|
|
||||||||
Review |
Department of Chemistry, 738 Hamilton Hall, University of Nebraska, Lincoln, NE 68588-0304. Fax 402-472-9402; e-mail dhage{at}unlinfo.unl.edu
| Abstract |
|---|
|
|
|---|
| Introduction |
|---|
|
|
|---|
According to the International Union of Pure and Applied Chemistry (1), affinity chromatography is defined as a liquid chromatographic technique that makes use of a "biological interaction" for the separation and analysis of specific analytes within a sample. Examples of these interactions include the binding of an enzyme with an inhibitor or of an antibody with an antigen. Such binding processes are used in affinity chromatography by first obtaining a binding agent, known as the "affinity ligand", that selectivity interacts with the desired analyte and then placing this ligand onto a solid support within a column. [See Refs. (2)(3) for reviews of supports and immobilization methods that can be used in making affinity columns.] Once this immobilized ligand has been prepared, it can be used for isolation or quantification of the analyte.
The immobilized ligand is the key factor that determines the success of any affinity chromatographic method. As implied by the definition given earlier for affinity chromatography, most of these ligands are of biological origin; however, the term "affinity chromatography" has also been used throughout the years to describe some columns that contain selective ligands of nonbiological origin. Examples of these nonbiological ligands are boronates, immobilized metal ion complexes, and synthetic dyes (e.g., triazine-related compounds). Terms such as "bioaffinity chromatography" and "biospecific adsorption" are occasionally used to specify whether the affinity ligand is really a biological compound. Regardless of the origin of the ligand, the type of ligand can be used to divide affinity techniques into various subcategories, such as lectin, immunoaffinity, dye ligand, and immobilized metal ion affinity chromatography, to name a few (2)(3). These and other affinity techniques will be examined in more detail later.
Another factor that can be used to distinguish between one affinity method and another is the type of support used within the column. In "low-performance (or column) affinity chromatography", the support usually is a large diameter, nonrigid gel, such as agarose, dextran, or cellulose. In "high-performance affinity chromatography" (HPAC),1 the support consists of small, rigid particles based on silica or synthetic polymers that are capable of withstanding the flow rates and/or pressures that are characteristic of HPLC systems (2)(4). Both low- and high-performance methods have been used in clinical methods. Low-performance affinity chromatography commonly is used for sample extraction and pretreatment because it is relatively easy to set up and inexpensive to use. However, the better flow and pressure stability of high-performance supports makes HPAC easier to incorporate into instrumental systems, which in turn gives it better speed and precision for the automated quantification of analytes.
| Direct Analyte Detection by Affinity Chromatography |
|---|
|
|
|---|
|
In addition to its simplicity, there are several other advantages to
using the direct detection mode of affinity chromatography. For
example, when this mode is performed on an HPLC system, the precision
is generally in the range of 15% and the run times are often as low
as 56 min per sample (for an example, see Fig. 1
)
(2)(4)(5). The greater speed of
these systems compared with many other ligand-based techniques (e.g.,
traditional immunoassays) largely can be attributed to the better mass
transfer properties and increased analyte-ligand binding rates that are
produced by the supports used in affinity columns. The precision of
this approach is partly the result of the reproducible sample volumes,
flow rates, and column residence times that are possible with modern
HPLC equipment. Another factor that leads to the good precision in
HPLC-based affinity methods is the reduced batch-to-batch variation,
which is the result of using the same ligand for the analysis of
multiple samples and calibrators. It has been reported in many studies
that several hundred injections can be performed on the same affinity
column, provided that proper elution and regeneration conditions have
been selected. In some cases, there have been reports were affinity
columns have been used for >1000 injections with no serious
signs of degradation (2)(5)(6).
One limitation of the direct detection format in affinity chromatography is that this requires the presence of enough analyte to allow the measurement of this substance as it elutes from the affinity column; in HPLC-based systems this is usually performed by on-line ultraviolet/visible absorbance or fluorescence detectors. Such a requirement tends to make the direct detection mode most useful when dealing with intermediate-to-high concentration solutes in clinical samples. However, it is also possible to use direct detection with trace sample components if the affinity column is combined with precolumn sample derivatization and/or more sensitive detection schemes, such as an off-line immunoassay or a suitable postcolumn reactor (5).
