The first installment of this article (IVD Technology, March/April 1999) covered the basic principles behind protein binding.¹ This second installment discusses common problems and the most common solutions. For clarity, the problems have been broken down into six generic groups (see Table I). Because IVD technology is wide and varied, there are no universal answers. No amount of discussion can replace experimental results. The purpose of this article is, therefore, to suggest starting points for experiments.

Table I. Common defects associated with nitrocellulose membrane–based test development and the most likely causes.The decrease in nitrocellulose surface area available for protein immobilization can be seen as the membrane pore size increases from 0.45 (a) µm through 3.0 µm (b) to 8.0 µm (c). Increased pore size leads to a faster wicking rate, but also means that the protein being applied travels a greater distance before it comes into contact with the nitrocellulose membrane wall. The capture line will thus be wider on an 8-µm membrane than on a 0.45-µm membrane, because of both the increased lateral wicking rate and the increased distance traveled by the protein before immobilization.

Nonspecific Signal

Nonspecific signal in nitrocellulose-based assays is a significant problem for the diagnostic industry. It can appear either generally over the entire membrane surface or, more seriously, at the capture line.

Causes. Nitrocellulose will bind proteins, which is the mechanism behind the capture line. The reasons for this binding were discussed in Part 1.

Nonspecific protein binding. Nonspecific protein binding can be caused by an unwanted interaction between the sample or conjugate and the capture line. Such interactions at the capture line can be the result of any of the common causes of protein attachment, including charge attraction, hydrophobicity, disulfide bridging, or a genuine nonspecific immunogenic binding.

Conjugate binding. Since conjugate particles are covered by a protein layer, the causes of nonspecific conjugate binding are similar to the causes of nonspecific protein binding. Additional problems can occur. The conjugate particles themselves are often permanently charged. In the case of gold conjugate, the conjugate is very susceptible to interaction with sulfhydryl moieties on the surfaces of proteins and membranes. The causes of and remedies for the interaction of conjugates are discussed in literature available from conjugate manufacturers.^2–4

Solutions. Nonspecific attachment to the membrane can normally be reduced by blocking with a protein (e.g., bovine serum albumin [BSA]), surfactant (Tween 20, Triton X-100, or sodium dodecyl sulfate [SDS]). The effect of charge can be overcome by either changing the pH of the test or increasing the ionic strength of the system.

Disulfide bridging is rarely seen in diagnostic products, although nonspecific immunogenic interactions do regularly occur. Standard immunologic blocking techniques, such as use of a similar protein as a blocking reagent, can solve this problem. For example, if a nonspecific interaction occurs between a specific mouse IgG and a human sample, blocking the sample with mouse serum would remove the false-positive signal.^5–7

Weak or Diffuse Capture-Line Intensity

If the level of capture reagent per unit area of membrane surface is too low, the capture line may be weak or diffuse. The problem may be simply mechanical; for instance, too-rapid movement of the membrane through the application striping system. However, the cause may be more complex.

Causes. It is often impossible to determine the cause of this problem by merely examining the test strip. The only solution is to work through the potential causes and find by experiment where the major problem lies.

Concentration of the capture reagent in the application solution is too low. The application solution diffuses away from the point of application. The capture reagent spreads with the application solution to an extent defined by the partition coefficient between the solid phase and the solute. If attachment to the solid phase is preferred, the capture reagent will spread less. If solubilization in the solute is preferential, then the capture reagent will spread further (see Figure 1).

Figure 1. A weak capture line indicates that the amount of protein bound to the membrane is too low.

Figure 2. A diffuse capture line can result when the capture reagent is washed away by the passage of analyte proteins and surfactant solutions.

 

Capture reagent is washed off the membrane by the sample. If the strength of attachment between the capture reagent and the membrane is too low, or the surfactant level in the running buffer is too high, the capture reagent may be physically removed from the membrane surface by the sample (see Figure 2).

