Proteases have been implicated in a number of malignant conditions, and researchers have observed increased secretion of proteases into the interstitial fluid around growing tumors. These proteases inevitably act on proteins, including those in the coagulation cascade leading to the formation of fibrin. Furthermore, fibrin is very frequently observed at the invading periphery of malignant neoplasms.¹ Malignant cells also characteristically possess high levels of plasminogen activator, which should induce local fibrinolysis.²

In view of the concurrent increase in the formation of fibrin and in the secretion of proteases in malignant conditions, it is logical to conclude that the measurement of serum fibrinogen degradation product (FDP) levels may represent a useful measure of malignancy. This article describes the results of early studies to explore the viability of an immunoassay, called Oncochek, for the detection of FDPs as indicators of the presence of various cancers.

Background

Tumor cells release proteases into interstitial fluid at a higher rate than normal cells.^3,4 At least four lines of evidence support the concept that this increased protease activity contributes directly to the invasiveness of tumor cells and to the destruction of the adjacent host tissue.^5—7

First, in the case of breast cancer metastases, four classes of proteases appear to be involved in disease progression.⁸ These include cysteine proteases (cathepsins B and L), aspartyl proteases (cathepsin D), collagenases (metalloproteases), and serine proteases (urokinase and plasminogen). Increased expression of the collagenases has been correlated with increased invasiveness of some tumor cells.

Second, down-regulation of these enzymes by genetic means reduces both the invasiveness and metastases of the tumor.⁹ Third, addition of tissue metalloproteinase inhibitors to tumor cells blocks cell invasion in vitro. Fourth, the administration of either natural or synthetic metalloproteinase inhibitors has been shown to prevent metastasis in a simple lung colonization model.⁹

Protease release by tumor cells can also result in the proteolysis of plasma proteins. Theoretically, the extent of proteolytic degradation of these proteins can be correlated with the activity of the tumor cells and used indirectly to evaluate their tumor burden or degree of malignancy.

Principle of the Technology

Taking advantage of the foregoing information, the authors have devised a method for detecting proteolytic degradation products of plasma proteins with minimal interference from the parent protein (the protease substrate) and have begun to explore the utility of the method as a cancer detection assay. Specifically, the method measures unique epitopes that are manifested secondary to proteolytic degradation of fibrinogen. These epitopes are either sterically or immunochemically unreactive in the native fibrinogen molecule.

Assay specificity is achieved by the use of two different antibodies in a two-site, solid-phase enzymometric assay. The more highly specific antibody, which is immobilized to the solid phase, consists of a murine monoclonal to a glycine-histidine-arginine-proline-leucine-aspartate-lysine-cysteine (GHRPLDKC) octapeptide. The first seven amino acids of this peptide represent an internal sequence within the ß-chain of fibrinogen, which is near the amino terminus and is exposed after initial plasminolysis (residues 15—21).¹⁰ After capture of the proteolytic degradation products of fibrinogen by the immobilized monoclonal antibody, the immune complex is detected by using a highly specific conjugate consisting of polyclonal antifibrinogen labeled with horse-radish peroxidase.

Assay Procedure. Calibrators, controls, and samples in a diluent buffer are added in 100-µl amounts to wells of an antibody-coated microwell plate. After a one-hour incubation at room temperature and a wash step, 100-µl aliquots of conjugate are added to each well. The plate is again incubated for 30 minutes at room temperature, washed, and incubated with 100 µl TMB substrate. After 15 minutes at 25°C the reaction is stopped by adding 100 µl 0.1 N HCl into each well. The absorbance at 450 nm is proportional to the level of FDP in the assay.

Analytical Performance

Calibration Curve. Calibrators for the assay were prepared by controlled plasminolysis of fibrinogen to produce FDPs, which were formed in a time-dependent fashion (data not shown). Intact fibrinogen that was not subject to prior treatment with plasmin was unreactive in the current assay. Plasmin-digested fibrinogen exhibited a curvilinear dose response over the concentration range of 0—1250 ng/ml fibrinogen equivalents (see Figure 1). Sera from cancer patients, which typically contained elevated FDP levels, exhibited dilutional parallelism and linearity to the FDP calibration curve over a dilution range from 5- to 80-fold (see Figure 2).

Figure 1. Standard curve for measurement of fibrinogen degradation products (FDPs) by Oncochek.

Figure 2. Dilution of a high-titer patient sample.

Specificity. The Oncochek assay selectively measures FDPs. Neither fibrinogen fragment D (FD), fibrinogen fragment E (FE), nor native fibrinogen (FG) register any significant overt cross-reactivity with the Oncochek assay (Table I). Absorbance values at 450 nm reflect the relative immunoreactivities in this assay system.

Table I. Analytical specificity of the Oncochek assay. Each component is added at a concentration of 500 ng/ml fibrinogen equivalents.

Figure 3(A) illustrates results indicating that FD affects FDP measurements in the Oncochek assay in a pattern consistent with noncompetitive inhibition or covert cross-reactivity.¹¹ This inhibition pattern is consistent with the mechanism that FD binds to the solid phase of capture antibody, thus reducing the antibody sites available for binding FDPs. The double reciprocal plots of FE and FG inhibition studies are consistent with the absence of interaction between MAb and FE and FG (see Figures 3(B) and 3(C)). They are also consistent with the results presented in Table I, which shows the lack of response by FE and FG in the Oncochek assay.

Figure 3. Selectivity of the Oncochek assay: (A) noncompetitive inhibition by fibrinogen fragment D; (B) double reciprocal plot for fibrinogen; and (C) double reciprocal plot for fibrinogen fragment E.

Clinical Performance. Results of the Oncochek assay indicate that FDP levels in the sera of patients with various types of cancer are significantly elevated in comparison to normals. For example, FDP levels in the sera of normal control subjects were compared with those in the sera of patients with five types of cancers. Each group consisted of 50 patients and included breast, colon, lung, ovarian, and prostate cancers. The data presented in Figure 4 were subjected to a receiver-operating-characteristics (ROC) analysis to assess the relationship between the sensitivity and specificity of the assay at various threshold concentrations of FDP. By ROC analysis, using an upper limit of normal corresponding to 96% specificity, sensitivities of 84, 82, 82, 34, and 60% were achieved for breast, colon, lung, ovarian, and prostate cancers, respectively (see Table II). If an elevation in the value of either the Oncochek assay or the organ-specific marker (or both) was used as a prediction of the presence of cancer, sensitivities approximating 90% or great-er were achieved for breast, colon, and lung cancers.

Table II. Clinical sensitivity of the Oncochek immunoassay compared to that of various organ-specific markers.

Results shown in Table II and Figure 4 suggest that the Oncochek immunoassay can detect multiple cancers with a high degree of specificity and clinical sensitivity. When it is used with established organ-specific markers, improved clinical sensitivity may be achieved for breast, colon, and lung cancers.

Figure 4. FDP levels in the sera of normal control subjects and patients with five types of cancer (n = 50).

Discussion

When used in conjunction with the recognized organ-specific tumor marker for breast, colon, and lung cancers, the unique epitope detected by the Oncochek immunoassay system appears to offer increased clinical sensitivity. The results for ovarian (CA 125) and prostate (PSA) cancers are less dramatic. Prospective clinical studies are needed to elucidate the interrelationships between Oncochek, the established tumor markers, and disease progression.

Biochemical specificity of the assay, which is central to its clinical utility, was ensured by use of a monoclonal antibody against a GHRPLDKC octapeptide conjugated to bovine serum albumin. The first seven amino acids of the octapeptide corresponds to amino acids 15 to 21 of the ß-chain of human fibrinogen.¹² This antibody recognizes fragment D, the proteolytic product of fibrinogen plasminolysis, but does not recognize either fragment E or intact native fibrinogen. Thus, in the current assay format, the immobilized monoclonal antibody will capture both fragment D and FDPs, but only FDPs are detected in the assay by the polyclonal antifibrinogen that has been labeled with horseradish peroxidase.

Conclusion

The novel immunoassay for cancer detection described in this article offers three major performance advantages. First, it enables immunochemical measurements of proteolytic degradation products in the presence of, and without interference by, the parent protein molecules (the substrate). Second, the assay detects multiple cancers with a high degree of specificity and sensitivity. Third, when used together with established organ-specific markers, the overall clinical performance may be improved. The authors are currently expanding their initial studies to include a larger patient population as well as other types of cancers and benign disease states such as inflammation and infection.

References

1. Hiramoto R, Berneck J, Jurandowski J, et al., "Fibrin in Human Tumors," Cancer Res, 20:592—593, 1960.

2. Ossowski L, Quigley JP, Kellerman GM, et al., "Fibrinolysis Associated with Oncogenic Transformation," J Exp Med, 138:1056—1064, 1973.

3. Sylven B, "Lysosomal Enzyme Activity in the Interstitial Fluid of Solid Mouse Tumour Transplants," Eur J Cancer, 4:463—474, 1968.

4. Sylven B, "Cellular Detachment by Purified Lysosomal Cathepsin B," Eur J Cancer, 4:559— 562, 1968.

5. Poole AR, Tiltman KJ, Recklies AD, et al., "Differences in Secretion of the Proteinase Cathepsin B at the Edges of Human Breast Carcinomas and Fibroadenomas," Nature, 273:545—547, 1978.

6. Keppler D, Abrahamson M, and Sordat B, "Secretion of Cathepsin B and Tumor Invasion," Biochem Soc Trans, 22:43—49, 1994.

7. Pietras RJ, Szego CM, Roberts JA, et al., "Lysosomal Cathepsin B—Like Activity: Mobilization in Prereplicative and Neoplastic Epithelial Cells," J Histochem Cytochem, 29:440—450, 1981.

8. Dickson RB, Shi YE, and Johnson MD, "A Novel Matrix-Degrading Protease in Hormone-Dependent Breast Cancer," Biochem Soc Trans, 22:49—52, 1994.

9. Goldberg GI, and Eisen AZ, "Extracellular Matrix Metalloproteinases in Tumor Invasion and Metastasis," in Regulatory Mechanisms in Breast Cancer, Lippman ME, and Dickson RB (eds), Boston, Kluwer Academic Publishers, pp 421— 440, 1990.