A second potential limitation of the direct detection mode is that samples and calibrators are analyzed sequentially by the affinity column rather than in batch mode. This makes the direct detection format most valuable in situations where low-to-moderate numbers of samples are being processed and/or fast turnaround times per sample are desired. It should be noted, however, that sequential analysis has the advantage of making affinity chromatography easier to troubleshoot than batch-mode techniques and easier to determine whether the assay is operating satisfactorily before patient samples are tested.
boronate affinity chromatography
Affinity methods that use boronic acid or boronates as ligands are
one group of chromatographic techniques that have been used
successfully with clinical samples. This group of methods, known
collectively as "boronate affinity chromatography", includes one of
the earliest reported quantitative applications of affinity
chromatography in the clinical laboratorynamely, the determination of
glycohemoglobin for the assessment of long-term diabetes management
(Fig. 2
) (7)(8)(9)(10)(11)(12)(13)(14)(15). At a pH above 8, most boronate derivatives
form covalent bonds with compounds that contain cis-diol groups in
their structure. Because sugars such as glucose possess cis-diol
groups, boronates are valuable for resolving glycoproteins (e.g.,
glycohemoglobin) from non-glycoproteins (e.g., normal hemoglobin).
|
The first use of a boronate affinity column for the determination of glycohemoglobin was by Mallia et al. (8) in 1981, where a low-performance agarose gel was used as the support and absorbance detection at 414 nm was used to quantify the retained and nonretained hemoglobin fractions in human hemolysate samples. Elution was performed by passing through the column a soluble diol-containing agent (i.e., sorbitol) that displaced the retained glycohemoglobin from the column; alternatively, a decrease in mobile phase pH could also be used for elution (9). After the initial report by Mallia et al., similar low-performance methods were reported or evaluated by other groups (10)(11)(12)(13). The same approach has since been adapted for use in HPAC and HPLC-based systems (9)(11)(14)(15).
In addition to hemoglobin, it is possible to use boronate columns to look at other types of glycoproteins in samples. For example, by monitoring the absorbance at 280 nm instead of 410415 nm, the technique used for glycohemoglobin can easily be modified to determine the relative amount of all glycated proteins in a sample (14). Alternatively, a particular type of glycoprotein can be examined by combining a boronic acid column with a detection method that is specific for the protein of interest, such as is done by using absorbance measurements at 410415 nm for the quantification of glycohemoglobin. Examples of this later approach include the use of boronic acid columns followed by an immunoassay for the detection of glycated albumin in serum and urine (16) or for the determination of glycated apolipoprotein B in serum (17).
lectin affinity chromatography
Lectins are another class of ligands that have been used for the
direct detection of clinical analytes by affinity chromatography. The
lectins are non-immune system proteins that have the ability to
recognize and bind certain types of carbohydrate residues
(18). Two lectins that are often placed into affinity
columns are concanavalin A, which binds to
-D-mannose and
-D-glucose residues, and wheat germ
agglutinin, which binds to
D-N-acetylglucosamine. Other lectins
that can be used are jacalin and lectins found in peas, peanuts, or
soybeans. These ligands commonly are used in the isolation of many
carbohydrate-containing compounds, such as polysaccharides,
glycoproteins, and glycolipids (2)(3).
One clinical application of lectin affinity chromatography has been in
the separation and analysis of isoenzymes. This is illustrated in Fig. 3
, where an HPLC column containing immobilized wheat germ
agglutinin was used to distinguish between the liver- and bone-derived
isoenzymes of alkaline phosphatase in human serum (19). This
method showed improved resolution of the isoenzymes vs a
low-performance affinity column (20) and gave good
correlation for a variety of patient samples when compared with a
solid-phase immunoassay for alkaline phosphatase (21).
