|Home | About | Journals | Submit | Contact Us | Français|
Isoelectric focusing (IEF) is a technique of exquisite resolution and high sensitivity. When applied to human biological fluids using conventional protein stains it is capable of detecting down to about 100 mg of protein/L. When combined with blotting and probing techniques it can get down to less than 1 mg/L. The exquisite resolution enables a greater discrimination between the various immunoglobulin abnormalities encountered in the clinical laboratory.
In the early 1970s at the Cancer Institute, Melbourne, we observed a unique electrophoretic pattern associated with a diagnosis of non-Hodgkin’s lymphoma, Hodgkin’s disease or chronic lymphatic leukaemia. The pattern was composed of multiple, closely spaced, low level bands in the gamma globulin region.1 Immunoelectrophoresis which was in routine use in our laboratory was unable to positively identify the bands due to their close spacing and as yet immunofixation, although applied to complement2 had not yet been applied to the investigation of gammopathies. In an attempt to elucidate the problem we turned to IEF and so began a 35 year journey with this wonderful technique.
Although IEF had been performed in density gradients for a number of years, it was in 1968 that three groups independently described IEF in polyacrylamide gels.3–5 It was the method of the last group that we modified and is the basis of the technique that I still use today.6 The next major breakthrough was the introduction of IEF-grade agarose7 that brought IEF to the routine clinical laboratory.
IEF can be carried out in immobilised pH gradients or using carrier ampholytes. Carrier ampholytes are used in free solution, e.g. in density gradients or capillaries (as in capillary electrophoresis, CE), or in a gel of polyacrylamide or agarose. Rarely other media, e.g. Sephadex, are used. This paper will mainly discuss the use of carrier ampholytes in agarose gel and their use in examining immunoglobulins and related free chains and fragments.
IEF gels may be stained for protein, immune fixed or blotted on to various media and probed with specific reagents, typically antibodies. The first two methods will not be covered in detail whilst blotting will be covered more extensively.
Our original apparatus had electrodes 6.5 cm apart. The gels were poured on to supports e.g. microscope slides. The gel was inverted and laid directly on to the electrodes, gel side down. In this case any fluid that accumulates at the electrodes due to electroendosmosis merely drips off the electrodes without interfering with the focusing. A more conventional apparatus incorporates a heat sink (cooling plate), leading to better temperature control. In this instance the gel is laid on the heat sink, gel side up, and the electrodes lie directly on the gel surface. Any fluid that accumulates at the electrodes due to electroendosmosis must be removed, typically by the use of wicks.
More recently, more sophisticated apparatus has become available commercially, e.g. Helena SPIFE or Sebia Hydrasys, which can be readily adapted to IEF. One of the best aspects of these types of apparatus is the temperature control in the gel that is obtained by the use of Peltier cooling, a major contributor to maintaining reproducibility. Whilst heat sinks can help to remove heat produced during IEF they fail to maintain a set temperature to the same degree that Peltier devices are able to do. A word of caution however: the ability to remove large amounts of heat can lead to temperature gradients within the gel. Even with thin (e.g. 0.5 mm) gels, modest heat dissipation can lead to a significant difference in temperature between the surface and the base of the gel with resulting loss of resolution. Also, significant differences in conductivity (discussed later) in different regions of the gel can lead to hot spots. As the isoelectric point (pI) of the carrier ampholytes and proteins varies with temperature one can see why a constant (known) temperature must be maintained.
These sorts of apparatus may allow automation of steps such as voltage ramping and sample application with improved reproducibility over manual control.
IEF is a very “forgiving” technique, especially with respect to sample application. Because IEF is a “focusing” technique, samples can (with few exceptions) be applied anywhere on the gel and not necessarily as a narrow band.
Samples may be applied as a droplet on to the surface, into wells moulded into the gel, or through templates. They may also be applied soaked into small pieces of filter paper or similar or via applicators which are part of the commercial automated instruments. Rarely the sample may be incorporated into the gel as a whole.
The heart of isoelectric focusing is the pH gradient developed by the carrier ampholytes. Some of the properties of an ideal carrier ampholyte include good buffering capacity at the pI (|pI – pKa| < 1), good conductivity at the pI, a molecular weight (MW) between 500 and 1000 Da and solubility in common protein precipitants. It is most important that an ampholyte mixture should contain a large number and even distribution of ampholyte species spanning the desired pH range. This is necessary to give an even pH gradient, conductivity and field strength across the gel. The field strength in the gel may be measured at (say) 1 cm intervals in a gel during a focusing run and should be relatively constant. However this only shows the gross conductivity, not what is happening on a micro scale.