10. Chung DC, Que BG, Rixon MW, et al., "Characterization of Complementary Deoxyribonucleic Acid and Genomic Deoxyribonucleic Acid for the ß Chain of Human Fibrinogen," Biochemistry, 22:3244—3250, 1983.

11. Suelter CH, A Practical Guide to Enzymology, New York, Wiley, p 248, 1985.

12. Haverkate F, and Timan G, "Protective Effect of Calcium in the Plasmin Degradation of Fibrinogen and Fibrin Fragments D," Thromb Res, 10:803—812, 1977.

That T. Ngo is president and CEO, Michael C. Cress is research scientist, and Ronald J. Moore is vice president of operations at AMDL, Inc. (Tustin, CA). The authors wish to thank Robert L. Oak and Thara Rhamankutty for their technical assistance in preparing this article.

In the 30 years since DNA binding to nitrocellulose membranes was first discovered, membranes have become an integral component of DNA-based assays in both medical diagnostics and forensic laboratory tests. In addition to solid-phase DNA analysis, the Western blots named for the analysis of electrophoretically transferred proteins to nitrocellulose have led to an explosion of membrane-based immunoassays.

Membrane Composition

The manifold varieties of membranes available in the market offer tremendous versatility. Membranes may be used as prefilters, as exemplified by borosilicate glass filters; for general biological sample filtration, as seen with cellulose acetate filters; or for high protein retention and DNA and RNA binding, as seen with nitrocellulose membranes.

Nylon membranes are inherently hydrophilic and work well with aqueous-based samples, but should not be used when maximum protein recovery is the goal. Polysulfone and poly-vinylidene fluoride (PVDF) membranes bind very little pro-tein, exhibit good flow rates, and are chosen depending on the need for solvent resistance. PVDF is highly resistant to most solvents, but polysulfone is generally used for aqueous-based biological samples. Polytetrafluoroethylene (PTFE) membranes are hydrophobic and highly resistant to solvents, acids, alkalides, and propellants. Finally, ultrafiltration membranes (molecular-weight cut-off filters) made up of either cellulose triacetate or polysulfone membranes are generally used for desalting, sample concentration, and deproteinization as well as buffer exchange.^1—4

Membranes offer not only a quick and simple medium for solid-phase fractionation of components but also the ability to easily detect the end products retained on them. These features have enabled the use of membranes as versatile solid supports in cell-based assays, immunoassays, and DNA-based assays, making membrane-based assays one of the fastest- growing classes of diagnostic tests.

In addition to their inherent physicochemical characteristics, membranes are being exploited for their active surface characteristics. The latter may be manipulated either before or after the manufacturing of the membranes.

Realizing the chemical capabilities of membranes, device manufacturers were quick to respond to the demands of assay development by placing them as the bottoms of the traditional plastic 96-well plates. The utility of membrane-bottom plates as versatile formats for assay and process development has in fact been further enhanced by integration with existing systems for liquid handling and detection in the microwell format.

Membrane synthesis and utilization is based on the assumption that a membrane is more than a simple solid support. Indeed, the choice of membrane should be dictated by its final utility to the assay system.

Depth Filters and Screen Filters

Membranes come in a variety of physical and surface chemistries. Simply speaking, they fall into two basic categories: depth filters and screen or microporous filters.^1—4 Photomicrographs of depth filters shown demonstrate their tortuous and complex flow path (see Figure 1).² The structure of the depth filter consists of a matrix of randomly oriented fibers bonded together to form complex channels. This feature provides a high loading capacity.

Figure 1. Photomicrograph showing the complex flow path of a depth filter. Photo Courtesy of Millipore Corp.

The obvious uses of a depth filter are as a prefiltration matrix to extend the life of a membrane filter, and in processes requiring high flow rates and retention of large particulate, such as waste water analysis and receptor assays. The depth filter's disadvantages include the inability to define pore sizes and the retention of large quantities of liquid, a serious impediment to its use with precious fluids.

The screen filter, or microporous membrane filter, as seen in Figure 2, is designed to retain particles on its surface, like a sieve.² The screen filter has a rigid surface with a precisely controlled pore size.

Figure 2. Photomicrograph of particles and bacteria collected on the surface of a microporous membrane (screen filter). Photo Courtesy of Millipore Corp.

The advantages of the screen or microporous membrane filter are low liquid retention and predetermined pore size. It is ideally suited to quantitative retention of particles in such applications as fluid sterilization, fluid clarification, monitoring of air particles, and bioassays. Obviously, low loading capacities lead to rapid clogging of screen filters in high-particle situations. The best functionality of filters is achievable by combining the characteristics of both the depth and screen filter. A depth filter is best used as a prefilter and the screen filter as a final filter.

Filter Selection

The choice of membrane type should be based on those physical and surface characteristics of the membrane that can directly affect assay performance variables such as sensitivity and reliability. Factors such as retention capabilities, hydrophobic or covalent interactions with the membrane surface, and signal-to-noise ratio of the end point analyzed, are highly dependent on the type of membrane.^1—5

A clear understanding of membrane characteristics therefore is essential to the identification of the right filter for biological assays and process development protocols. The membrane of choice should meet the following criteria:

• Ability to bind the component of choice, leading to high specificity.

   

• Ability to retain the immobilized molecule in a biologically active state.

   

• Detection mode compatible with signal-to-noise ratios, leading to high sensitivity.

   

• High tensile strength to withstand transfer conditions, both physical and chemical.

Biomolecule immobilization optimally occurs under clearly defined conditions of binding, either with hydrophobic or covalent interactions or electrostatic attraction. Depending on pore size, nitrocellulose membranes can bind anywhere from 80 to 150 µg of protein per square centimeter² of membrane.⁴ Variables determined by the mode of biomolecule transfer—such as ionic strength, pH, and high currents used in electrophoretic transfer—affect the degree of sample retention.

High specificity of signal can be achieved only with membranes that allow few if any nonspecific interactions with the active surface. Blocking agents that make the active sites unavailable to further interactions generally increase the signal-to-noise ratio of the analyzed end point.⁵

Detection protocols developed to date include radioisotopic, chromogenic, chemiluminescent, and fluorogenic modes on membrane filters. Nitrocellulose membranes work well with all detection modes except fluorescence.⁴

Nylon membranes such as Nytran (Schleicher & Schuell, Keene, NH) and Biodyne membranes (Pall Gelman Sciences, Port Washington, NY) are generally superior to nitrocellulose membranes in durability and in the nonradioactive (chemiluminescent or fluorescent) detection of DNA, RNA, and proteins.¹ Biodyne nylon membranes are available with four types of chemically integral surfaces, providing specific adsorption parameters. Each of the four types of nylon membranes has distinct chemical groups, and the groups that correspond to differences in surface charge and chemical behavior.

Specific Biodyne membranes perform better in certain detection methods. The Biodyne B membrane is the membrane of choice for radioactive detection. The Biodyne A and the Biodyne Plus membranes provide optimal results with nonradioactive detection systems. The FBI laboratories use the Biodyne A membrane for the restriction fragment length polymorphism (RFLP) profile of human genomic loci from blood stains.⁶ This chemiluminescent detection protocol is in fact the gold standard for forensic laboratories evaluating RFLP profiles.

The Membrane-Bottom Plate

The physicochemical properties, versatility, and proven success of membranes as solid supports with chemically active groups made them obvious candidates for integration into existing high-throughput plastic microwell formats. The integration of membranes into microwell plates yielded membrane-bottom plates.

Membrane integration into the microwell format extended the possible applications of membranes by making possible their use with preexisting accessories designed for microwell formats, such as liquid handling, pipetting operations with existing instrumentation, and detection in commercially available radioisotopic, colorimetric, and chemiluminescence plate readers. Assay development in such microwell formats thus offers a versatile tool for high-throughput sample processing on membranes. Exemption from regulatory approval, or a Class I exempt registration, has been granted in the United States for the membrane-bottom plate used in at least one clinical diagnostic test.

Following is a discussion of the use of currently available 96-well membrane-bound microplates in biological assay and process development systems. Also included is an evaluation of the properties of some of the other available membranes, projecting their potential applications into newer product formats, such as 96-well membrane-bottom microplates.

Solid-phase assay systems have the advantages of target immobilization localized in sufficiently high concentrations, fractionation of components without centrifugation, and ease of handling end products on a manageable matrix. Already representing a sizable proportion of the laboratory and diagnostic test market, immunoassays are also moving increasingly into nonmedical applications, including environmental sample analysis, agricultural assays, and food safety testing. Immunoassays are being employed in drug discovery as well, especially in the elucidation of the possible pathways of biological function. Examples of a few of these assays are explored below.

Cell Proliferation Assays. Silent Monitor plates (previously manufactured and sold by Pall Gelman Sciences but now available from Nalge Nunc International, Rochester, NY) have been shown to support the growth of the eukaryotic Dictyostelium cells.⁷ Membrane-cultured Dictyostelium cells are easily processed for a colony-blot hybridization to detect specific RNA transcripts.

MultiScreen HV plates (Millipore Corp., Bedford, MA) also support eukaryotic cell growth. In experiments measuring tritiated thymidine uptake in the MultiScreen HV plates, HB-124 cells (a murine hybridoma cell line, ATCC, Rockville, MD) were easily cultured and processed for subsequent liquid scintillation counting. Figure 3 shows a typical increase in DNA synthesis measured by thymidine uptake.⁸ The micro-porous membranes thus effectively allow nutrients to be available for cell growth on the solid support.

