gene cohort. This platform can measure the preferred 250–500 genes with replicates and controls.

Key Application

One application that is currently being developed with the CombiMatrix system is a tumor classification and drug response prognostic (see Figure 2). The central focus of this product is the development of a prognostic database. Multiple tumor samples are analyzed with the CombiMatrix CustomArray to generate gene expression signatures. For each of the gene expression signatures, a corresponding detailed clinical history exists, which includes all relevant information including type and grade of tumor, response of the patient and tumor to drug therapies, and outcome. Once the database has been built and validated with blinded samples, it can serve as a powerful diagnostic tool to aid physicians in managing cancer patients.

For example, a biopsy or tumor sample from a diagnosed patient is analyzed with the CombiMatrix CustomArray; a gene expression signature is provided and then compared with expression signatures in the database. Though each expression signature is expected to be unique, certain patterns and trends will be similar to the signatures in the database. An analysis of the common patterns will provide information to physicians indicating that the patient exhibits an expression signature that is characteristic of a particular tumor type and grade, and a response probability to various drug treatment options. This product and other similar products will change the way patients are managed in a clinical setting.

Conclusion

The introduction of microarray-based applications in clinical laboratories has been a slow process, due to myriad problems involving technical, regulatory, and financial reasons. Nonetheless, it has become apparent that microarray platforms may revolutionize clinical diagnostics as they enable laboratory professionals to move away from single-analyte analysis and focus on complex multianalyte applications. The clear winner in this process will be the patients, as the healthcare industry learns to manage the complex processes of human diseases with 21st-century tools.

(Left to right) Amit Kumar, PhD, is president and chief executive officer at CombiMatrix Corp. (Mukilteo, WA) and chairman at CombiMatrix Molecular Diagnostics. Michael Opel, PhD, is a senior scientist at CombiMatrix Molecular Diagnostics. They can be reached at akumar@combimatrix.com and mopel@cmdiagnostics.com, respectively. Matthew Moore, PhD, and David Baunoch, PhD, (not pictured) were formerly at CombiMatrix Molecular Diagnostics.

The nucleic acid test (NAT) market has grown significantly in size and diversity in recent years. Generally, nucleic acid tests fall into three categories: those based on direct hybridization and detection using specific nucleic acid probes; those based on signal amplification, where specific probes bind the target sequence and the resulting signal is amplified; and those involving target amplification, typically achieved through enzymatic means. The third category is characterized chiefly by amplification methods that offer enhanced sensitivity, such as the polymerase chain reaction (PCR).

Since its invention in the 1980s, versatile PCR has found a multitude of commercial applications in the life sciences and in vitro diagnostics. More than 20 years later, PCR remains at the center of NAT technology. It is commonplace in the clinical laboratory and essential as a research tool. PCR provides the benchmark for nucleic acid amplification, but many alternative methods have been devised and published that use novel and innovative approaches to achieving amplification and detection of target sequences. Such alternatives include nucleic acid sequence–based amplification, rolling circle amplification, Qß replicase, and simultaneous strand displacement amplification. The majority of these other approaches have application in infectious-disease diagnostics. Although some have proven clinical utility, many may have been developed in an attempt to circumvent patent issues.

More recently, research has concentrated on novel detection techniques and equipment designs that make PCR and other amplification technologies more amenable to current molecular diagnostic requirements. High-throughput automation of sample extraction, amplification, and detection has been the principal focus, primarily driven by blood bank screening for infectious diseases such as HIV and, lately, by genomic screening and single-nucleotide polymorphism (SNP) analysis. Point-of-care (POC) nucleic acid tests have been an elusive goal until recently, largely owing to the complexity of molecular assays and the technical challenges they present with regard to sample preparation and control of assay reproducibility and reliability. Nevertheless, fully integrated POC nucleic acid platforms are emerging. Several multinational companies in the diagnostics and information technology industries have invested significant resources in the development of complex bioengineering strategies centered on microfluidic and bioelectronic sensor technologies. These devices are expensive and currently used only in high-profile applications such as biodefense.

Figure 1. Diagram of a lateral-flow immunoassay rapid test strip.

An alternative, and more cost-effective, approach to POC nucleic acid testing is to use a lateral-flow platform. Lateral-flow immunoassays, which are exemplified by pregnancy test devices, represent a significant portion of today's immunochemical POC market. Their speed, low cost, and simplicity of use make lateral-flow immunoassays the only true point-of-care test for now. These chromatographic devices employ nanoparticles coated with materials that bind to an analyte, such as an antibody or antigen, within a sample. This analyte-nanoparticle complex flows laterally through a series of overlapping membranes until it is captured on an antibody or antigen capture line. A visual result is achieved within a few minutes. Even an unskilled operator can rapidly interpret the test result without any need for complex or expensive equipment (see Figure 1).

Nucleic acid lateral flow (NALF) uses nucleic acid hybridization to capture and detect nucleic acid amplification products in a manner akin to lateral-flow immunoassays. Strips used in the technique do require modified test and capture lines and conjugate pad components. This approach combines the advantages of lateral-flow platforms with those of traditional nucleic acid tests, as described in this article.