The lateral wicking rate of the membrane is too high, causing spread of the capture reagent after application. If the membrane used has a very high lateral wicking rate, the applied protein will diffuse very rapidly from the point of application. As in all physical interactions, there is a time-related function—the faster the protein is moving in a lateral direction, the wider the capture line will be. The protein solution will also penetrate vertically into the membrane, and most of the protein that penetrates will be wasted, as only the material trapped in the top 10 µm of the nitrocellulose membrane will be visible. This combination of lateral and vertical wicking can cause a weak capture line. The apparent concentration of the analyte in solution is a function of the lateral wicking rate of the test (apparent concentration 1/wicking rate).² This equation means that as the wicking rate increases, the apparent concentration of the analyte drops dramatically.⁸

A reduction in the lateral wicking rate (by use of a smaller-pore nitrocellulose or a viscosity modifier) results in a stronger test line. However, it may cause an increase in nonspecific signal, a change in the test sensitivity cut-off, or a significant increase in test time.

The capture reagent binds only weakly to the membrane. The binding of proteins to membranes can be influenced by optimization of the application buffer. However, some capture reagents are difficult to attach to membranes, either because of their size or their surface properties. In these cases where the binding is very weak, there are only a limited number of solutions. The most common is cross-linking the required reagent to a carrier protein.

The affinity constant between the capture reagent and the target analyte is too low to support efficient capture. The use of low-affinity antibodies is, fortunately, normally avoidable. Careful screening of antibodies during development is an absolute necessity. In cases where the use of a low-affinity antibody is the only option, then the use of a small-pore membrane coupled with a low lateral wicking rate will maximize the interaction between the sample and the analyte.

Solutions. Whatever the cause of weak or diffuse capture-line intensity, the solution lies in achieving a higher concentration of capture reagent at the desired point.

Use a smaller-pore membrane. The use of a smaller-pore membrane will increase line sharpness and intensity for two reasons. First, the smaller the pore, the greater the surface area of the material, and the greater the surface area, the higher the concentration of capture reagent that attaches to it. Second, the smaller the pore, the slower the wicking rate.

Use a different membrane. Different membranes have different binding characteristics for different capture materials. Selecting a membrane that has better binding characteristics for the particular capture reagent chosen improves the capture line appearance.

Use a higher protein concentration. Increasing the concentration of the capture reagent in the application buffer may cause a higher concentration of the capture reagent to attach to the membrane. This would allow a higher level of analyte to stick in the capture zone and hence improve signal intensity.

Change the application buffer. Modification of the application buffer adjusts the point of equilibrium between the amount of capture reagent attached to the membrane and the amount remaining in solution. Optimization of the buffer following the principles outlined in Part 1 of this article gives the maximum adsorption of the capture reagent to the membrane.¹

Reduce the lateral wicking rate. Slowing the lateral wicking rate by the introduction of a viscosity modifier or by the use of a smaller-pore membrane effectively increases the concentration of the analyte in solution and hence improves the appearance of the capture line.

Cross-link the capture reagents. If the molecular weight of the applied protein is low (or if there are unfavorable surface properties), the use of a cross-linking agent can significantly enhance the level of protein binding observed. The cross-linking itself can either precede or follow application of the sample to the membrane.

The most common method is to cross-link the capture reagent to a carrier protein that will not cause any nonspecific interactions prior to application. The proteins typically used for this purpose are BSA or keyhole limpet haemocyanin (KLH). It is possible to use a cross-linking agent after capture reagent application (e.g., a glutaraldehyde wash step). However, the developer should ensure that all active groups introduced to the membrane are efficiently blocked before performing the test (e.g., in the case of glutaraldehyde, an ammonium sulfate wash blocks any free aldehyde groups). More specific examples of cross-linking chemistries can be found in the literature or in suppliers' catalogs.^10–12

Uneven Capture-Line Wetting

The uneven rewetting of the capture line can have very serious consequences for the developers of rapid diagnostic assays. If the membrane rewets unevenly, the capture line will be seen to be striped. In the worst case, it is possible for the capture line to be more hydrophobic than the surrounding membrane (normally due to the removal of membrane surfactant in washes). In this case it is possible for "submarining" to occur—this is when the sample runs through the membrane until it reaches the capture line, the higher hydrophobicity of the capture line effectively stopping lateral flow. The sample may then run along the plastic support of the membrane (which is more hydrophilic), and then reenter the membrane above the capture line where the membrane is more hydrophilic. This can result in a capture line that has no or very little sample penetration, with obvious problems for test sensitivity and selectivity.