Figure 3. Cell culture and processing with the MultiScreen system. (A) Thymidine uptake versus time in culture for glass-fiber disk and MultiScreen membrane production of cells.. HB-124 cells, in standard culture media containing tritiated thymidine, were added to individual wells (1 x 10⁵ cells/well) of sterile MultiScreen HV plates, and the plates were incubated at 37°C in a 5% carbon dioxide atmosphere. Nonspecific binding controls were prepared by adding tritium-labeled thymidine to wells containing media without cells. The total volume per well was 0.2 ml; 0.2 ml of ultrapure Milli-Q system water was added to unused wells of the plate. At indicated times, the MultiScreen plates were removed from the incubator and processed. The results show a typical increase in DNA synthesis with time measured by thymidine uptake. (B) Thymidine uptake effect of stimulation by serum. HB-124 cells were added to individual wells (5 x 10⁴ cells/well) of a sterile MultiScreen-HV plate and incubated at 37°C in a 5% carbon dioxide atmosphere for four days to deplete the medium of nutrients. On day 4, 50 µl of 10 different dilutions (0.01—8%) of bovine calf serum were added, and the incubation continued overnight. At the end of serum stimulation, tritiated thymidine (concentration: 0.05 µCi/well) was added and incubated for two hours. The plates were read following processing. The results demonstrate that following nutrient depletion of HB-124, cells are stimulated by as little as 0.1—0.5% serum in a dose-response fashion. Courtesy of Millipore Corp.

Cell-Based Receptor-Binding Assays. Receptor-binding assays require extensive incubation, washing, and quantitative collection of filtrates and filters. The hydrophilic microporous Durapore PVDF membranes in MultiScreen formats are used in several cell-based receptor-binding assays. In this application, the PVDF filters allow for the separation of bound labeled ligand from free labeled ligand.

Figure 4 shows the inhibition of 125I-angiotensin II binding with a prototypical experimental drug in adrenal gland preparations.⁹ 125I-labeled angiotensin was added to adrenal membranes in MultiScreen-HV 0.45-µm plates and incubated for 90 minutes at room temperature. Nonspecific binding was measured in the presence of 125I-angiotensin II, where antagonists were also present. This rapid screening system has been standardized for routine determinations of the binding affinities of newly developed chemical compounds.⁹

Figure 4. Cell-based receptor-binding assay. 125-I angiotensin II was added to adrenal membranes in MultiScreen HV (0.45 µm) plates and incubated 90 minutes at room temperature. Each 250-µl incubate consisted of the following (final concentration): 50 mM Tris, 120 mM NaCl, 5 mM MgCl2, 1 mM dithiothreitol, 0.05% bovine serum albumin, 0.1 nM 125I-angiotensin II and 8—15 µg adrenal membrane protein. Antagonists were added in concentrations from 10 nM to 100 µM. Nonspecific binding was measured in the presence of 0.1 µM SAR1, Ile8-angiotensin II. Binding was terminated by applying vacuum to filter plates. Receptor-ligand complex trapped on filters was washed three times with 300 µl ice-cold wash solution (50 mM Tris, 120 mM NaCl, 5 mM MgCl2, 1 mM dithiothreitol). Filter disks were dried, punched out, and counted in gamma counter (52% efficiency). Courtesy of Millipore Corp.

Enzyme Assays. Membrane-bottom microwell plates facilitate exploitation of novel matrices, like beads, in traditional immunoassays. The surface area for ligand binding is greater with beads than with simple coating of the wells of a traditional microwell plate. The coated immunosorbent beads may provide less steric hindrance and therefore improved sensitivity of the immobilized antibody.

This principle is evident in a report published by Orthner and colleagues of the American Red Cross (Rockville, MD).¹⁰ Their assay was designed to measure enzymatically active activated protein C (APC) in plasma samples using MultiScreen Durapore type DV membranes. APC is a serine protease and a naturally occurring antithrombotic enzyme. It functions as an anticoagulant by proteolytically inactivating coagulation factors VIIIa and Va. APC bound to a monoclonal antibody on immunosorbent beads was analyzed following its elution through Durapore MultiScreen plates. The immunosorbent bead protocol proved to yield a much more sensitive assay than did the monoclonal antibody-coated microwell system.

Immunosorbent resins may be used in the creation of microcolumns in 96-well membrane-bottom microwell plates. With the increase in surface area, the beads very likely favored an increased diffusion of reactants, such as APC to antibody-coated porous beads, resulting in improved assay sensitivity. Similarly, minicolumns have also been used with Alumina to quantify c-AMP phosphodiesterases, and with Sephadex G-50 for the purification of DNA.^11,12

Nylon membranes in the Silent Screen plates available from Nalge Nunc International (and previously in the Silent Monitor plates) have proved to be excellent supports for immunoassays in the rapid diagnosis of influenza A infection.¹³

Efficient Filtration and Quantitative Filtrate Collection. Federal law mandates that neonates be tested for phenylketonuria and galactosemia, disorders that result in the infant's inability to produce enzymes necessary in the breakdown of certain food products. The bacterial inhibition assay and a fluorometric assay are currently available to test for these enzyme deficiencies.¹⁴ Infant blood samples collected on a Guthrie card (Schleicher & Schuell) are punched out and tested for the levels of phenylalanine by the bacterial inhibition assay described by Guthrie and Susi.¹⁴ This test provides only a semiquantitative reading, is subject to interference by antibiotics, and lacks automatability. Several fluorometric tests have now been developed to replace it.

Hoffman and associates pioneered the use of fluorometric detection of the enzyme digestion end point in the 1980s.¹⁵ Both the Shield Diagnostics Quantase system (available in the United States and Canada from Wallac, Inc., Gaithersburg, MD) and the AP1300 System from Astoria Pacific Corp. (formerly Alpkem, Inc., Clackamas, OR) were developed for automated fluorometric assays for phenylalanine, galactose, and other mandated tests. The fluorometric assay system, however, is very sensitive to blood debris and contaminating fibers from the Guthrie cards. Filtration through membrane-bottom plates clears the reactants of filter fibers or blood debris, enabling a clear filtrate to be read fluorometrically.¹⁶ The Millipore Durapore plate made with PVDF filters is used with both commercially available fluorometric systems. Quantase system kits include MultiScreen DV plates, which are registered in the United States as Class I exempt devices specifically for PKU and galactose screening.

Purification of PCR Products. Purification or fractionation of large-molecular-weight DNA, such as PCR products, from oligonucleotides and unincorporated nucleotides is easily accomplished by column chromatography through resins such as Sephacryl-500HR (Pharmacia LKB, Uppsala, Sweden). Such a procedure may be easily adapted into high-throughput 96-mini-spin-columns by packing the prewetted Sephacryl-500HR filtration matrix into either the Silent Monitor (Nalge Nunc International) or the MultiScreen filtration plate (Millipore Corp.).¹⁷

Packing of the column in each well is achieved through centrifugation of the resin-loaded membrane-bottom plate to achieve a column height of about three-quarters of the depth of the well. PCR products are applied to the center of each column with a multichannel pipette, then centrifuged using a microplate rotor. Both the flow-through material and the material eluted with buffer contain pure PCR products. In the 96-well membrane-bottom plate, therefore, column chromatography of single samples evolved into spin-column chromatography of 96 samples in parallel.

High-Throughput Preparation of Yeast Artificial Chromosome DNA. The yeast artificial chromosome (YAC) is an important vector for cloning large pieces of DNA. With DNA fragments as long as 250 to 300 kilobases, YAC grown in a microwell plate system does not reach optimal density levels. DNA preparation from YAC cells is laborious and time consuming.

The use of MultiScreen plates allows high cell densities and simplifies DNA extraction from YAC cells.^18,19 As illustrated in Figure 5, the MultiScreen 96-well filter-bottom plate is first embedded in solid growth media poured into the lid of a 96-tip yellow-tip rack. With additional medium available through dialysis of nutrients by the membrane, YAC cell densities of much greater than 10⁷ cells per well are achieved. YAC DNA extraction is accomplished through the zymolase treatment of the YAC cells embedded in agarose plugs in the MultiScreen plate followed by dialysis of the DNA through the micro-porous membrane of the MultiScreen plate.

Figure 5. High-throughput preparation of yeast artificial chromosome (YAC) DNA using MultiScreen filtration plates. (A). Growing YACs to high density (growth plate). The MultiScreen filter-bottom plates are embedded in solid culture media so that the yeast, while isolated in wells, is able to take advantage of more medium by dialysis of nutrients through membrane. This technique produces 10 times the cells from each clone, for about 3 µg of DNA. (B). Extracting DNA (dialysis plate). This method is a modification of a standard protocol using agarose blocks for DNA preparation for pulsed-field gel electrophoresis. The blocks are formed within the wells of the MultiScreen filtration plates. Dialysis through the membrane bottom allows DNA preparation without manipulation of the blocks. For high throughput, several plates may be si-multaneously dialyzed in a simple "flooding" chamber. Courtesy of Millipore Corp.

High-Throughput Synthesis of Drugs. Membrane-bottom plates are now being used in the process development of 96-well automated synthesis of small-molecule drug libraries. Figure 6 demonstrates the setup of the Protogene polypropyl-ene filter-bottom plate with a polypropylene filter (Protogene Laboratories, Inc., Palo Alto, CA) used in the synthesis of 96 different chemical libraries for drug discovery.²⁰ The polypro-pylene filter-bottom microwell plate may be substituted for assays where resin usage is combined with the filtration mode, as for nucleic acid purification, desalting, or ion-exchange chromatography. The option of substituting a variety of filters allows a more versatile adaptation of the Protogene filter plate to both assay and process development. Indeed, other commercially available and user-fabricated membrane-bottom plates are also being used as reaction vessels in nucleic acid extraction and purification, as well as in the parallel synthesis of 96 different chemical libraries per unit.

Figure 6. The Protogene Laboratories, Inc., filter-bottom plate, a V-bottom polypropylene plate. (A) Each well ends in a capillary exit designed to prevent any cross-talk. The format allows use of the plate as a filter, a spin-column support, or an active membrane support in a 96-well format. (B) The plate is a reaction chamber in the 96-well parallel automated synthesis of small-molecule drug-library by combinatorial chemistry protocols.

Future Developments

In addition to the membrane types currently available in microwell plates, functionally specific membranes developed for future products may be useful for high-throughput sample preparation for DNA diagnostics. Huang and associates have published a report of DNA preparation in a high-throughput format in the Silent Monitor plates (now available as Silent Screen plates).²¹ In plasmid and cosmid DNA preparation in 96-well deep-well plates, the cell debris from lysed bacterial cells is mixed with a polyelectrolyte protein-precipitating reagent (Affinity Technologies, Inc., Fairfield, NJ) and processed for DNA purification by filtration through the Biodyne membranes of Silent Screen plates. Furthermore, there is an acute need for DNA sample preparation before blood is accepted for genotyping and before screening tests are carried out for infectious agents in donated blood.