Detection with NALF

Figure 2. NALF capture strategies: antibody stripe with labeled amplicon (a), and oligonucleotide probe (b); streptavidin stripe with biotinylated amplicon (c), and biotinylated oligonucleotide probe (d); passive adsorption of BSA-oligonucleotide probe (e), and an unlabeled oligonucleotide probe (f).

Nucleic acids can be captured on lateral-flow test strips in an antibody-dependent or antibody-independent manner (see Figure 2). Antibody-dependent capture, as schematized in the first two drawings in the figure, involves an antibody capture line and a labeled amplicon or oligonucleotide probe of complementary sequence to the amplicon—for example, goat anti-dinitrophenol (DNP) and DNP-label. Antibody-independent alternatives offer more potential for multiplexing, minimize the likelihood of batch-to-batch variation, and may be lower cost. One such method uses noncovalent interaction between two binding partners, exploiting, for example, the high affinity and irreversible linkage between a biotinylated probe or amplicon and a streptavidin line.

A much favored and more simple approach is to immobilize oligonucleotide capture probes directly onto the nitrocellulose lateral-flow membrane. This can be achieved by passive adsorption of a bovine serum albumin–labeled oligonucleotide probe or, preferably, an unlabeled oligonucleotide probe. All of these methods employ standard lateral-flow immunoassay striping equipment and yield strips with long-term stability, often at room temperature.

Figure 3. NALF detection chemistries. In all cases, the oligonucleotide probe may be replaced with an antibody and an appropriately labeled oligonucleotide probe or amplification primer. With an enzyme, an enzyme-probe complex converts substrate into a colorimetric product. With gold, a nanoparticle-probe conjugate provides a visual signal. With Qdot, a Qdot-probe conjugate emits fluorescence. With UPT, a UPT-reporter conjugate is excited at 980 nm, and visible green light is emitted at 500 nm. With a lipsome vesicle, a vesicle-probe conjugate releases dye.

Enzymatic Detection Signals. In theory, any detection chemistry used in lateral-flow immunoassays is applicable also to NALF. The only limitation would be its ability to conjugate the signal molecule of interest to an appropriate antibody or oligonucleotide detection probe. A broad range of NALF detection signals with quantitative and multiplexing capabilities have been reported (see Figure 3).

Reverse-hybridization enzymatic strip assays are the forerunners of today's NALF tests. In these tests, enzyme-labeled probes are hybridized to complementary nucleic acid target species on the surface of a nitrocellulose or nylon membrane. The result is a hapten-antibody-enzyme complex such as, for example, biotin-streptavidin–alkaline phosphatase. Complex wash and substrate-incubation steps are necessary to develop a readable colorimetric signal. Therefore, conversion to conventional lateral flow that will be suitable for nonlaboratory POC NAT applications is a relatively complex matter. Lateral-flow immunoassay–based Fluidics-on-Flex technology (Epocal Inc.; Ottawa, ON, Canada) provides an example. This system incorporates an integral fluidic circuit and pump manifold.

Nanoparticle Detection Signals. Three so-called bead technologies have been used successfully in NALF, namely, colloidal gold, latex, and paramagnetic nanoparticles. Both gold and latex give rise to colorimetric signals visible to the naked eye or semiquantifiable via inexpensive readers. Latex can be manufactured in any color, whereas 2- to 250-nm gold nanoparticles have a characteristic red color that results from surface plasmon resonance.

Gold is the lateral-flow nanoparticle of choice, mainly because of its small size, sensitivity, and robust manufacturing methods.¹ It can be conjugated to antibodies and oligonucleotides and labeled with small binding moieties such as biotin or DNP. Gold NALF platforms generally use 30- to 80-nm nanoparticles conjugated to an antibiotin antibody. This gold conjugate is then complexed with a biotinylated amplicon or sequence-specific oligonucleotide detection probe that, when captured by means of an oligonucleotide capture probe or an antibody-hapten-based capture method, yields a signal in the form of line. The optimal size and concentration of the gold nanoparticles used depends on assay specifications, including the application, line intensity, color (cherry red or purple), linear response to target concentration, and uniformity of multiplexed signals.

Paramagnetic NALF is similar to gold NALF but uses 100- to 200-nm superparamagnetic nanoparticles.² These nanoparticles emit a nonvisual signal when they are subjected to a magnetic field; interpretation requires a specialized reader. Gold and superparamagnetic NALF can detect as little as 1 fmol of synthetic target, sensitivity an order of magnitude greater than that of labor-intensive gel electrophoresis. And superparamagnetic NALF promises further improvement. As with all nanoparticle detection technologies, quality reagents are a key prerequisite. Nanoparticles should be uniform in shape and size and remain free of aggregate.