|
A variety of other glycoproteins also have been studied and quantified by the use of lectin affinity columns. For example, low-performance columns based on concanavalin A have been used to separate apolipoprotein A- and apolipoprotein B-containing lipoproteins in human plasma (22), to study the microheterogeneity of serum transferrin during alcoholic liver disease (23), and to characterize the carbohydrate structure of follicle-stimulating hormone and luteinizing hormone under various clinical conditions (24). A combination of concanavalin A and wheat germ agglutinin columns also has been used to identify changes that occur in asparagine-linked sugars on human prostatic acid phosphatase during prostate cancer (25).
protein a or protein g affinity chromatography
A third class of ligands that have been used in direct analyte
detection by affinity chromatography are antibody-binding proteins such
as protein A and protein G. These are bacterial cell wall proteins
produced by Staphylococcus aureus and group G
streptococci, respectively (26)(27)(28). These
ligands have the ability to bind to the constant region of many types
of immunoglobulins. Protein A and protein G bind most strongly to
immunoglobulins at or near neutral pH, but readily dissociate from
these solutes when placed in a buffer with a lower pH. These two
ligands differ in their ability to bind to antibodies from different
species and classes (3)(26)(29); for
example, human IgG3 binds much more strongly to
protein G than protein A, and human IgM shows no binding to protein G
but does interact weakly with protein A (3). A recombinant
protein known as protein A/G, which blends the activities of these
ligands, also is available for use in affinity columns
(3)(30).
The ability of protein A and protein G to bind to antibodies make these
good ligands for the analysis of immunoglobulins, especially
IgG-class antibodies, in humans. The first clinical uses of
these ligands in an HPLC system were methods based on immobilized
protein A for the analysis of IgG in serum samples
(31)(32). A similar method for the determination
of IgG in serum has been developed based on high-performance protein G
columns (33). Yet another study used a combination of two
affinity columns, one containing immobilized protein A and the other
containing anti-human serum albumin (HSA) antibodies, for the
simultaneous analysis of IgG and albumin in serum for the determination
of albumin/IgG ratios (Fig. 4
) (34). An additional application of protein A and
protein G has been as secondary ligands for the adsorption of
antibodies onto supports to be used in immunoaffinity chromatography,
as discussed in the following section. This particular method can be
used when high antibody activities are needed or if it is desirable to
replace the antibodies in the affinity column frequently
(5)(6)(35).
|
immunoaffinity chromatography
Of all the types of affinity chromatography, those that use
antibodies or antibody fragments as ligands make up the largest and
most diverse group of affinity methods in clinical testing. This is a
combined result of the specificity of antibodies and the relative ease
with which they can be obtained to a wide variety of analytes. The term
"immunoaffinity chromatography" (IAC) is used for an affinity
chromatographic method in which the stationary phase consists of an
antibody or antibody-related reagent (5)(35).
When such a technique is performed as part of an HPLC system, the
resulting method can be referred to as "high-performance
immunoaffinity chromatography" (HPIAC)
(5)(35).
Several examples of direct analyte detection by HPLC-based IAC are
described in Ref. (5). Some clinical applications that have
been reported include methods developed for anti-idiotypic antibodies
(36)(37), glucose-containing tetrasaccharides
(38)(39), granulocyte colony-stimulating factor
(40), HSA (34)(41), IgG
(42), immunoglobulin E (43), interferon
(44)(45), tumor necrosis factor-
(45), interleukins (45)(46),
ß2-microglobulin (47), and
transferrin (48). One such example (i.e., the determination
of fibrinogen in human plasma) is illustrated in Fig. 1
(49). In this particular case, the amount of fibrinogen in
the retained peak was determined by the measurement of its absorbance
at 280 nm. The sample was a 20-µL aliquot of plasma diluted 1:10. The
retained peak appeared at 6 min, and the time between sample injections
was 15 min, which included 9 min for column reequilibration
(49).