Figure 1A shows an electropherogram at 200 nm of a sample of Ampholine® 3.5–10 after CE at pH 10 in a borate buffer. The many peaks and troughs indicate that at the micro level there is an uneven distribution of ampholyte species. In contrast, Figure 1B shows an electropherogram of a sample of Pharmalyte® 5–8 run under similar conditions with an almost completely smooth distribution. As most human serum immunoglobulins are spread over a pI range of 4.5 to 8.5 this would appear to be a much better solution to use. In practice pH 5–8 does not quite encompass the pI range of all the serum proteins so I employ a mixture of pH ranges based on 5–8 with added 8–10.5 and some full range 3–10 to detect any high pI immunoglobulins (and fragments) plus albumin and α1-antitrypsin at the low pI end. Figure 1C shows the electropherogram of such a mixture and again although there is a number of peaks and troughs at the acidic end, the region where we expect the immunoglobulins to focus is essentially smooth. The effective pH range of this mixture is about 4–10.
Figure 2A shows a typical pattern obtained using this mixture with serum samples. On the left is a serum from a healthy individual, the two middle tracks are specimens containing a monoclonal IgG and on the right is a mixture of four different haemoglobin variants as pI markers. Note that in the normal specimen, the polyclonal IgG is completely smooth, and that the lines in the spectrotype (see below) of the two monoclonal IgGs are equally spaced. In contrast, Figure 2B shows the same normal specimen and one of the monoclonal IgGs, together with a monoclonal IgM and the same haemoglobin pI markers focused in an Ampholine-containing gel. Note the lines in the polyclonal IgG region of tracks 1 and 3, and the disturbance in the spectrotype of the monoclonal IgG in track 2. These changes (that might be interpreted as superior resolution in tracks 1 and 3) are due to the uneven distribution of ampholyte species in the Ampholine (Figure 1A).
In common with most proteins, immunoglobulins show microheterogeneity due to synthetic error, incomplete carbohydrate addition and post-synthetic modification. Such modifications include cleavage of N-terminal amino acid residues, deamidation of basic amino acids and loss of carbohydrate chains including sialic acid.8 The net effect is that a single (monoclonal) band seen on conventional electrophoresis resolves into a series of sharp lines on isoelectric focusing. The pattern of lines is known as a spectrotype.
Patterns are generally of three types:
In an oligoclonal response, one group may be present at much higher concentration and thus predominates, so much so that the pattern seen on zone electrophoresis is that of a single band. Conversely all the groups may be present at low concentration with again only one group being visible – “the tip of the iceberg”. It is important to have adequate sensitivity to detect these low level groups to ensure that an oligoclonal pattern is not reported as monoclonal.
The classical spectrotype of a monoclonal IgG is one of about six equidistant lines approximately 0.05 pH units apart. This difference in pH is similar to the expected change in pI when a molecule of 160 kDa loses an amino group from a basic amino acid.8
The classical spectrotype of a monoclonal free light chain is one major and one or two minor bands (Figure 4). In this case the distance between bands is greater than in the intact IgG because a single charge difference in a lower molecular weight molecule (about 25 kDa) gives a greater difference in pI.
IgM has limited solubility in dilute ionic strength solutions, i.e. IgM is a euglobulin. This limited solubility can result in the following behaviour:
Whatever the mechanism, intact IgM never completely focuses and does not show a spectrotype. However IgM may be converted from the 900 kDa native pentamer to a 180 kDa monomer by mild reduction. The monomer being quite soluble, focuses well and has a characteristic spectrotype similar to monoclonal IgG (Figure 3, track 4). (This same mild reduction procedure may also be used to solubilise those IgMs that precipitate at the origin on zone electrophoresis or fail to wash out on immunofixation.)