A number of methods for improving the sensitivity of nanoparticle NALF have been investigated. Detection probes labeled with multiple hapten moieties have been used to form large signal-enhancing lattices by binding specifically to multiple gold nanoparticle conjugates. Combined with real-time PCR, this method can achieve a visual sensitivity similar to that of fluorogenic instrument–based probe methods.³ Also, nanoparticle NALF may be combined with DNA dendrimer signal enhancement, resulting in DNA dendrimers that are branched nucleic acid species with 2 to 900 identical labels per dendrimer. These dendrimers can improve biological assay sensitivities up to 200-fold, depending on the dendrimer size, the application, and the nature of the assay.⁴

Methods are now available that allow gold and superparamagnetic particles to be coupled to oligonucleotide primers or probes.⁵ This approach may help to minimize steric hindrance and maximize gold NALF assay sensitivity. Combining such methods with oligonucleotide capture probe immobilization may also eliminate the need for antibody. This antibody-free NALF format can reduce the number of assay components and, in some cases, device cost. With further optimization, this system may improve NALF sensitivity, specificity, and reproducibility. It is also possible to prepare oligonucleotide gold and paramagnetic conjugates that are stable at elevated temperatures. Such conjugates can tolerate thermal PCR cycling conditions, allowing them to be included in amplification reactions.

Emerging Detection Chemistries. More-pioneering detection approaches that draw on liposome and fluorescence methodology are in early stages of development. Liposome nanovesicles constitute a lipid bilayer that can be covalently linked to antibodies and oligonucleotides. These transparent spheres can be used to encapsulate aqueous signals such as dyes in a controlled manner. When employed in NALF testing, oligonucleotide-tagged liposome nanovesicles release dye to yield a visual capture line. A prototype device has been able to detect as few as five viable Cryptosporidium oocysts.⁶

The use of fluorescence-based lateral-flow immunoassay reporters is on the increase, and several of them have been demonstrated in NALF applications. For example, a dual fluorescein- and biotin-labeled oligo probe has been used to detect single-stranded amplicon generated by cycling probe technology.⁷ However, the utility of such standard fluorophores is limited by high background fluorescence, the need for a complex reader, and the number of spectrally diverse fluors available for multiplexing.

UPT-NALF is an alternative approach that uses up-converting phosphor reporters, which are approximately 400-nm particles composed of rare earth lanthanide elements that are embedded in a crystal. These particles emit visible light after excitation with infrared radiation in a process called up-conversion. Up-converting occurs only in the phosphor lattice, so autofluorescence of other assay components is virtually nonexistent. UPT has been used to develop rapid prescreening and hybridization-based confirmatory tests for the detection of human papillomavirus type 16, a marker for cervical cancers.⁸

A third fluorescence approach uses quantum dots, known also as Qdots.⁹ These nanometer-sized semiconductor nanocrystals have extraordinary optical fluorescence properties that enable them to be as much as a thousand times brighter than conventional dyes. Nanocrystal size determines their color. Their emission profile is narrow and symmetrical, resulting in minimum crosstalk. Qdots are visualized under ultraviolet (UV) light and can be tuned to allow excitation by means of the same long-wavelength UV lamp. The development of water-soluble Qdots that can be conjugated to antibodies (Qdot bioconjugates) has made Qdots amenable to lateral-flow immunoassay applications.¹⁰ Initial feasibility testing has employed dot-infused hcG pregnancy tests and is likely to be extended to spectrally multiplexed assays and next-generation NALF applications in the near future.

Detection limits equivalent to or better than those of current gold-standard nucleic acid tests and lateral-flow immunoassays are essential for many POC nucleic acid applications. All of these emerging technologies claim to improve on nanoparticle NALF and enzymatic detection signal sensitivity by two to three orders of magnitude, but they are currently limited by the need for suitable readers or more early-stage development.

Figure 4. Multiplexed NALF, represented in (a) by a schematic showing conversion of an influenza typing panel to that platform, and in (b) by sevenplex detection of synthetic amplicon representing 14 different influenza genotypes: (1) InfA N1 H1; (2) InfA N1 H3; (3) InfA N1 H5; (4) InfA N1 H undefined; (5) InfA N2 H1; (6) InfA N2 H3; (7) InfA N2 H5; (8) InfA N2 H undefined; (9) InfAB negative, H undefined; (10) InfA H1 N undefined; (11) InfA H3 N undefined; (12) InfA H5 N undefined; (13) InfA HN undefined; (14) InfB HN undefined.

NALF Applications

Chief areas of imminent POC application of this new NALF technology, and worthy of discussion here, are low-density multiplexed detection and rapid detection of SNP genetic indicators.

Low-Density Multiplexed NALF. Automated high-throughput DNA chip and array technology is well suited for multiplexed detection of very high numbers of samples or probes. However, a need remains for low-density multiplexing platforms for POC detection in the region of 2–25 targets. Low-density multiplexed NALF should fill this niche. This method is designed to detect multiplexed PCR amplicons, as in the Templex assays developed by Genaco Biomedical Products Inc. (Huntsville, AL). A simple multiplexed housed device comprises a single lateral-flow strip, sample port, and conjugate pad, and multiple oligonucleotide capture probe stripes.