Causes. Capture-line wetting may be uneven because of a defect in the membrane itself or because of a flaw in the test-strip manufacturing process.

Membrane drying is uneven. The rewetting of a nitrocellulose membrane is usually dependent upon the degree of drying the membrane has undergone. If the relative rates of drying vary across a membrane due to variability of the drying conditions to which the membrane was exposed during manufacture, the rate at which the membrane rehydrates will vary across the membrane.

Table II. The efficiency of different blocking techniques.

The application of an aqueous sample can also affect the distribution by washing any water-soluble residues away from the point of application. Any variation in the character or concentration of these residues will affect the rewetting rate of the membrane.

Hydrophobicity of the membrane is uneven. Perhaps the most significant factor for uneven capture-line wetting is hydrophobicity variation in the membrane. The rate of membrane rehydration is strongly influenced by the presence of hydrophobic or hydrophilic residues on the membrane surface. These residues can be introduced by membrane posttreatments (e.g., the introduction of a rewetting agent), hydrophilic materials added during manufacture, or additives in the striping buffer. The distribution of these hydrophilic materials is a factor in the evenness of the initial application and any subsequent migration of the hydrophilic materials through the membrane during storage.

Membrane pore structure is inconsistent. The pore structure of nitrocellulose membranes is a function of the parameters in the casting machine during the casting process. Uneven airflow within the casting machine may cause a variation within the membrane. As nitrocellulose-casting machines normally use laminar airflow, uneven airflow is typically seen as a variation across the width of the machine. This can be seen as a variation in performance between adjacent rolls cut from the larger master roll.

The developer cannot solve this problem without performing 100% quality control checks on incoming materials. The best compromise is to ensure that across-machine variation is evaluated adequately during the development process.

Solutions. As mentioned above, if a defect in the membrane is the cause of uneven rewetting, the only recourse is more-rigorous inspection of raw materials. If the problem is a result of the test manufacturing technique, however, there are several possible solutions.

Change the application buffer. Introduction of a mild surfactant to the striping buffer ensures that the capture line is in an evenly hydrophilic environment. This encourages even rewetting of the capture line. The choice of surfactant and its concentration is critical; effective results can be obtained using a low-concentration (~0.1%) SDS or sodium dodecylbenzolylsulfonate solution.

Perform a membrane-blocking step. The most common way to ensure even wetting of the capture line is to use a blocking technique. Blocking the membrane with a material that promotes rewetting of the membrane ensures rapid and even membrane rewetting. The effect of these blocking agents has been evaluated in product support literature.^13,14 The developer should investigate a range of blocking agents to find the most efficient for any particular test.

The method used for membrane blocking is also a significant factor in the success or failure of the blocking step. There are three points where a blocking agent can be applied (see Table II). The choice of method for inclusion of the blocking agent depends on the efficiency of the blocking step coupled with the cost of achieving the solution. As a compromise, therefore, the second technique (inclusion of the blocking agent in the conjugate pad or in a sample application pad) is often chosen. While it is less efficient than blocking the entire membrane, the procedure is operationally simple.

Change the membrane. A membrane with a surfactant posttreatment is more likely to show uneven capture line rewetting than a membrane without one. Application of the capture reagent solution may wash the surfactant away from the membrane surface. The removal of the additional rewetting agent can significantly affect the rewetting properties of the capture line. When no additional rewetting agent is added to the membrane surface, the line rewetting is likely to be consistent. However, the overall speed of rewetting may be slow.

Capture Line Too Thick

If the capture line is too thick, test results may be difficult to interpret (see Figure 3). The capture line may well show significant intensity variation across its width, probably with an intense front and back edge. For an inexperienced user the potential variation across the width may be confusing.

Figure 3. Problems with protein binding are typically visible in the capture line of an assay's test result, as in these examples.

 

Causes. At first glance the cause of too thick a capture line may seem obvious: too thick a line is being applied. In reality, this is only one of many possible causes.

The capture line is spreading too far after its application. If the protein being applied favors remaining in solution rather than attaching to the solid phase, it is possible that the protein molecules will move with the solvent front of the application buffer. In such a case the capture line may be extremely wide or have sharp edges with a relatively diffuse middle portion. The latter effect (colloquially known as a "coffee ring effect") is caused by migration of the protein with the solvent and increasing concentration of the protein as the solvent evaporates.