IgA, IgD and IgE contain much more carbohydrate than IgG or IgM and thus have more complicated patterns (Figure 3, track 3). Spectrotypes of up to 40 closely spaced lines may be observed for monoclonal IgA. Typically there is proportionality between the width of the band seen on zone electrophoresis and the number of lines in the spectrotype. Occasionally on zone electrophoresis a monoclonal IgA may manifest as a “multi banded” pattern. On IEF, the multiline pattern of the spectrotype tends to fluctuate in intensity but the classical pattern of closely spaced lines is maintained. Often the patterns may be simplified by treating the sample with neuraminidase typically reducing the number of lines in the spectrotype, sometimes down to a pattern similar to IgG whilst moving the spectrotype cathodally (higher pI). This same treatment may be used with zone electrophoresis to “move” a monoclonal IgA from the beta region into the gamma region to aid in identification and/or quantitation. Monoclonal IgD and IgE can be similarly treated.
There are occasions when the carbohydrate on these proteins inhibits antibodies from binding to the monoclonal proteins, leading to false negatives on immunofixation. Again treatment with neuraminidase often alleviates this problem. This is especially important in the diagnosis of heavy chain disease. The absence of a light chain reaction with a monoclonal protein is not sufficient to establish a diagnosis of heavy chain disease. These proteins being incomplete, heavy chains must be shown to have a lower molecular weight than their normal counterparts.
Heavy chain disease proteins through faulty synthesis are incomplete heavy chains that lack the ability to bind to light chains. Typically they have a characteristic spectrotype, irrespective of their heavy chain type. The pattern is characteristically one of multiple lines, up to 20 or more, regularly spaced but more widely spaced than IgA or even IgG. This is due to the lower MW of the chain. The pattern on zone electrophoresis is quite diffuse. In fact the width of a “band” on zone electrophoresis is generally related to the number and/or pI spread of the bands seen on IEF.
Post-synthetic proteolytic cleavage of immunoglobulins may occur in certain circumstances, e.g. when plasmin is produced during fibrinolysis. IgG3 is a natural substrate of plasmin and is cleaved to Fab and Fc-like fragments. If the glomerulus becomes permeable to higher MW proteins, IgG and plasmin may pass into the urine where the enzyme can cleave the immunoglobulin. This can be exacerbated if the urine is concentrated The Fc fragment has a lower pI (faster electrophoretic mobility at pH 8.6) than the parent molecule and the Fab a higher pI (lower mobility at pH 8.6, often showing post-gamma mobility). The Fc fragment reacts with anti-heavy chain sera and not with anti-light chain, whereas the Fab fragment reacts with anti-light chain and weakly or not at all with anti-heavy chain sera. Thus a monoclonal Fab fragment may be falsely interpreted as Bence Jones protein but on IEF can be distinguished by a different spectrotype (not shown). If the IgG3 is polyclonal, a sharp, “monoclonal”-like Fc band on electrophoresis and spectrotype on IEF is seen, whereas the Fab-like fragments are diffuse, albeit shifted cathodally. This Fc band/spectrotype should not be mistaken as monoclonal IgG.
It may seem trite to remind the reader that polyclonal immunoglobulin consists of a large number of monoclonal immunoglobulins.9 If enough monoclonal immunoglobulins are produced in a polyclonal response, the gamma globulin pattern on zone electrophoresis appears “smooth”. Similarly, provided “quality” carrier ampholytes are used as detailed earlier, the immunoglobulin region of the IEF plate is also smooth. On the other hand, if the number of monoclonal immunoglobulins involved is much lower, say less than 10, as can occur as a response to carbohydrate antigens and in some disease states (e.g. immunodeficiency diseases and cancers, especially those of lymphoid tissues), the pattern seen on zone electrophoresis is one of multiple sharp bands, or an irregular “shape” of the gamma region. Obviously the lesser the number of monoclonal proteins, the easier it is to observe the bands. This has been termed the oligoclonal pattern. On IEF the pattern in these patients is one of groups of lines. Even with partial overlap, distinct groups can still be recognised.
As the number of monoclonal proteins forming the response and/or the electrophoretic overlap increases, the oligoclonal pattern becomes increasingly more dif cult to recognise on zone electrophoresis. However on IEF an oligoclonal pattern can still be recognised. As the number of clonal products increases further and/or there is more overlap it becomes increasingly dif cult to recognise individual groups and the pattern progresses to one of (apparent) random lines. However it must be remembered that each of these lines is part of a group. On occasions these patterns may be simplified by immune fixing with light chain antisera when groups may be revealed. The author feels that there is a lack of appreciation that the IEF pattern in an oligoclonal response is formed from (many) monoclonal spectrotypes, especially so in the CSF literature.