A gold-conjugate-based prototype device incorporates seven different target and complementary probes. It has been used to detect synthetic target mixes representing 14 different influenza genotypes in 20 minutes at room temperature (see Figure 4). Preliminary data demonstrate detection of as little as 1 fmol of synthetic nucleic acid sequence, and one-tenth of a standard Templex reaction. A bidirectional housing that incorporates two longer test strips, or a multidirectional housing such as tri- and quad-NALF strips in a tee or cross configuration, can theoretically multiplex much higher numbers of samples. The extent of multiplexing actually is limited by the availability of nonstandard sizes of lateral-flow membranes, the rate and volume of flow through these, and the amount of gold conjugate required. However, preliminary evidence suggests that 24-plexing is feasible with the use of a quad strip format.

Figure 5. SNP NALF detection strategies. (a) In competitive allele-specific short oligonucleotide hybridization, biotinylated and unlabeled mutant and wild-type sequence-specific hybridization probes compete at the SNP site, resulting in sequence-dependent presence or absence of signal at a streptavidin capture line. (b) Alternatively, chimeric PCR primers that incorporate hexapet tags are used to synthesize PCR amplicons in an allele-specific manner, with the resulting amplicons being captured at room temperature on complementary striped hexapet oligonucleotide probes.

The major hurdle to be overcome in multiplexed NALF is the nonspecific signal inherent in NAT platforms comprising large numbers of probes and amplification primers. Minimization of this can be achieved through careful optimization of primer and probe sequences, concentrations, and positioning of capture lines.

SNP NALF. Single-nucleotide polymorphisms are important indicators of human genetic disease, strain genotypes, and drug resistance. A number of rapid SNP NALF detection techniques amenable to POC are in development.

Most PCR-based SNP diagnostics use a single primer pair to make amplicons with variable internal regions containing one or more single-base mismatches. One such approach involving NALF technology is competitive allele-specific short oligonucleotide hybridization (CASSOH).¹¹ This method discriminates at the capture level by means of biotinylated sequence-specific hybridization probes designed to contain the SNP base of interest. Competition between labeled and unlabeled mutant and wild-type probes results in the presence or absence of signal at a streptavidin capture line in a target-sequence-dependent manner (see Figure 5).

A disadvantage of the CASSOH system is the requirement for separate reactions for each target sequence and a postamplification temperature ramping. However, an alternative approach has been devised that uses allele-specific PCR and chimeric PCR primers that incorporate hexameric repeat tags termed hexapet tags.¹² These tags, designed to exhibit minimal cross-reactivity, have been used to demonstrate specific hybridization-based capture of amplicon at room temperature using NALF strips striped with complementary hexapet tag sequences (see Figure 5b). This system discriminates between alleles at the amplification level, which brings its own disadvantage: multiple allele-specific primers are required.

Figure 6. Duplex detection of SNPs using hybrid nucleic acid probes in a platform developed by BBInternational (Cardiff, UK). (a) In the schematic representation, SNP-specific oligonucleotide detection probes (stripes) capture single-stranded PCR amplicon. The antibiotin-gold conjugate (red circle) and biotinylated detection probe (green circle) are immobilized in a dry-conjugate-pad format. (b) The test images reveal housed full-dipstick discrimination of factor II wild-type and mutant genotype sequences (C = wild type, G = mutant).

BBInternational (Cardiff, UK) is developing a noncompetitive PCR NALF platform that discriminates at the detection level. Short immobilized hybrid nucleic acid probes are designed to have carefully optimized melting temperatures. This system has demonstrated discrimination at room temperature without the need for allele-specific primers, and detects asymmetric PCR amplicon or amplicon that has been rendered single-stranded by lambda exonuclease enzyme digestion (see Figure 6).

All of these SNP-based NALF platforms potentially can be used to multiplex detection of two or more SNPs in a single device. In all cases, the key to success is careful primer and probe design.

NALF Formats

Methods are available that allow for the manufacture of generic NALF strips.¹³ Such strips include a bridge oligo with a first region complementary to a generic striped oligonucleotide probe, along with a second sequence-specific region complementary to the amplicon of interest (see Figure 7). Detection of different target sequences requires redesign of the bridge oligo, but the same generic NALF strip can be employed, which theoretically minimizes strip redesign.

An extension of this generic approach is the development of combination lateral-flow immunoassay strips that utilize nucleic acid base pairing to facilitate detection of immunochemical targets (Figure 7b). Detection involves a generic striped oligonucleotide probe capturing a complementary oligonucleotide–human antibody conjugate. The analyte of interest is sandwiched between the capture conjugate and a cross-reacting antigen-label conjugate.

NALF Assay Design

Figure 7. Generic NALF strips take two basic forms. (a) In a bridging-probe NALF, detection is facilitated by a bridging oligonucleotide probe with a first portion complementary to a labeled PCR product and a second portion complementary to a striped capture oligonucleotide. (b) The immunochemical-NALF combinatorial sandwich assay approach involves bringing the patient antibody (red circle) between an antigen-labeled conjugate detection reagent and an antihuman antibody–oligonucleotide conjugate.