Leukosorb medium, a fibrous material that selectively binds leukocytes (see Figure 7), is used in Pall leukocyte reduction filters (Pall Gelman Sciences). The leukocyte reduction filters were developed with the Leukosorb matrix for the depletion of leukocytes from transfusion products. The Leukosorb membrane in a 96-well membrane-bottom format is suitable for use in sample preparation or diagnostic tests on leukocytes. Such tests may include genotyping of the leukocytic DNA and identification of intracellular viruses harbored in the leukocytes retained by the Leukosorb medium.

Figure 7. Leukosorb medium is a white, highly wettable, fibrous matrix designed for use in procedures requiring separation of leukocytes from clinical fluids such as blood. Photo Courtesy of Pall Gelman Sciences.

The Hemasep V membrane (Pall Gelman Sciences) provides a noncentrifugal method of sample preparation.²² In this method, plasma is vertically separated from small quantities of whole blood. Thus extracellular viruses, such as hepatitis viruses and human immunodeficiency viruses, could theoretically be identified from the plasma of the blood separated on the membrane.

The Immunodyne ABC membrane (Pall Gelman Sciences) also has potential applications in 96- or higher multiple-well formats. A chemically activated nylon 6,6 membrane, it covalently binds proteins or amino-terminated oligonucleotides. Covalently bound molecules are especially useful in biosensor applications.

Conclusion

Although membrane-bottom plates have been around for just over a decade, they are now used as much more than merely solid supports for reactants. Membranes offer a higher surface area than traditional plastic surfaces. Their porosity increases their versatility. Surface-modification capabilities for reactive groups and charges on membranes allow ionic, hydrophobic, or covalent binding with target molecules. Membranes that are chemically resistant to strong organic solvents enable high-throughput organic synthesis of chemicals.

The inclusion of membranes in 96-well membrane-bottom formats has combined all the advantages mentioned above with all the capabilities of 96-well formats. The latter include use in robotic workstations, with modularization and automated liquid handling and assay end point reading. Future developments include a move beyond 96-well formats into 384- or higher multiple-well formats for high-throughput manipulation.

References

1. "Applications Guide for Pall Membranes," Port Washington, NY, Pall Gelman BioSupport Div., 1996.

2. "Millipore Corporation Catalog," Bedford, MA, Millipore, Corp., 1997.

3. "Alltech Bioscience Lab Resources," 1st ed, Deerfield, IL, Alltech Bioscience, 1997.

4. "Products for Life Science Research," Keene, NH, Schleicher & Schuell, 1995—1996.

5. Dubitsky A, "Blocking Strategies for Nylon Membranes Used in Enzyme-Linked Immunosorbent Assays," IVD Technol, 3(4):53—59, 1997.

6. Guisti AM, and Budowle B, "A Chemiluminescent-Based Detection System for Human DNA Quantitation and Restriction Length Polymorphism (RFLP) Analysis,"Appl Theoret Electrophoresis, 58:89—98, 1995.

7. Maniak M, Saur U, and Nellen W, "A Colony-Blot Technique for the Detection of Specific Transcripts in Eukaryotes," Anal Biochem, 176:78—81, 1989.

8. "MultiScreen Methods: Thymidine Uptake Assays Using the MultiScreen System," lit #TB038, Bedford, MA, Millipore Corp., 1994.

9. "MultiScreen Methods: Cell-Based Receptor Binding Assays Performed with the MultiScreen Assay System," lit #MM011, Bedford, MA, Millipore Corp., 1996.

10. Orthner CL, Kolen B, and Drohan WN, "A Sensitive and Facile Assay for the Measurement of Activated Protein C Activity Levels in Vivo," Thromb Haemost, 69(5):441—447, 1993.

11. Daniels DV, and Alvarez R, "A Semiautomated Method for the Assay of Cyclic Adenosine 5'-Monophosphate Phosphodiesterase," Anal Biochem, 236:367—369, 1996.

12. "MultiScreen Methods," Bedford, MA, Millipore Corp., in press.

13. Duverlie G, et al., "A Nylon Membrane Enzyme Immunoassay for Rapid Diagnosis of Influenza A Infection," J Virol Methods, 40:77—84, 1992.

14. Guthrie R, and Susi A, "A Simple Phenylalanine Method for Detecting Phenylketonuria in Large Populations of Newborn Infants," Pediatrics, 32:328, 1963.

15. Hoffman GL, Laessig RH, Hassemer DJ, et al., "Dual-Channel Flow System for Determination of Phenylalanine and Galactose Application to Newborn Screening," Clin Chem, 30(2):287—290, 1984.

16. Elvers LH, Diependaal GAM, Blonk HJ, et al., "Phenylketonuria Screening Using the Quantase Phenylalanine Kit Combination with a Microfilter System and the Dye Tartrazine," Screening, 3(4):209—223, 1995.

17. Wang K, Gan L, Boysen C, et al., "A Microtiter Plate-Based High-Throughput DNA Purification Method," Anal Biochem, 226:85—90, 1995.

18. MacMurray AJ, Weaver A, Shin H-S, et al., "An Automated Method for DNA Preparation from Thousands of YAC Clones," Nucleic Acids Res, 19(2):385—390, 1991.

19. "MultiScreen Filtration System: High-Throughput Preparation of YAC DNA Using MultiScreen Filtration Plates," lit #TB061, Bedford, MA, Millipore Corp., 1991.

20. Molinari RJ, "Solid-Phase Synthesis of Small Molecule Drug Libraries Using Second-Generation 96-Well Array Synthesizer," presented at the Second Annual Solid-Phase Synthesis Meeting, Coronado, CA, February 6—7, 1997.

21. Huang GM, Wang K, Kuo C, et al., "A High Throughput Plasmid DNA Preparation Method," Anal Biochem, 223:35—38, 1994.

22. Alter J, "Single-Step Vertical Plasma Separation of Whole Blood for Test and Sample Prep," Genet Engineer N, November 15, 1996.

Asha A. Oroskar, PhD, is president of Oros Technologies, Inc. (Oak Brook, IL), and head of industrial research at the Center for Biotechnology, Northwestern University (Evanston, IL).

In the Flowmetrix system, a single sample can contain up to 64 discrete microsphere sets differentiated by their unique red-orange fluorescence. Each set carries the reactants of a distinct bioassay. Illustration by Keith Kasnot

In the human body, an effective immune response to even the simplest antigen requires the properly orchestrated activity of tens, if not hundreds, of distinct molecules including cytokines, enzymes, and eventually antibodies. This complex system of biochemical control employs thousands of biomolecules to coordinate the many processes necessary for proper function. Disease, however, alters the balance of these control elements, and it is this deviation from normal that forms the basis of much of today's diagnostic laboratory medicine.

Current laboratory methods often look at only one piece of this complex biochemical puzzle. Rarely, though, are diseases confined to a single, isolated molecular abnormality. The predictive value of diagnostic testing is increased significantly when it supplies many relevant pieces to the puzzle. For example, maternal alpha-fetoprotein (AFP) was initially used as a screen for fetal neural tube defects. It was subsequently discovered that AFP could identify approximately 20% of fetuses with Down's syndrome. Unfortunately, a high false-positive rate and low predictive value made reliance on AFP alone of questionable benefit. When combined with human chorionic gonadotropin and free estriol however, approximately 70% of fetuses with Down's syndrome could be identified prenatally.¹ It is likely that other markers will be discovered which could raise this percentage even higher.

Similar examples of synergistic testing exist throughout laboratory medicine, and more will be developed as additional biochemical interrelationships are described. However, the economic environment of today's clinical laboratory clearly demands cost-efficient testing. Future in vitro diagnostics must deliver the advantage of advanced multiple analyte quantitation in a simple, cost-effective format. To meet this demand, Luminex Corp. (Austin, TX) has developed a flow cytometry—based approach to multiplexing assays, called the Flowmetrix system.

System Description

Flowmetrix is a unique platform currently capable of performing 64 different assays in a single specimen aliquot using a flow cytometer and advanced digital signal processing hardware and software.^2—4 The system depends on the signal processor's ability to classify polystyrene beads (microspheres) dyed with distinct proportions of red and orange fluorophores when illuminated by the flow cytometer (see Figure 1). The reagents of each separate assay are allocated to the individual sets of microspheres, all 5.5 µm in diameter. A green fluorescent reporter molecule is used for quantitation of the analyte. The flow cytometer is gated at 5.5 µm so that only fluorescence associated with the bead is measured, eliminating the need for wash steps. This also eliminates any interferences caused by icteric, hemolytic, or lipemic samples. In fact, assays performed in whole blood are not affected by cellular elements, since few components will enter the 5.5-µm gate and none will have the unique combination of different fluorescences that identify the unique bead sets.

Figure 1. The surface of each microsphere contains multiple carboxyl groups that function as sites for covalent ligand attachment. In their interior, precise proportions of red and orange fluorophores identify discrete microsphere sets.

Flowmetrix differs from other microsphere-based multiplex systems in that only a single size of microsphere is used. Classification is based solely on the fluorescence of multiple dyes within the beads. The number of possible bead sets is established by the number of lasers exciting different dyes in the microspheres. For instance, the 64 assays currently performed by the Becton Dickinson Facscan increases to several hundred performed simultaneously on the dual-laser Facscalibur. This number could reach into the thousands and beyond with this year's introduction of a solid-state benchtop analyzer designed specifically for Flowmetrix applications. By combining small diode lasers with inexpensive digital signal processors and microcontrollers, it is expected that this instrument will perform rapid, multicolor analysis for about one-fifth the cost of conventional flow cytometers ($20,000 compared to $100,000).

The reagent components of the Flowmetrix system are also economical. Fluorescent dyes and microspheres are relatively inexpensive, and reagent usage on these very small microspheres is commonly 1% or less than that of microwell-based assays. Preanalytical specimen preparation is also markedly decreased, resulting in lower costs for labor and consumables. With these advantages, multiplexed analysis can be performed for the cost of most single immunoassays, enabling physicians to make use of the predictive power of synergistic testing for potential disease-causing agents (see Table I).