Both large and small analytes can be determined by the use of direct
detection in IAC. Furthermore, it is possible to utilize immunoaffinity
columns either separately or in combination with other affinity
columns. This was demonstrated in the previous section in the
discussion of the dual-column immunoaffinity/protein A method for the
analysis of HSA and IgG in serum (34). A similar approach
recently has been used with fluorescent labeled samples and up to 10
separate immunoaffinity columns connected in series for the
simultaneous determination of various cytokines in clinical samples
(Fig. 5
) (45). In theory, this multicolumn format could be
used with even larger numbers of compounds and with other combinations
of analytes. The fact that a single sample aliquot is required for all
of the columns makes this technique appealing in situations where only
a limited amount of a patient's sample may be available for analysis.
|
When used as part of an HPLC system, the direct detection of analytes as they elute from immunoaffinity columns usually involves monitoring their ultraviolet/visible absorbance. However, special methods for the detection of low-concentration analytes have also been devised that use precolumn derivatization to place fluorescent tags (40)(45)(46) or radiolabels (39) onto sample solutes before injection. Alternatively, the column eluate can be collected in fractions and later analyzed by an immunoassay (43)(47) or receptor assay (46) that is specific for the species of interest. In addition, specialized methods can be combined with IAC to monitor compounds that elute in the nonretained fraction of the sample. This latter approach was used recently in a method that combined an HPIAC column and flow injection analysis for the determination of urinary albumin/creatinine ratios. This technique used an anti-albumin immunoaffinity column for the capture and detection of HSA and a Jaffé-based colorimetric reactor for the quantification of creatinine in the portion of the sample that was not bound by the antibody column (30). This is yet another approach that could be useful when only small amounts of sample are available and information on several clinical analytes is desired.
columns based on miscellaneous ligands
In addition to the ligands that have been mentioned already,
several other types of ligands have also been used for the direct
detection of clinical analytes by affinity chromatography. For example,
an immobilized heparin column has been used for the determination of
antithrombin III in human plasma (50)(51).
S-Octylglutathione has been reported as a ligand for the
separation and analysis of glutathione S-transferase
isoenzymes in human lung and liver samples
(52)(53). And finally, immobilized
p-aminobenzamidine has been used for the separation of human
plasminogen species, with the addition of an immobilized urokinase
column for on-line detection (54).
| Chromatographic Immunoassays |
|---|
|
|
|---|
Many of the same labels that have been used in traditional immunoassays have also been used within chromatographic-based immunoassays. For example, enzyme labels such as horseradish peroxidase, alkaline phosphatase, and glucose oxidase have all been used in such methods. Other labels that have been reported include fluorescent tags such as fluorescein, Texas red, or lucifer yellow; chemiluminescent labels based on acridinium esters; and liposomes impregnated with fluorescent dye molecules (5). The detection of these labels generally is performed on line as they elute in the nonretained or retained peaks of the immunoaffinity column; however, fraction collection and off-line detection can also be used when required.
competitive binding immunoassays
There are several different methods for performing chromatographic
immunoassays, but the most common method uses a competitive binding
format. The easiest approach is to mix the sample with a labeled
analyte analog (i.e., the label) and to inject the mixture
simultaneously onto an immunoaffinity column that contains a relatively
small amount of antibody. This format, known as a "simultaneous
injection competitive binding immunoassay", is the most common
approach for chromatographic immunoassays. A specific example of this
method is shown in Fig. 6
, in which theophylline was measured in serum by a
flow-injection competitive binding immunoassay that used
carboxyfluorescein-impregnated liposomes as the label. The injected
samples contained 50 µL of serum diluted 1:100 and combined in a 1:2
ratio with a working solution of the label. The total cycle time
between samples was 16 min, and good correlation was noted vs a
fluorescence polarization immunoassay (59). Other clinical
analytes that have been measured by simultaneous injection competitive
binding immunoassays include human chorionic gonadotropin
(60), thyroid-stimulating hormone (60), HSA
(61), IgG (62)(63), testosterone
(64), and transferrin (61)(65);
additional studies with theophylline have also been reported
(66)(67)(68)(69).