The concentration of total or individual protein in urine can vary over several orders of magnitude, from g/L to μg/L in health and disease. Confining the discussion to Bence Jones protein, a laboratory should be able to detect down to 1 mg/L which is below the level of detection by zone electrophoresis. Typically using zone electrophoresis (and say a sample of 1 μL per cm track width) one can easily see a band of 1 g/L and maybe as little as 0.2 g/L using the usual protein stains. In IEF, because it focuses the protein in sharper zones, the level of detection is between 20 and 100 mg/L. This can be further decreased by applying more sample but only by increasing the salt load (unless desalted first) with possible disturbance of the pH gradient.
There are various ways of increasing the sensitivity down to the mg/L level:
When a urine showing the “light-chain ladder” is focused, it shows bands similar to those seen on zone electrophoresis with no sharp lines of the Bence Jones spectrotype. Even BJP that have been proteolytically cleaved show a series of sharp lines, quite different from the broad bands of the tubular proteinuric specimens. A problem can occur when very low levels of Bence Jones protein are superimposed over one of the bands of a “light chain ladder” but the IEF pattern is much easier to interpret than a zone electrophoresis pattern.
The pattern (spectrotype) of a monoclonal immunoglobulin is a group of regular spaced lines, much like the rungs of a ladder. The number of lines and complexity of the pattern may vary from immunoglobulin class to class, from the simplest typical monoclonal free light chain (Bence Jones protein) to the more complicated patterns of the carbohydrate-rich IgA, D and E. The oligoclonal pattern is a number of separate monoclonal spectrotypes, sometimes overlapping, sometimes showing as distinct groups. As more “monoclonal” immunoglobulins are produced, the groups overlap until finally a smooth polyclonal pattern is observed.
The advantages of IEF over zone electrophoresis are:
In biological fluids such as serum that contain g/L amounts of protein, IEF with stains such as Amido Black or Coomassie Blue is sensitive enough. Immune fixation can be used to identify abnormal bands.
When dealing with biological fluids such as urine or CSF in which much lower concentrations of protein can be expected, some method of increasing sensitivity has to be employed. This has generally involved concentration, typically up to 100-fold, often coupled with immune fixation and/or silver staining. Another promising technique is to “blot” the separated proteins on to a membrane, usually nitrocellulose or polyvinylidene fluoride (PVDF). The immobilised proteins may then be “probed” with various reagents, often antibodies conjugated with an enzyme tag.
Blotting was first applied to DNA17 and is known as Southern blotting after its originator. Later the technique was applied to RNA (“Northern blotting”) and subsequently to proteins18 (“Western blotting” or “immunoblotting”, possibly a misnomer). Methods of detection include autoradiography of 125I-labelled second antibodies and visual methods such as fluorescence-labelled or enzyme-labelled second antibodies, typically horseradish peroxidase or alkaline phosphatase visualised using appropriate substrates.
When using agarose techniques for separation of proteins, the blotting step can be a simple “squash”: a suitably moistened membrane is placed over the gel, followed by absorbent material, a glass plate and a weight. As the gel “squashes”, the fluid containing the protein passes through the membrane into the absorbent pad. The proteins bind to active sites in the membrane via a hydrophobic mechanism and are immobilised.
I use a technique (unpublished) which I term “evaporative” blotting described in detail later.
The next step is to “block” any residual active sites on the membrane using a non-specific protein (e.g. gelatine, bovine serum albumin or dried milk protein) or non-ionic detergent to ensure that subsequent protein reagents do not bind to the membrane. Blocking steps usually take between 10 minutes and an hour. Any weakly bound proteins may be competitively displaced by blocking agents and lost. They can sometimes be stabilised by chemical crosslinking to the membrane, e.g. with glutaraldehyde, prior to blocking.
After blocking, the membrane is probed with antibody. The antiserum is diluted in neutral buffer containing non-ionic detergent and carrier protein (e.g. bovine serum albumin). Antibody dilution is dependent on a number of factors including antibody titre, sensitivity required, non-specific background, incubation time, availability and cost. In a one-step technique, this primary antibody is labelled, the membrane is washed to remove any unreacted antibody and placed in substrate solution to reveal the protein.