Most thermal and isothermal NAT amplification technologies can be readily converted to NALF by adaptation of probes and labeling moieties. In many cases, amplification and detection probe sequence redesign is unnecessary; thus, development time and cost are minimized, and so may be the time required to secure regulatory approval. The incorporation of amplification and detection controls is also relatively straightforward, just a matter of including one or more additional capture lines and complementary probes and/or primers.

Designers of NALF assays must take into account general lateral-flow immunoassay design criteria that influence assay sensitivity and nonspecific signal.¹⁴ Nucleic acid–compatible sample and conjugate pads and blocking materials must be chosen. For example, nucleic acids will attach to glass-fiber materials, causing sample retention and inhibiting probe release. However, these effects can be minimized through the application of novel pad materials and blocking strategies.

Nitrocellulose membrane pore size and flow rate also are key, as there is a trade-off between assay time and efficient target-probe hybridization. NALF detection is not particularly prone to sample effects, because the sample often is diluted in upstream processes such as amplification. Gold NALF assays can detect amplification reactions involving blood products and crude bacterial cell lysates without significant signal inhibition. Thus, unlike with lateral-flow immunoassays, complex sample-pad materials for upstream processes such as blood separation are not always necessary.

Manufacturing Implications and Limitations

Future NALF devices are likely to include dry-format amplification and detection reagent technology and versatile lateral-flow housing designs. Antibody- and oligonucleotide-nanoparticle conjugates can already be supplied in dry, soaked conjugate pad, and sprayed reagent line formats. Technology also is available to combine PCR and nanoparticle NALF components into a dry pellet that can be reconstituted upon sample addition.

Standard chromatographic lateral-flow materials are used for NALF strip manufacture, and strong similarity between NALF probe and antibody immobilization methodologies allows standard automated lateral-flow manufacturing equipment to be employed. Equipment costs, therefore, are minimized. Also, there is no need to retrain production staff. Unmodified oligonucleotides are significantly cheaper than striped antibodies, and oligonucleotide-striped membranes have long-term stability at room temperature. Successful NALF detection chemistries will be those operated and stored at room temperature.

Figure 8. A diverse range of lateral-flow housing designs.

Today's lateral-flow immunoassay designs can be extremely elaborate, incorporating novel functions such as separate sample and buffer ports; buffer bags and blister packs that entail piercing or wick insertion; multidirectional flow; multiple analyte detection; and moving parts (see Figure 8). Standard injection-molded housings often hide complex mechanisms such as wash steps behind simple push-button fascias. Similar approaches are being used to develop novel closed-tube NALF housing designs that will meet future POC needs in the clinical, environmental, and food diagnostic sectors. BBInternational, for example, is developing a housing that allows cross-contamination-free transfer of a PCR reaction to a NALF test strip.

NALF assay detection is rapid. Detection has been achieved in as little as 5 minutes, the speed depending on the application and the detection scheme. A major future challenge for manufacturers will be to incorporate rapid, simple, and relatively inexpensive upstream sample extraction and amplification processes that will reduce the overall time to result. Future NALF devices are likely to integrate emerging technologies that shrink these process bottlenecks to a matter of minutes. The RapidCycler 2 instrument from Idaho Technology Inc. (Salt Lake City) already achieves amplification in 15 minutes, but it comes with a high price tag. Microfluidic PCR devices of the future may prove to be simpler, less expensive, and more amenable to NALF. Prototypes are emerging that have no moving parts or complex in situ components such as pumps, valves, or wells, and that perform PCR in less than 5 minutes with the same yield as a typical 90-minute thermal cycler protocol.

Similarly, rapid and expensive benchtop extraction equipment is reaching the market. The PlasmaGen APR-510-S from Atmospheric Glow Technologies Inc. (Knoxville, TN), for example, provides one-step extraction of DNA or RNA from a variety of dry sample matrices in 1 to 2 minutes, and the QuickGene-810 of Fuji Photo Film Company, Ltd. (Tokyo), isolates DNA and RNA from whole blood in 6 minutes. Another approach uses rapid cycles of hydrostatic pressure to extract biological material. All of these methods yield high-quality nucleic acid from low-abundance samples and are potentially NALF-compatible. However, a need for simple, low-cost alternatives remains.

A variety of manufacturers offer magnetic-bead systems that do not require centrifugation, but these systems generally do require open-tube transfer from the lysis tube to the amplification vessel. Some progress has been made with the development of chaotropic archiving materials. These so-called papers contain chemicals that lyse cells, denature proteins, and protect nucleic acids from nucleases, oxidation, and UV damage. Unfortunately, unfavorable wash steps are necessary. Tubes coated with a solid-phase matrix that irreversibly binds nucleic acid, allowing extraction and PCR in the same reaction vessel, are another step in the right direction.¹⁵ Future solutions may lie in improved amplification technologies that can tolerate cruder sample extracts.