Table I. Comparison of intralaboratory cost factors for performing a cardiac marker panel on the Flowmetrix system versus doing so using other methods. For nonmultiplexed tests, costs increase in direct proportion to the number of analyses performed.

Synergistic Panel for Myocardial Infarction

An example of the multiplexed testing made possible by the Flowmetrix system can be seen in the analysis of biochemical markers for myocardial infarction. These markers have traditionally been evaluated individually rather than as synergistic groups. However, by combining the sensitivity of myoglobin with the specificity of troponin I and the general utility of creatine kinase-MB (CK-MB), the positive and negative predictive value of the group far surpasses that of any of the individual tests.^5—7 Since digoxin toxicity can mimic myocardial infarction, digoxin was also added to the panel.⁸ Despite the different molecular sizes and reference ranges of these analytes, all are measured simultaneously from a single aliquot and reported in real time (see Figure 2). If other markers, such as fatty acid binding protein, glycogen phosphorylase BB, or atrial natriuretic peptide demonstrate clinical utility, they could be easily added to the panel.

Figure 2. Laser illumination of microspheres (A) elicits red-orange internal fluorescence (B and C), identifying each microsphere set. Reporter molecules added with the patient sample quantitate the reactions by the intensity of their associated green fluorescence (D).

The system could also allow simultaneous measurement of biomolecules that regulate blood flow and pressure (e.g., aldosterone, renin, prorenin, antidiuretic hormone), since these analytes may have heretofore unrecognized diagnostic and predictive value. Since all of these analytes could be assayed at once, a previously unrecognized clinical efficacy could be unveiled.

Multiplexed Flowmetrix assays begin with the separate development of each component assay. For the cardiac panel, distinctly colored bead sets were coupled to the appropriate ligand through either chemical or avidin-biotin linkages. All reactants were then titrated to generate standard curves covering the physiologically significant ranges of analyte concentrations. Sensitivity and range can be balanced by altering the degree of saturation of capture ligate on an individual bead set, as well as by the number of microspheres used in the assay development. Regardless of format, a typical coupling to a microsphere set uses 2 ng of reactant per assay. Approximately 50 ng of green fluorescent reporter molecule is required per assay.

Each assay was designed to provide limits of detection significant to confirm or disprove occurrence of myocardial events as well as to cover physiologically significant ranges for each ligate (see Table II). For the most sensitive quantitative results, Flowmetrix assays require incubation times typical of immunoassays (approximately 15 minutes). However, semiquantitative data can be available in less than 5 minutes due to the near-liquid phase kinetics of soluble microsphere-based reactions. Multiple results on each individual patient require less than 10 seconds instrument analysis time. Correlation with predicate technologies has been excellent on every application attempted, including cardiac markers, hormones, tumor markers, allergy testing, deoxyribonucleic acid (DNA)—based assays, and others. The system is a rapid, powerful, and flexible method for analyzing multiple biomolecular events in real time.

Table II. Functional limits of detection and dynamic range for a cardiac marker panel on the Flowmetrix system.

With the Flowmetrix system, extensive risk-stratification panels could be developed. For example, a test for an individual tumor marker is now quite expensive. As components of a multiplexed Flowmetrix test, however, entire panels could be delivered in a cost-effective manner. Markers such as free and bound prostate-specific antigen (PSA) could easily be combined with other tests such as CK-BB to increase predictive value. In addition, cost constraints have made it difficult for many patients to have their coronary risk properly categorized. A cholesterol test alone is inadequate for risk stratification. The Flowmetrix system will allow complete coronary risk analysis, including high-density lipoprotein, low-density lipoprotein, apolipoproteins, lipoprotein a, and other markers to be performed simultaneously, at a reduced cost. As other important analytes are identified, such as C-reactive protein, they can be easily added with no system modification. In fact, virtually any other test grouping is possible, including cytokines, DNA-based testing, allergy, autoimmunity, or combinations thereof.

How it works: Flowmetrix assay development

The Flowmetrix system can perform multiplexed analysis of up to 64 different reactions simultaneously by using a flow cytometer and digital signal processor to perform real-time analysis of multiple microsphere-based assays. There are three major components of the system: a benchtop flow cytometer, microspheres, and computer hardware and software.

Figure 3. Preparing for a capture-sandwich format assay on the Flowmetrix system: (A) microsphere preparation; (B) reporter preparation; (C) positive sandwich assay result recorded by the flow cytometer.

The flow cytometer analyzes individual microspheres by size and fluorescence, simultaneously distinguishing three fluorescent colors: green (530 nm), orange (585 nm), and red (>650 nm). Microsphere size, determined by 90° light scatter, is used to eliminate microsphere aggregates from the analysis. Red and orange fluorescences are used for microsphere classification, and green fluorescence is used for analyte measurement. The Flowmetrix system is currently configured for the Becton Dickinson Facscan, a multiparameter flow cytometer that uses a single 488-nm excitation laser (Becton Dickinson Immunocytometry Systems, San Jose).

To prepare a multiplexed assay, individual sets of microspheres are conjugated with the target molecules required for each reaction (see Figures 3 and 4). For example, one microsphere set might contain 15% red and 85% orange fluorescence whereas a different set might contain the exact opposite ratio. Targets may be antigens, antibodies, oligonucleotides, receptors, peptides, enzyme substrates, or other types of molecules.

Figure 4. A competitive inhibition format assay on the Flowmetrix system: (A) microsphere preparation; (B) reporter preparation; (C) negative result shows reporters binding to the microsphere; (D) positive result shows reporters binding to competitive target molecules.

A fluorescent reactant is then prepared for each target molecule. A reactant may be any molecule that will bind to the target molecule, including oligonucleotides, antigens, antibodies, receptors, and so on. After optimizing the parameters of each assay separately, the assays can be multiplexed by simply mixing together the different sets of microspheres. The fluorescent reactants are also mixed to form a cocktail for the multiplexed reactions.

The microspheres are then reacted with serum, for example, followed by the cocktail of fluorescent reactants. After a short incubation period, the mixture of microspheres—now containing varying levels of green fluorescence on their surfaces—are analyzed with the flow cytometer. Data acquisition, analysis, and reporting are performed in real time on all microsphere sets included in the multiplex. As each microsphere is analyzed by the flow cytometer, the microsphere is classified into its distinct set, based on red and orange fluorescence, and the green fluorescence value is recorded (see Figure 2). One hundred individual microspheres of each set are analyzed and the mean value of the green fluorescence is reported. Using a typical standard curve, true quantitative results can be achieved, with precision enhanced by the analysis of 100 microspheres per data point.

Conclusion

Luminex has miniaturized and multiplexed the long-established microsphere assay format, markedly reducing reagent and consumable costs as well as sample requirements. Because the significant computing power driving Flowmetrix is also essentially free, the only unaffected costs associated with diagnostic testing are those involving sample collection, transport, and handling. In practice, therefore, it will be important to maximize the information derived from each sample. This will enable practitioners not only to avoid repetition of the expensive preanalytical costs, but also to take full advantage of the diagnostic power of synergistic testing.

As is the case with other technologies that are being driven by rapid developments in the microelectronics industry, significant performance gains should be a predictable component of the Flowmetrix system.

References

1. Haddow JE, Palomaki GE, Knight GJ., et al., "Reducing the Need for Amniocentesis in Women 35 Years of Age or Older with Serum Markers for Screening," New Engl J Med, 16:1151-c, 1994.

2. McDade RL, and Fulton RJ, "True Multiplexed Analysis by Computer-Enhanced Flow Cytometry," Med Dev Diag Indust, 19(4):75—82, 1997.

3. Fulton RJ, McDade RL, Smith PL, et al., "Advanced Multiplexed Analysis with the Flowmetrix System," Clin Chem, 43(9):1749—1756, 1997.

4. Gordon RF, and McDade RL, "Multiplexed Quantification of Human IgG, IgA, and IgM with the Flowmetrix System," Clin Chem, 43(9):1799— 1801, 1997.

5. Antman EM, Tanasijevic MJ, Thompson B, et al., "Cardiac-Specific Troponin I Levels to Predict the Risk of Mortality in Patients with Acute Coronary Syndromes," New Engl J Med, 18:1342— 1349, 1996.

6. McLaurin MD, Apple FS, Voss EM, et al., "Cardiac Troponin I, Cardiac Troponin T, and Creatine Kinase MB in Dialysis Patients without Ischemic Heart Disease: Evidence of Cardiac Troponin T Expression in Skeletal Muscle," Clin Chem, 43(6):976—982, 1997.

7. Wu AHB, Feng YJ, Contois JH, et al., "Comparison of Myoglobin, Creatine Kinase MB,
and Cardiac Troponin I for Diagnosis of Acute Myocardial Infarction," Ann Clin Lab Sci, 26: 291—300, 1996.

8. Harrison's Principles of Internal Medicine, 13th ed, Isselbacher, Braunwalk, Wilson, et al. (eds), New York, McGraw-Hill, 1994.

Michael Spain, MD, is medical director at Luminex Corp. (Austin, TX).

Two years ago, FDA identified analyte specific reagents (ASRs) as its highest regulatory priority in the field of IVDs. Now the agency has issued its final rule on the subject. Published in the Federal Register on November 21, 1997, the rule will become effective on November 23, 1998.¹

It has been several years since FDA first asserted its authority to regulate so-called "home-brew" assays directly through the premarket approval (510(k)) process. While the agency has not disavowed its jurisdiction over laboratories that produce ASRs, the final rule takes a different direction. At least for now, the agency has abandoned the notion of regulating home-brew assays directly. Instead, under the ASR regulation, it has shifted the regulatory focus away from the assays developed in laboratories in favor of regulating the reagents purchased by laboratories to develop those assays.