In the two-step procedure, the primary antibody is unlabelled. After the wash step the membrane is probed with enzyme-labelled second antibody which is an antiserum directed against immunoglobulin of the species in which the primary antiserum was raised. Dilutions and incubation times are typically of the same order as for the primary antibody. Again following the incubation, the membrane is washed and placed in substrate solution to reveal the bands. A typical staining time of 10 minutes can be varied to increase or decrease sensitivity.
In Figure 4 the IEF pattern of two urines that have been concentrated 100-fold shows the presence of Bence Jones protein. Figure 5A shows these same urine concentrates which had been diluted 100-fold together with a normal urine (unconcentrated) and a 10 mg/L monoclonal IgG as a control. Following IEF the separated proteins were “evaporative blotted” on to PVDF (see later section) and probed in a two-step procedure. As can be seen, the Bence Jones bands are easily identified. Figure 5B shows one of these “urines” probed with specific antisera in the first step, confirming the BJP as a κ light chain. The sensitivity of the method is less than 1 mg/L per band.
Whether one uses a one-step or two-step procedure depends on the circumstances. The one-step procedure is probably the method of choice when screening native (unconcentrated) fluids, e.g. CSF for oligoclonal bands using an enzyme-labelled anti-human IgG. However when the sample needs to probed with a number of specific antisera, it is probably more appropriate to use unlabelled primary antisera followed by an enzyme-labelled secondary antiserum. In this way a large number of specific antisera may be accommodated at a lesser cost.
Nitrocellulose membrane used originally18 and in a number of procedures today is somewhat brittle and easily damaged. PVDF is much more pliable and easy to handle. It is extremely hydrophobic, exemplified by the fact that it will not “wet” when placed in aqueous solutions. It has to be first placed in methanol, subsequently into water and then into buffer solutions. Once a gel has been blotted on to it, the PVDF membrane may be dried making it again hydrophobic, but the areas where protein has adsorbed remain hydrophilic and wettable. In this condition the membrane may be probed “dry”. Due to the hydrophobic nature of the membrane “at large”, the “active sites” are not available to proteins in the probing solution so the membrane does not need to be blocked, as it does for nitrocellulose, and the amount of washing is decreased. For this “dry probe” procedure to work adequately, the membrane must be completely dry.
This property led to the “evaporative blot” procedure mentioned earlier. In this instance following electrophoresis or IEF, the PVDF membrane is placed over the gel as in the “squash blot” technique, but the adsorbent sheets are omitted. The gel covered by the membrane is then dried at 50 °C. As the fluid evaporates through the membrane it carries the separated proteins with it allowing them to bind to the “active sites”. Once evaporation is complete, the membrane is rinsed in water to release it from the gel, dried and is now suitable for “dry probing”. The patterns shown in Figures 5A and 5B were processed by this method.
The use of blotting is not restricted to increasing the sensitivity for dilute protein solutions. A variation of the technique is filter affinity transfer,19–21 in which mono or polyspecific antibodies are initially adsorbed on to the membrane and any unreacted sites blocked. The blotting procedure is then carried out in the normal way but in this instance only those proteins that are complementary to the antibody are bound to the membrane. The membrane is then probed with various antisera. This technique has many applications but has particular use in light chain typing of low levels of monoclonal immunoglobulin in the presence of large amounts of immunoglobulin of a different class, e.g. low levels of monoclonal IgD in the presence of normal levels of polyclonal IgG.20
In comparison with immunofixation, blotting and probing can involve fewer steps and be made less labour intensive, results can be available in a shorter time (as little as 30 minutes), and the cost of antisera can be less as they are used at high dilution.
IEF is an indispensable technique in the investigation of immunoglobulin abnormalities. It is simple to perform and with experience the patterns are easily interpreted. It can be combined with immune fixation to confirm immunoglobulin identity and to increase sensitivity of detection. Combined with blotting and probing, a sensitivity of less than 1 mg/L can be achieved. In terms of resolution IEF is superior to zone electrophoresis and is the method of choice for determining the clonality of immunoglobulins.
Competing Interests: Iso Laboratories Pty Ltd is a manufacturer of IEF gels and equipment. Mr Cornell is a consultant to Helena Laboratories (Aust) Pty Ltd.