Some elements of the NALF technologies described above draw on skills known in the art, while others are patent-protected or based upon carefully guarded know-how. When choosing a NALF developer, it is important to consider the impact that early immunochromatographic lateral-flow patents may have on the business arrangement. BBInternational has negotiated an agreement with Inverness Medical Innovations that offers protection for contract development and manufacturing customers on a selection of lateral-flow patents within the Inverness portfolio.

Conclusion

Recent technological advances in nanoparticle and alternative detection chemistries, along with the development of flexible platforms that allow multiplexing and SNP detection, make NALF a serious contender for a place in the future nucleic acid POC diagnostic test market. This rapid, simple, and inexpensive detection methodology has potential for application in rapid infectious-disease strain typing, human genetic disease diagnosis, and environmental field-site testing. Development of sensitive and accurate NALF-type POC tests will most likely be achieved through interdisciplinary partnering between molecular biology specialists, experienced immunochemical lateral-flow assay developers and manufacturers, and experts in sample extraction, amplification processes, and equipment design. This approach will allow full exploitation of this flexible technology and bring the industry a step closer to true POC nucleic acid tests.

(Left to right) Joanna Seal, PhD, is senior project leader of the nucleic acid division, Helen Braven, PhD, is a project leader in research and development, and Paul Wallace, PhD, is technical director at British Biocell International (Cardiff, UK). The authors can be reached at joseal@britishbiocell.co.uk, helenbraven@britishbiocell.co.uk, and paulwallace@britishbiocell.co.uk, respectively.

Three U.S. companies have been vying for control of Vision Systems Ltd. (VSL; Melbourne, Australia), a contract manufacturer for IVD instruments in the cancer detection market on an OEM basis. Despite a flurry of activity during the past couple of months, many issues still remain unresolved.

The commotion began in August when VSL and Ventana Medical Systems Inc. (Phoenix) entered into a merger implementation agreement. Ventana was prepared to pay approximately $346 million for VSL, hoping to combine the instrument technologies of both companies, expand reagent development and manufacturing capabilities, and expand sales and support infrastructure to accelerate growth and product development.

In September, Cytyc Corp. (Marlborough, MA) jumped into the fray, announcing that it would initiate a tender offer to acquire VSL for approximately $374 million. Cytyc was eager to add VSL's Novocastra-branded antibodies to its portfolio, and incorporate VSL's product development unit in order to strengthen its position in the global diagnostics market.

“Vision presents an ideal opportunity to leverage our sales, service, and laboratory support infrastructure,” said Patrick J. Sullivan, Cytyc's chairman, president, and chief executive officer, in the company's announcement.

Despite expectations of a bidding war, Ventana announced in mid-September that it would not raise its bid for the company, and would instead file a patent suit against VSL. In its complaint, Ventana alleged that Vision's Bond X and maX OCR instruments infringe its patent for scheduling algorithms, which enable optimal sequencing of multiple tests. The suit did not affect Cytyc's plans, which reaffirmed its intention to buy VSL.

The issue was stirred up further on September 26 when rumors began to circulate that Danaher Corp. (Washington, DC) was also interested in acquiring VSL. The Wall Street Journal reported that Danaher might pay as much as $400 million for VSL. Danaher officials acknowledged that the company had engaged in preliminary discussions with and had been conducting limited due diligence on VSL regarding a competing bid. Proving it was not out of the game yet, Ventana announced two days later that it had acquired 12% of VSL's outstanding shares.

Not to be outdone, on September 29, Cytyc increased its offer to $517 million, offering shareholders a 53% premium over the Ventana agreement price. “We decided to increase our offer to underscore our commitment to completing this transaction as soon as possible,” said Sullivan. He also announced that the offer had already secured prebid acceptances from shareholders for approximately 29.6 million shares of VSL.

Danaher and Ventana countered with an announcement on October 1 that they were engaged in discussions “regarding a potential cooperative effort to acquire VSL.” However, the companies also noted that “there can be no assurance that Danaher and Ventana will reach an agreement with respect to a cooperative effort to acquire VSL, or that Danaher will make an offer of any nature for VSL.”

On the same day, VSL sent Ventana a notice that unless it issued a more favorable counteroffer by October 3, it would terminate the merger agreement. Since no counteroffer was made, VSL terminated its agreement with Ventana. Meanwhile, Ventana announced that it was still continuing to explore its options regarding VSL, including possible cooperation with Danaher.

Meanwhile, Cytyc filed an application with the Australian Takeovers Panel to set aside Ventana's acquisition of VSL shares. Cytyc also wrote to the Australian Competition & Consumer Commission with concerns about the negative competition consequences it says an acquisition of VSL by Ventana would have.

“The combination of Vision Systems and Ventana would be a combination of the number-one and number-two players in the market,” said Sullivan. “It will result in a combined market share of around 85%, blocking healthy competition.” Sullivan also urged the VSL board to endorse the Cytyc offer, but that has not yet happened.
 