In its final version, then, the stated goal of the ASR regulation is to ensure the quality of the reagents used by IVD manufacturers and laboratories that make their own home-brew assays. This approach largely mirrors that of FDA's first proposed rule for ASRs, which was published on March 14, 1996, shortly after an advisory panel endorsed the ASR concept. The final rule amends two FDA regulations, 21 CFR 809.10 and 864.4010, and adds two new sections, 21 CFR 809.30 and 864.4020.

ASRs Defined. One of the most important parts of the rule is its definition of ASRs. The language of the final rule defines 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.²

This definition differs from the 1996 proposal by adding the term ligand, because ligands bind the reagents to the analytes. The final rule also clarifies the fact that binding between ASRs and their analytes may take place through either physical or chemical means. And FDA agreed with industry comments supporting the inclusion of nucleic acids in the ASR definition. But the most important change is the addition of the term diagnostic, making it clear that a substance can be an ASR only if it is intended for diagnostic use.

Application to Manufacturers. Not everything that falls within this definition will be regulated under the rule. ASRs are excluded from the provisions of the rule when they are sold to IVD manufacturers or organizations that use the reagents to make tests for purposes other than providing diagnostic information to patients and practitioners (e.g., forensic, academic, research, and other nonclinical laboratories).

The final rule classifies or reclassifies the majority of ASRs as Class I medical devices exempt from 510(k)s. (The FDA Modernization Act of 1997 independently exempts Class I devices from 510(k) notices.) However, ASRs will be subject to general controls. Accordingly, suppliers of ASRs must satisfy the following requirements:

• Register with FDA and provide it with a list of the ASRs they supply to laboratories for use in developing in-house tests.

   

• Comply with the requirements of FDA's quality system regulation (21 CFR 820), including good manufacturing practices (GMPs), as applicable.

   

• Comply with the requirements of the medical device reporting (MDR) regulation (21 CFR 803).

The ASR industry includes many small device manufacturers, some of which have not previously been subject to GMPs. Some companies commented that the proposed rule, by requiring GMP compliance, would harm these companies. FDA has not been sympathetic to this view. In its preface to the regulation, the agency declares that "the size of a company that commercially markets ASRs will not exempt that manufacturer from compliance with appropriate CGMPs."³

Regulatory Requirements

The final rule also imposes restrictions on the distribution, use, and labeling of ASRs. ASRs are restricted devices under section 520(e) of the Federal Food, Drug, and Cosmetic Act and are subject to the restrictions set forth in the new regulation. ASRs may be sold only to IVD manufacturers, clinical laboratories regulated under the Clinical Laboratory Improvement Amendments of 1988 (CLIA) and qualified to perform high-complexity testing, and organizations that use the reagents to make tests for purposes other than providing diagnostic information to patients and practitioners.⁴

The regulation differs from the 1996 proposal by referring to laboratories regulated rather than certified under CLIA. This clarification allows ASRs to be sold to state and Department of Veterans Affairs laboratories, which are not covered by CLIA.

Labeling. The final regulation also imposes new requirements for the labeling of ASRs.⁵ For example, the labeling, advertising, and promotional materials for Class I 510(k)-exempt ASRs must include the statement: "Analyte Specific Reagent. Analytical and performance characteristics are not established." Similarly, the labeling, advertising, and promotional materials for Class II and Class III ASRs must include the statement: "Analyte Specific Reagent. Except as a component of the approved/cleared test (name of approved/cleared test), analytical and performance characteristics of this ASR are not established."

In addition, advertising and promotional materials for ASRs must include the identity and purity of the ASR, including its source and method of acquisition, and the identity of the analyte. These materials may not include any statement regarding analytical or clinical performance. These restrictions may make it more difficult to promote ASRs.

FDA believes that "except for those ASRs sold to in vitro diagnostic manufacturers, almost all ASRs will require relabeling."⁶ For this reason, the final regulation allows manufacturers and suppliers up to one year to deplete their current labeling stock before they must comply with the new labeling requirements. FDA states that all ASR manufacturers or suppliers must review their labeling, including promotional materials, to determine compliance with the new labeling requirements.⁶

In its economic analysis, FDA estimates that redesigning and reviewing the new labeling will take only four hours, and cost just $89.50 per product.⁶ This estimate is far too low. Given the number of steps necessary to draft, revise, and review labeling in accordance with the agency's quality system regulation, four hours significantly understates the work load for most companies.

In-House Use. The restrictions on ASRs are not intended to apply to products developed by laboratories for their own in-house use. FDA's preface notes that "the focus of this rule is the classification and regulation of ASRs that move in commerce, not tests developed in-house by clinical laboratories or ASRs created in-house and used exclusively by that laboratory for testing services."⁷

However, results generated by such tests will need to be accompanied by a disclaimer regarding the lack of FDA approval. The new rule requires laboratories that develop in-house tests to provide a disclaimer stating: "This test was developed and its performance characteristics determined by (laboratory name). It has not been cleared or approved by the U.S. Food and Drug Administration."⁸ The legal basis for imposing this requirement is unclear. Laboratories will need to assess the impact of this statement on reimbursement coverage.

Another restriction relates to ordering tests. In-house tests that are developed using ASRs can be ordered only by practitioners licensed by the relevant state. FDA specifically rejected direct consumer access to test results generated with ASRs.⁹

ASR Classification. Although most ASRs will fall into Class I, some will not. ASRs used in blood-banking tests classified as Class II devices, where the underlying tests have already been cleared for marketing pursuant to a 510(k), will themselves fall into Class II. These Class II blood-banking tests fall into two categories: FDA-required screens for diseases with a low potential for transmission (e.g., treponema pallidum nontreponemal test reagents, which are used in tests that aid in the diagnosis of syphilis); and tests used electively by blood banks to screen for diseases that are likely to be transmitted to subsets of blood-unit recipients known to be at greater risk of infection (e.g., certain cytomegalovirus serological reagents).

In addition, the rule identifies another subset of ASRs as Class III, and therefore subject to premarket approval (PMA) requirements. This subset consists of ASRs incorporated in tests intended to diagnose contagious diseases that are highly likely to be fatal and where accurate diagnosis offers an opportunity to mitigate the risk to the public health. Examples include ASRs used in tests to diagnose human immunodeficiency virus/acquired immune deficiency syndrome or tuberculosis. Class III ASRs also include those incorporated in Class III tests intended to establish the safety of blood and blood products, such as hepatitis assays, including genetic tests intended to ensure the safety of the blood supply.

In order to determine the substantial equivalence of Class II ASRs, or the safety and effectiveness of Class III ASRs, FDA intends to review the performance of both the ASRs and the tests of which they are components. Thus, FDA expects that most Class II and Class III ASRs will not be marketed as independent components. Under the ASR rule, there will be little incentive for companies to commercialize most Class II or Class III ASRs.

Genetic ASRs. One of the most controversial topics covered in the rule is the regulation of genetic tests. Agreeing that ASRs used in such tests should be regulated as Class I 510(k)-exempt devices, FDA states that it "does not believe there is a scientific basis to distinguish between tests based on the use of DNA and tests based on the use of other proteins or substances, or between tests based on the use of DNA and tests based on the use of other molecular diagnostic technologies."¹⁰ Accordingly, "FDA does not intend, at this time, to regulate ASRs used in genetic testing differently from other restricted Class I medical devices that are exempt from premarket notification requirements."¹¹

However, FDA adds that the issuance of these regulations does not preclude it from someday reevaluating whether additional controls may be needed for genetic testing or for the ASRs used in such tests. If further developments in the field result in significant uses of ASRs in genetic assays, the agency may reevaluate the need for additional controls.

Relation to Other Regulations. In the preamble to the final rule, FDA discusses how this regulation relates to the definition and proposed regulation for immunohistochemistry reagents and kits (IHCs) and to the compliance policy guide (CPG) on the distribution of research- and investigational-use products. FDA explains that IHCs can be marketed in a variety of ways, depending on their labeling and intended use. When an IHC is developed as a kit or system for IVD use (with a proposed intended use, indications for use, instructions for use, and performance characteristics), it would be subject to review as a Class I, II, or III device according to its intended use, as outlined in the IHC regulation proposed in June 1996.¹² When an IHC is developed and marketed as an ASR (intended for ASR use only, with no instructions for use, and no defined performance characteristics), it would be exempt from premarket notification and review, but would be subject to general controls and ASR restrictions.¹³

FDA also mentions that, in August 1992, it invited comment on a draft CPG relating to research-use only (RUO) and investigational-use only (IUO) devices. (Curiously, FDA does not mention the 1996 "final" version of the CPG.) In that 1992 draft, FDA described its enforcement policy concerning RUO and IUO IVDs that are commercialized for diagnostic or prognostic purposes. Without providing any indication of when a final CPG might be issued, FDA states that any such guide will be consistent with the ASR final rule.¹⁴

Conclusion

The ASR rule affects many companies in the diagnostic arena: suppliers of ASRs, IVD manufacturers and laboratories that purchase ASRs, and laboratories that develop their own home-brew assays. The rule also affects companies selling kits that compete with ASRs. And it offers IVD companies a new strategic option: the chance to weigh whether speedier access to market via the ASR route offers better potential for success than selling a complete diagnostic kit.

In 1996, describing their agency's proposal to regulate ASRs, FDA officials commented that it would have far-reaching impact. Looking over the final version, IVD manufacturers will have to agree they were right.

References

1. Federal Register, FR 62(225):62243—62260, November 21, 1997.

2. Code of Federal Regulations, 21 CFR 864.4020.

3. FR 62(225):62252, November 21, 1997.

4. 21 CFR 809.30(b).

5. 21 CFR 809.10(e) and 809.30(c)—(e).

6. FR 62(225):62257, November 21, 1997.

7. FR 62(225):62249, November 21, 1997.

8. 21 CFR 809.30(e).

9. 21 CFR 809.30(f).

10. FR 62(225):62247, November 21, 1997.

11. FR 62(225):62245, November 21, 1997.

12. FR 61(116):30197—30200, June 14, 1996.

13. FR 62(225):62250, November 21, 1997.

14. FR 62(225):62250—62251, November 21, 1997.

Jeffrey N. Gibbs is a partner in the law firm of Hyman, Phelps & McNamara (Washington, DC).