The capture reagent is washed away when the sample is applied. If the physical attachment of the protein to the membrane is too weak, or if a surfactant is present in the system, the capture line itself can be washed away as the sample wicks up the membrane. The displaced protein can then reattach to the nitrocellulose above the capture line.

Too much protein is applied. If we assume that the protein saturates the available nitrocellulose, the effect of excess protein levels will be the spread of the protein capture line.

The application aperture is set too wide. The protein will normally attach to the membrane at the point of application. If the settings are such that the sample is applied over a wide area, then the capture line will be wide.

The membrane is dried insufficiently after capture line application. If the membrane is dried insufficiently, the capture reagent will not be efficiently immobilized.¹ The application of the test sample may wash the capture reagent away from the point of application. The capture line may therefore become significantly wider due to the spreading of the capture reagent.

Solutions. The possible means of achieving a line of the desired thickness involve varying many of the parameters already discussed. Changes in the membrane, the buffer, the blocking agent, or the manufacturing conditions may be appropriate.

Change the application buffer. Optimization of the application buffer, both to minimize protein stability in solution and to maximize the viscosity of the application buffer, will produce the sharpest capture line on the membrane. The slower the protein solution flows, the greater the chance that the protein will bind close to the point of application (see Figure 4).

Figure 4. Varied results from capture lines of 1mg/ml mouse IgG applied using different buffers: (a) 10 mmol phosphate, pH 7.2; (b) 10 mmol phosphate + 3% methanol, pH 7.2; (c) 10 mmol phosphate + 150 mmol NaCl + 3% methanol, pH 7.2; (d) 50 mmol phosphate + 150 mmol NaCl + 1% BSA, pH 7.2; (e) 50 mmol phosphate + 150 mmol NaCl, pH 7.2; (f) 50 mmol phosphate + 150 mmol NaCl, pH 6.0. All samples were detected by a 40 nmol gold-conjugated goat antimouse IgG antibody.

 

Use a smaller-pore membrane. Smaller-pore membranes have a greater surface area of nitrocellulose per unit area of membrane, and the wicking away from the point of application is therefore slower than in larger-pore membranes. The combination of these effects means that the protein line is relatively narrow.

Change the membrane. Selecting a membrane with higher binding capacity will improve line sharpness.

Change the membrane blocking conditions. The blocking agents used are normally chosen because they interfere with protein binding. If the blocking materials are present in high quantities, the capture line binding may experience some interference effects. Lowering the concentrations of the blocking agents and investigating alternatives may achieve a better result.

Use less protein. Reducing the protein content of the system will reduce the area of nitrocellulose occupied by the capture reagent. The protein binds to the first available unoccupied surface of nitrocellulose.

Set application aperture more narrowly. Applying the protein solution from a narrower aperture will initially apply the protein to a smaller area of membrane, encouraging the formation of a thinner line.

Increase drying of the membrane. The strength of protein binding to nitrocellulose has long been linked to drying. Drying the membrane more vigorously after the protein reagent has been applied may well therefore reduce the chance of the capture protein being washed away when the sample is applied.

Capture Line Too Thin

A capture line that is too thin may give a false-negative test result. This case may well be true for cases where there is significant proportion of eye disease in the target market, this would be especially prevalent where the target market is the elderly of many third world areas. A line that is very thin would be very difficult to read accurately. Perhaps the optimal line width would be in the region of 0.8 to 1 mm wide. If the capture line is of the order of 0.2-mm wide then reading even a strong positive result can be difficult.

Causes. The causes of too thin a capture line are mainly the opposites of those mentioned above.

Protein binds too rapidly after application. Binding of the protein to the membrane immediately after application results in a capture line that is very narrow, although the application-buffer front will be significantly wider than the capture line itself.

Protein in the application system is insufficient. If the protein shows good binding properties to the membrane surface, a low concentration of protein in the capture reagent gives narrow lines following application. All the protein will bind in a very small area before it has had a chance to diffuse away from the application point.

The application aperture setting is too narrow. With optimal protein binding, the protein binds immediately to the area to which it is applied and does not spread. If the capture reagent is applied over a very narrow area, the protein is unlikely to move away from the application point before binding.