Steven Gutman

FDA recently released two draft guidances: one for commercially distributed analyte specific reagents (ASRs) and another for IVD multivariate index assays (IVDMIAs).

In 1997, FDA issued a rule that defined and classified ASRs, established labeling requirements, and imposed restrictions on ASR sale, distribution, and use. According to this rule, ASRs must follow the same marketing requirements as other IVD devices. In response to IVD manufacturers' concerns, FDA's new draft guidance provides further clarification by expanding upon ASR marketing practices, research, and investigational use.

“What is ironic about all of this is that there is an underlying disconnect in the IVD industry,” says Tom Tsakeris, president at Devices and Diagnostics Consulting Group Inc. (Rockville, MD). “There is historical unevenness between what is required of a traditional IVD company and what is required of laboratories.”

Leif Olsen, a regulatory affairs specialist at Hogan & Hartson LLP (Washington, DC), agrees. “FDA created this situation in a way, since it has been uneven in how it enforces the regulation,” he says. “But it is an imperfect world, and FDA cannot think of everything ahead of time.” He goes on to say that while he does not blame FDA, he feels this guidance is about two years overdue.

Confusion has existed regarding IVDMIA regulation when developed and used by laboratories. This is partially due to FDA's regulation of laboratory-developed tests that use ASRs and other commercially available FDA-regulated components. The draft guidance addresses the definition and regulatory status of IVDMIAs, as well as premarket pathways and postmarket requirements for the tests, since FDA does not believe that IVDMIAs fall within the reach of laboratory-developed tests.

According to Glen Freiberg, president of RQC Consulting (San Diego), this is an important distinction: “If a laboratory manufactures all or part of an assay without using a manufacturer-supplied ASR, disclosure is not required.” He continues, “In these cases, physicians will have no idea that the assay does not have FDA clearance. OIVD's new draft guidance will not affect this without a new regulation or clear enforcement directive.”

However, change can be costly for laboratories. “There is such a long history here of FDA oversight of laboratories,” Tsakeris says. “Laboratory services have spawned up without previous regulation. With the new guidance, companies will find themselves suddenly defined by FDA as having developed a new device.”

Steven Gutman, MD, director at the Office of In Vitro Diagnostic Device Evaluation and Safety (OIVD), says that IVD companies should be motivated to demonstrate solid science regardless. “Companies need good core science to get FDA clearance, but they also need it to convince healthcare practitioners to order their tests as well as to ensure reimbursement,” he says.

Gutman also suggests that these documents should “not be overread.” He says the ASR guidance “simply clarifies the parameters for use of ASRs,” and the IVDMIA guidance applies to a “narrow niche of home-brew tests that FDA has decided not to apply enforcement discretion to, but to regulate as medical devices.” In addition, he emphasizes that this is not a fundamental change in policy with regard to home-brew assays. As far as ensuring equal enforcement, OIVD will rely on trade complaints within the industry. “Companies will turn each other in,” Gutman says.

However, industry analysts feel that this approach is an example of poor governance by FDA. “FDA has spent a lot of resources, time, and effort to develop these guidelines,” says Freiberg. “Publishing them and then basing enforcement on competition complaints does not serve the public health in an efficient manner.”

Likewise, other industry experts like Olson stress that further regulation can make costs escalate or prevent certain tests from hitting the market. “While it is a good idea, it is hard to predict the consequences in advance,” he emphasizes.

Additional information about these guidances can be accessed via OIVD's Web site at www.fda.gov/cdrh/oivd/index.html.

Gordon H. Smith

If Senator Gordon H. Smith (D–OR) could have it his way, selling direct-to-consumer (DTC) genetic tests in the United States will become more difficult in the near future. Responding to growing concerns about the lack of oversight of such DTC tests and the unproven science behind them, Senator Smith, who is chairman of the Senate Special Committee on Aging, called for a yearlong investigation by the Government Accountability Office (GAO) into companies that are selling DTC genetic tests.

At a recent hearing held by the Special Committee on Aging, GAO presented the findings of its investigation. During the past year, GAO purchased tests from four Web sites and created fictitious consumers by submitting DNA samples for analysis. According to GAO's report, “The results from all the tests purchased mislead consumers by making predictions that are medically unproven and so ambiguous that they do not provide meaningful information to consumers.”

“I am deeply disturbed by GAO's finding that consumers are being misled and exploited by this modern-day snake oil,” said Senator Smith in his opening statement at the hearing. “And I am shocked to learn how little the federal government is doing to help consumers make informed decisions about the legitimacy of these tests.”

As a result of GAO's findings, Senator Smith and the committee have been exploring ways to improve the regulation of DTC genetic tests and to warn consumers about such tests.

“The senator is looking at the current gaps in the regulations of genetic testing and the threat that it has to public health,” said Kimberly Collins, a spokesperson for the Special Committee on Aging. “Since the hearing, Senator Smith has continued to work with FDA and CMS to look into the safety and efficacy of these laboratory tests.”