At FDA, the performance of microbiological IVD products has traditionally been measured against "gold standards" made up of routine biological cultures. Although a manufacturer's premarket notification (510(k)) may claim substantial equivalence to any number of predicate devices, when it comes to demonstrating performance characteristics FDA expects these to be measured relative to such gold standards.

A problem with this long-standing practice is that biological cultures are subject to the influence of a wide range of variables that make their use as gold standards somewhat problematic. To begin with, laboratories use a wide variety of growth media, any of which can affect organism recovery. Culture methods are useful only for viable organisms, the growth of which can be influenced by such parameters as the initial organism load, the strain of organism, incubation, atmospheric conditions, and antibiotics or over-the-counter preparations used by patients. Sample handling, storage, and transport can also effect culture results. Above all, the successful culturing of any organism relies heavily on the technique of the laboratorian, thus opening the door to wide variations in the quality of culture results.

For all of these reasons, it is difficult to establish definitively how an assay performs relative to a "routine" biological culture. The outcome of such an effort depends on the quality of the culture result. Moreover, advanced technologies whose performance equals or exceeds that of routine cultures present problems for both FDA and manufacturers, including how to compare the performance of such technologies against the standard, and how to present the relevant data.

Last October, Kimber Richter, deputy director of FDA's Office of Device Evaluation (ODE), announced that the agency intends to hold a panel meeting early this year to discuss its use of gold standards. It is encouraging that ODE should recognize the potential problems with culture-based gold standards in this way. Together with the agency's efforts at reengineering and modernization, hopefully we will see appropriate changes in what FDA uses as a gold standard, and how it does so.

Certainly, a part of this panel discussion should center on the technical difficulties inherent in using biological cultures as gold standards. But some part of the meeting should also focus on FDA's overall policies for the use of such standards and how the agency uses--or should use--medical practice guidelines in deciding what restrictions to require in product labeling. The panel should also discuss how far FDA should be permitted to dictate, directly or indirectly, the practitioner's use of an assay. The need for a discussion of these issues can be demonstrated by examining current FDA policy and practice regarding rapid assays for group A strep and group B strep.

In the case of strep A, FDA's policy, based on a recommendation from the American Academy of Pediatrics' Red Book, is that the results of all rapid direct-antigen assays must be backed up by comparison to culture. In adopting such a policy, the agency seems unmoved by the fact that organizations issuing such recommendations are traditionally slow to respond to advances in technology. Moreover, such guidances are typically written with patient management in mind and are not intended to be used as part of a substantial equivalence (510(k)) review.

It is problematic when FDA misunderstands the viewpoint of a guideline-writing organization, or misapplies guidelines intended to advise practitioners. One result can be agency requirements for product labeling statements that restrict practitioners' use of an assay and amount to an indirect regulation of the practice of medicine--an area FDA professes to stay out of. Similarly, the agency's blanket application of recommendations written by external groups that don't keep up with advancing technology hinders both industry and the practitioner.

FDA's policy does not allow manufacturers to claim superiority to a gold standard even when the company's data would support such a claim. Nevertheless, the agency does permit manufacturers to display data demonstrating such performance. Such a policy appears to be a mixture of not disagreeing with the data combined with retaining the sanctity of an old standard. But the policy presents a problem when new technology produces assays with sensitivity equal to or better than the standard. It allows poor performers to look good, and prevents better performers from standing out. And it is difficult to ensure uniform compliance by all manufacturers.

Medical practitioners encounter these problems in the product labeling that requires them to perform culture back-up in spite of product data that would not logically require it. Certainly, it makes little sense to universally back up a more-sensitive test by using a less-sensitive one. And recent studies looking at medical practice related to the diagnosis and treatment of strep A infection have demonstrated that patient management is not compromised when a rapid assay of adequate sensitivity and specificity is not backed up by culture.

Although FDA may intend that practitioners should exercise their own judgment about the use of such diagnostics, the labeling that the agency requires leaves little latitude for such professionals. In a litigious society such as ours, any practitioner who chooses to deviate from the instructions provided in product labeling runs a high risk of liability for his or her actions.

In 1996, FDA issued a safety alert to medical practitioners about the proper use and interpretation of direct antigen assays for group B strep in testing pregnant women and infant urine. The purpose of the alert was to warn practitioners that some of the group B strep (GBS) tests on the market had unacceptably low sensitivity, and that the agency believed more data were needed to show a correlation between antigen in infant urine and GBS treatment.

The agency subsequently informed all manufacturers of strep B tests that if they intended to keep their products on the market, they would have to submit performance data comparing the products to a new, more-sensitive standard: broth culture. Several manufacturers withdrew from the market on their own initiative, and many practitioners interpreted these actions to mean that FDA would soon withdraw all such tests from the market. In clarification, Steve Gutman, director of the ODE Division of Clinical Laboratory Devices, told industry representatives that the agency did not intend to remove products from the market as long as data relative to the new standard were provided. Pending receipt of those data, products that performed according to their stated specifications could continue to be used in accordance with their labeling.

FDA's alert tarred all rapid strep B assays with the same brush, implying that none had sufficient sensitivity to be clinically effective. Industry was never given a clear opportunity to demonstrate otherwise. Many published and soon-to-be-published studies have concluded that rapid assays of appropriate sensitivity are very useful in the management of GBS disease. Those manufacturers whose products remain on the market continue to work with FDA to define and establish what is necessary to address the agency's concerns.

Among practitioners, however, there remains considerable confusion about the intent of the safety alert. Some continue to wonder if all strep B assays will be withdrawn from the market, and many are concerned that FDA's actions have created a new and unanticipated liability dilemma. Just as in the case of strep A, FDA's strep B alert defines a standard for use that removes much of the practitioners' decision-making ability and places them in the position of using an assay in a manner inconsistent with its labeling.

And once again, as in the case of strep A, FDA's handling of strep B assays raises questions about what aspects of a product the agency should consider as part of its product clearance process. The agency should review its use of external guidelines. Contrary to the impression one would get from reading the GBS management guidelines issued by the Centers for Disease Control and Prevention, there is more than one way to effectively manage GBS disease. New studies continue to demonstrate that there are efficacious management alternatives that make use of a judicious combination of rapid assays, risk factors, and culture.

One can ask whether FDA's authority over product approvals should be permitted to cause such a cascade of problems for manufacturers and practitioners. Perhaps FDA's policy was valid for older technologies, but it now needs to be changed. The agency would do better to restrict its labeling requirements to factual matters related to a product's performance. As long as these facts are presented accurately, the agency should not dictate further the clinical use of the product. Such an approach would allow practitioners to evaluate which products provide the performance needed for the diagnosis and treatment of diseases.

With all the changes FDA is undergoing, it is heartening to hear that the agency is looking into how best to ensure realistic and uniform assessment of a product's performance. Hopefully such needed changes will be carried out in a timely fashion.

Roger Briden is director of regulatory affairs at Biostar Inc. (Boulder, CO).

Diagnostics manufacturers should be watching the Web site of FDA's Center for Devices and Radiological Health (http://www.fda.gov/cdrh) for a new list of IVD product classifications.

According to Steve Gutman, director of the Division of Clinical Laboratory Devices (DCLD) at FDA's Office of Device Evaluation, the list is expected to be posted on the Web site "early in 1998."

Addressing a Barnett International conference on "Clinical Development for IVDs" in December by speakerphone, Gutman said that DCLD had been reviewing all Class I and Class II IVDs with an eye toward reclassifying many products. That effort was helped by the November passage of the FDA Modernization Act of 1997, which exempts manufacturers from having to file premarket notifications (510(k)s) for most Class I products.

DCLD's approach is to "allow most Class I devices to be down-classified to exemptions," Gutman said. "But there will be some exemptions to the exemptions—that is, reserved products that will not be exempted from 510(k) requirements." The modernization act permits the agency to require 510(k) submissions for Class I products if they are intended for a use that is of substantial importance in preventing impairment of human health, or if they present a potential unreasonable risk of illness or injury.

Noting that reclassification could have a significant effect on how manufacturers must handle their future premarket submissions, Gutman urged the audience to look over the new list carefully and submit comments to the agency.

Ultraprecise manufacturing equipment, high-quality biochemicals and reagents, and specialized contract services for diagnostics manufacturing will be among the featured offerings of the IVD Suppliers Pavilion at Medical Design & Manufacturing West 98 Conference and Exposition, January 20— 22 at the Anaheim Convention Center (Anaheim, CA).

New products that will be on exhibit include a line of microplate washers by Tri-Continent Scientific (Grass Valley, CA), a designer and manufacturer of liquid-handling instruments and components for clinical and biotechnological uses. Kinematic Automation, Inc. (Twain Harte, CA), a designer and builder of manufacturing equipment for the test-strip industry, will feature diagnostic strip cutters, laminators for diagonal strip cutters, and dispensing systems.

Chemicon International, Inc. (Temecula, CA), a producer of bulk-order and custom antibodies and purified reagents for the R&D and manufacturing markets, will feature its line of animal sera, monoclonal and polyclonal antibodies, neurochemicals, and adhesion molecules. Precision processing equipment for test-strip manufacturing will be exhibited by BioDot, Inc. (Irvine, CA), a manufacturer of material-handling, dispensing, and processing systems used to produce rapid diagnostic tests. Medtox Diagnostics, Inc. (Burlington, NC), a contract manufacturer of custom reagent tubes and packaging supplies, will display its line of test strips and ampules for lyophilized and liquid reagents.

SurModics, Inc. (Eden Prairie, MN), will demonstrate its PhotoLink surface-modification technology, which immobilizes molecules to create unique surfaces on polymers, metals, and glass. The company will also show its StabilCoat immunoassay stabilizer for increasing the shelf-life of dried components, and its StabilZyme conjugate stabilizer for maintaining the activity of conjugates in a solution.

In addition to its IVD Suppliers Pavilion, MD&M West 98 will offer a comprehensive technical conference including a variety of presentations of interest to diagnostics manufacturers. The conference will begin January 20 with a special workshop on current regulatory trends, including a detailed analysis of FDA design control requirements. Other regulatory sessions will include a mock FDA inspection workshop, a seminar on practical considerations for compliance with FDA's quality system regulation, and a presentation on how to obtain and keep the CE mark for medical products.