In addition, on the day of the hearing, Senator Smith announced that the Federal Trade Commission, in conjunction with FDA and CMS, released an alert cautioning consumers about DTC genetic tests.

Tightening the regulation of DTC genetic tests would be a welcome change and a vast improvement over the current situation. As GAO's report pointed out, since there is no genetic testing specialty under CLIA, there are no specific requirements or unique standards for laboratories that perform genetic tests. Such minimal oversight makes it difficult for consumers to determine whether a genetic test provides meaningful, scientifically based information.

“There's still more that needs to be done to ensure that there's adequate regulation by FDA of home-brew laboratory tests,” said Collins. “The senator is still considering different options to ensure that there is such adequate regulation. The senator is in communication with FDA and CMS to see what needs to be done to ensure the safety of consumers. He's also looking at whether a legislative option is needed or if there needs to be a genetic test specialty under CLIA.”

Industry analysts agreed with the senator's assertions that the regulations for DTC genetic tests need to be changed, which is long overdue.

“If a DTC genetic test has not been cleared or approved by FDA, the test should not be made available without a physician's order and physician interpretation,” said Glen P. Freiberg, president of RQC Consulting (San Diego). “At the same time, a laboratory-developed test that is ordered by a physician may still not have test-to-test, lot-to-lot, or lab-to-lab reproducibility. The CLIA rules don't do enough to ensure this. Consequently, an update to the regulations is required that treats all tests in the same way, requiring manufacturing reproducibility, test reproducibility between labs via proficiency, and a determination of clinical utility by a physician or lab, not the government.”

A copy of the report can be accessed via the GAO's Web site at www.gao.gov/new.items/d06977t.pdf. Additional information can be accessed via the Senate Special Committee on Aging's Web site at http://aging.senate.gov/public/.

The personalized medicine movement recently got a major proponent on Capitol Hill. Senator Barack Obama (D–IL) introduced the Genomics and Personalized Medicine Act of 2006 (S. 3822). The bill proposes to increase funding for research on genomics, expand the genomics workforce, provide a tax credit for the development of diagnostic tests that can improve the safety or effectiveness of drugs, and reaffirm the need to protect genetic privacy.

“What's already happening in the area of genomics itself and what will likely happen in the next 5–10 years touches on a number of Senator Obama's health priority issues,” says Dora Hughes, MD, the senator's health and education policy advisor. “So it was a natural for him to pursue this issue further.”

According to Obama aides, the bill would create incentives to accelerate private-sector innovation in genomics. The legislation seeks to spur such innovation by allocating $150 million for research in genomics. It also provides a 100% tax credit for private research to develop companion diagnostic tests that can improve the effectiveness or safety of certain drugs. In addition, this bill would modernize FDA's outdated process for reviewing genomic tests.

“The companion diagnostic test credit is a major fixture in the bill,” says Hughes. “The senator recognizes that the private sector has championed the genomics area. We expect that as we continue to develop more of these tests and treatments, we will continue to rely appreciably on the private sector to make the investment.”

Obama aides said that while the incentive takes the form of a tax credit in the bill, what it ultimately may be in the legislation that passes is still unclear. Other suggestions include an extension of patent protections and market exclusivities. The National Academy of Sciences (Washington, DC) has expressed an interest in formally studying the incentive issue; it has already conducted a study on genome patent law. The Institute of Medicine (Washington, DC) is also developing a genome roundtable, and it may be willing to discuss this issue.

“No one is sure what the final incentive will be,” says Hughes. “There are a number of risks that the companies will take. While obviously there are a number of benefits for them as well, this will be an important way to continue to accelerate work in the genomics area. Hopefully we will be able to collaborate with other senators to develop a bipartisan compromise bill. It's hard to predict what the other senators' priorities will be and what preferences they will have with regards to the incentive. But Senator Obama will continue to support the inclusion of a section on some type of financial incentive for the private sector.”

The bill has been referred to the Senate Committee on Finance, largely because of the tax credit provision. The other sections of the bill have been referred to the Senate Committee on Health, Education, Labor, and Pensions. Obama aides believe that it is highly unlikely that the bill will pass this year. Nonetheless, the senator and his staff will continue to increase awareness and education about the genomics issue, and work with outside groups to refine the bill's provisions even further. They also plan to work with other senators on a bipartisan level to develop a compromise bill that will be introduced and hopefully passed in 2007.

“There's enough interest and activity going on in this area, whether it's the hearings on direct-to-consumer genetic tests by Senator Gordon Smith (D–OR) or the bill on laboratory-developed tests by Senator Edward Kennedy (D–MA),” says Hughes. “So it's reasonable to expect that the parties will come together, plus a number of others who've expressed interest privately to develop a compromise bill.”

The bill can be accessed via the following Web site: http://thomas.loc.gov/cgi-bin/query/z?c109:s.3822:. More information about Senator Obama's involvement in this bill can be accessed at his Web site at http://obama.senate.gov/.