Presentations on statistical tools will provide an overview of the techniques commonly used in device manufacturing, together with case studies demonstrating their usage in process development, validation, and control to meet FDA and ISO requirements. In another series of sessions, experts will discuss how small and midsized companies can do the research needed to launch a business venture in China.

The exhibit halls will be open 10:00 a.m.—5:00 p.m. on Tuesday and Wednesday, January 20 and 21, and 10:00 a.m.—3:30 p.m. on Thursday, January 22. Registration can be arranged in advance or on-site during the show. For information on attending, contact Canon Communications llc, 3340 Ocean Park Blvd., Ste. 1000, Santa Monica, CA 90405; 310/392-5509, fax 310/392-4920. Or, stop by the expo center.

Making instruments work invisibly with blood substitutes should be a primary goal for manufacturers of blood substitutes and related instrumentation, according to a panel of laboratorians.

"The goal should be to have instrumentation that can provide reliable results without interferences caused by the use of blood substitutes," said Melvin Glick, PhD, director of the toxicology section at Wishland Memorial Hospital and of the toxicology section and clinical chemistry laboratory at the Veterans Administration Hospital, Indianapolis.

Glick's comments were made last November at a Chicago-area conference on the impact of blood substitutes on clinical laboratory systems sponsored by the American Association for Clinical Chemistry (AACC).

"In order to interpret the results of current lab tests, laboratorians need to know what blood substitutes patients have received and in what concentrations," said Demetra Callas, PhD, a postdoctoral fellow at Loyola University (Chicago). "But what we're looking for is instrumentation that is not subject to interferences, or a reagent that will clear blood substitutes from patient samples so that it's not necessary to identify those samples as containing blood substitutes."

Referring to the problems caused when blood substitutes by different manufacturers are combined, Glick suggested that each manufacturer engineer its product to include a unique molecular handle. "That would enable it to be pulled out and distinguished from others." He also recommended that instrument manufacturers provide a list of tests that should not be performed in the presence of blood substitutes. "When manufacturers launch new products, they should provide the claims and specifications of their instruments for all analytes as part of their premarket submissions to FDA," Glick said.

"For every instrument, diagnostic manufacturers should provide standard interference protocols and profiles, interference correction ranges, and instructions on how to set up experiments," said John Chapman, DrPH, professor of pathology and laboratory medicine at the University of North Carolina School of Medicine and executive director of the core and clinical chemistry laboratories for the University of North Carolina Hospitals. "It's essential that manufacturers educate their laboratory customers about any problems using blood substitutes on their instruments."

"Blood substitute manufacturers also have to provide more information," said Chapman. "Labs need to have an up-front label or other mechanism for determining what kind of substitute has been used in a sample. They also need to know what the in vivo clearance rate is for every blood substitute, and how to clear such products from patient samples."

Biosensors that promise to measure blood glucose easily—and painlessly—are the holy grail of insulin-dependent diabetics. Diabetics might gain relief from the finger stabs necessary to draw blood for glucose tests and, in so doing, achieve longer and better-quality lives.

The successful developer of such a monitor stands to reap a financial windfall. The worldwide market for glucose-monitoring products exceeds $2 billion and is growing annually by double digits. More than half of this total comes from sales in the United States—and about 90% is from the disposable glucose reagent strips used in finger-stab monitoring. Painless monitors would grab an enormous chunk of this market.

There are several front-runners in this race. Cygnus, Inc. (Redwood City, CA), is developing a wrist-worn monitor called the GlucoWatch, which would use electro-osmosis to draw glucose molecules from the skin into a dermal patch for measurement. SpectRx (Norcross, GA) and Integ (St. Paul, MN) are fabricating devices that would sample interstitial fluids. The SpectRx device would use off-the-shelf glucose strip chemistry to test the fluid; Integ's LifeGuide would rely on an infrared photometer to measure glucose in the sample. Another company, Biocontrol Technologies (Pittsburgh), is developing an infrared-based device, the Diasensor 1000.

The engineering efforts by these companies have attracted the attention of several big guns in the IVD industry. Becton Dickinson (BD; Franklin Lakes, NJ) and Yamanouchi Pharmaceutical (Tokyo) are partnering with Cygnus to market the GlucoWatch—Yamanouchi in Japan and Korea, BD in the rest of the world. Meanwhile, Abbott Laboratories (Abbott Park, IL) has purchased exclusive worldwide rights to the SpectRx technology except in Singapore and the Netherlands, where the company has coexclusive rights.

The problem is that many companies are having a tough time getting their technologies to work. Biocontrol has had more than its share of problems. The company's tabletop spectrophotometer is designed to recognize a person's glucose patterns through the use of a light beam that passes through the skin of the forearm into the blood and is then reflected back to a sensor. A microprocessor is intended to interpret the data and calculate the blood glucose level.

Less than two years ago hopes were high that the company would soon get the device approved by FDA, but in February 1996 an advisory panel recommended against approval. At the meeting, the company produced successful data on only 8 patients out of the 85 enrolled in its clinical trials. The big problem was calibration. One possible reason may be that optical "signatures" of glucose levels may be as specific to patients as their fingerprints.

But according to Biocontrol spokesperson Susan Taylor, calibration was not a serious problem in a home clinical study completed late last year. "We had significantly better results from the calibration process done in the home study than in the laboratory setting," Taylor says.

The advance of such noninvasive monitors may benefit from a "sense of Congress" statement written into the FDA Modernization Act of 1997, which was signed into law in November by President Clinton. The act states that the "availability of a safe, effective, noninvasive blood glucose meter would greatly enhance the health and well-being of all people with diabetes across America and the world."

A. Paul Harding, vice president for sales and marketing at Integ, believes the statement is the result of lobbying. "That wording probably got put in at the behest of companies that are developing true noninvasive systems, a field where they're having tremendous difficulty achieving any sort of reasonable accuracy or precision."

Neither the SpectRx nor the Integ device is truly noninvasive—but both do promise to be painless. SpectRx fires a laser into the outer layer of the skin, creating micropores about 80 µm across and 20 µm deep. Integ uses a needle to penetrate the same part of the skin; there is no pain because this outer layer has very few nerve endings or capillaries.

With the SpectRx device, interstitial fluid flows into the micropore, is sampled, and is then passed to Abbott's MediSense test-strip technology, which is currently available for conventional finger-stab blood glucose testing. Integ's LifeGuide system uses light radiation rather than test strips to analyze the fluid.

Integ was hoping to be in FDA review by now but ran into problems about mid-1997. "Over the past four months we've been focusing on identifying sources of error in the system," says Harding. When the company will submit a product approval application to FDA, however, is not known. "We're basically building up a new set of instruments, and until we test them and know their performance it's impossible to predict when we'll go to FDA."

SpectRx hopes to makes its submission to FDA sometime in 1999. "We anticipate this will be a 510(k) submission, where we will just have to show substantial equivalence to predicate devices," says company spokesperson Bill Wells. "We've published data that show a very high correlation between our device and the finger-stick test."

At this point, Cygnus appears to be best positioned to commercialize its technology. The company advocates a noninvasive approach that uses low-level electrical energy to pull interstitial fluid out of the skin to a small disposable pad located between the GlucoWatch hardware and skin. The presence of glucose molecules in the fluid triggers an electrochemical reaction with a reagent in the pad, generating an electric current that is measured by an application-specific integrated circuit.

The company recently completed its first set of clinical trials and is now preparing to begin the second and final round. The studies, which involve subjects with diabetes, are being conducted by independent clinical investigators. According to Cygnus spokesperson Susan Snyder, the trials are expected to last at least through the first half of this year.—G.F.

A new technology developed at the University of Washington promises diagnostic tests with shirt-pocket portability and immediate results.

Sinclair Yee, PhD, and colleagues at UW (Seattle) have developed a light sensor that combines circuitry for spectral analysis with a kind of biochemical paint attached to the end of a probe. When dipped in a patient sample, such as blood, the paint draws target molecules into the light. "The sensor could be the size of a pencil," says Yee, a professor of electrical engineering.

Prototype of UW's surface plasmon resonance probe. Photo Courtesy Sinclair Yee

 

The electronics for the probes, which will both detect the presence and measure the concentration of specific chemicals, has been licensed to start-up company Ikonos Corp. (Portland, OR). Using proprietary biochemistry, Ikonos is developing an array of sensors. "We plan to bring the lab to the patient in the shirt pocket of the doctor," says Christophe Sevrain, company president.

The compact size of the technology means the probe is brought to the specimen, not the other way around. ER physicians and paramedics might immerse it in a patient fluid sample to test for the bacteria that cause salmonella poisoning, for example, or the enzymes associated with a heart attack.

"Imagine putting modules into a calculator-sized instrument," Sevrain says. "One could be for blood analysis. Another could be for urine analysis. They would both use Dr. Yee's technology, but the different modules would have our molecular imprinting chemistry. It's a good combination of two different technologies."

Chemical agents on the surface of the probe attract and bond target molecules. White light is selectively absorbed by the target chemicals, if they are present. The absorbed wavelength of light, called the wavelength of resonance, indicates the composition of the sample to a concentration of 10 parts per billion.

The basic technology, called surface plasmon resonance, has been around for about 15 years. The problem has been reducing its size and cost—about that of a console television priced at more than $200,000—so that a practical instrument could be made. The UW researchers have achieved that dual goal by harnessing planar geometry, which has shrunk the probe and associated processor to the size of a calculator with a likely price of less than $2000.

The probe, which is shaped like an orchestra conductor's baton, contains a glass-fiber core less than half a millimeter in diameter with planar electronics built around it. "All the integrated circuits are built along a plane," Yee notes.

The device developed at UW illuminates the sample with two beams of light. One scans for the target chemical. The other serves as a reference to record changes in the sample, such as changes of temperature, that might affect measurements. Subtracting the two signals ensures accurate results.

The technology is still in the prototype stage, but that could soon change. Sevrain expects that Ikonos will begin manufacturing a product incorporating the UW electronics and its own proprietary biochemistry by the end of this year.