|
An alternative format that has been explored for clinical testing involves the application of only sample to the immunoaffinity column, followed later by a separate injection of the label. This method is known as a "sequential injection competitive binding immunoassay" (61)(70). One advantage of the sequential injection approach is that even an unlabeled preparation of analyte potentially can be used as the label, provided that this species produces a sufficient signal for detection; this method is particularly useful for complex samples that contain analytes at moderate-to-high concentrations in complex mixtures. Another advantage of this technique over the simultaneous injection format is that there are no matrix interferences present during detection of the label because it is never in contact with the actual sample. However, the sequential injection method does require an additional step vs the simultaneous injection method for the separate application of label to the immunoaffinity column. One strength of both the simultaneous and sequential injection competitive binding methods is that they can be used equally well for either small or large analytes.
sandwich immunoassays
The sandwich immunoassay, or two-site immunometric assay, can also
be performed as part of an affinity chromatographic system
(71)(72)(73)(74). In this technique, two different types of
antibodies that bind to the analyte of interest are used. The first of
these two antibodies is attached to a chromatographic support and is
used to extract the analyte from samples. The second antibody
contains an easily measured tag and is added in solution to the
analyte either before or after sample injection; this second antibody
serves to place a label onto the analyte, thus allowing the amount of
analyte on the immunoaffinity support to be quantified as it and the
label are eluted from the column.
One clinical application in which a chromatographic sandwich
immunoassay has been used is in the determination of intact parathyrin
(PTH) in plasma (73)(74). Fig. 7
shows a typical calibration curve and correlation plot for this
technique. This particular method involved incubation of plasma samples
combined in a 2:1 ratio with a working solution of anti-(134 PTH)
acridinium ester-labeled antibodies for 1 h. After the incubation,
a 100-µL aliquot of this mixture was injected onto an immunoaffinity
column containing anti-(4468 PTH) antibodies, thus producing
formation of sandwich immune complex within the column. The amount of
retained PTH was then determined by examination of the signal produced
by the labeled antibodies as they were eluted at pH 3.0 and passed
through a postcolumn chemiluminescence reactor. The total time per
sample injection was 6.06.5 min, and the limit of detection was 0.2
pmol/L PTH (73)(74). Other analytes that have
been examined by chromatographic sandwich immunoassays are some
antigen-specific antibodies (71) and IgG (72).
|
Like its traditional solid-phase counterpart, an important advantage of
the chromatographic sandwich immunoassay is that it produces a signal
for the bound label that is directly proportional to the amount of
injected analyte (Fig. 7a
). The use of two types of antibodies in a
sandwich immunoassay gives this technique higher selectivity than
chromatographic-based competitive binding immunoassays. The main
disadvantage of the chromatographic sandwich immunoassay is that it can
be used only for analytes such as large peptides or proteins that are
large enough to bind simultaneously to two separate antibodies.
one-site immunometric assays
This is the third format that has been used to automate
immunoassays by affinity chromatography. This approach has been used in
determining such agents as thyroxine (55) and
-(difluoromethyl)ornithine (75). In this technique, the
sample first is incubated with a known excess of labeled antibodies or
Fab fragments that are specific for the analyte of interest. After
binding between the analyte and antibodies has occurred, this mixture
is applied to a column that contains an immobilized analog of the
analyte. This column serves to extract any antibodies or Fab fragments
that are not bound to the original analyte. Meanwhile, those antibodies
or Fab fragments that are bound to analyte from the sample will pass
through the column in the nonretained peak, which is then detected and
used for analyte quantification. Like the chromatographic competitive
binding immunoassays, this method is able to detect both small and
large solutes. However, like a chromatographic sandwich immunoassay, it
also gives a signal for the nonretained label that is directly
proportional to the amount of analyte in the original sample. One
disadvantage of this approach is that relatively pure and highly active
labeled antibodies/Fab fragments must be used to provide a low
background signal.
| Affinity Extraction |
|---|
|
|
|---|
off-line affinity extraction
Off-line extraction is the easiest method for combining an
affinity column with another analytical technique. This approach
typically involves the use of an affinity ligand that is immobilized
onto a low-performance support (e.g., activated agarose) that is packed
into a small disposable syringe or solid-phase extraction cartridge.
After the affinity column is conditioned with the necessary application
buffer or conditioning solvents, the sample is applied and nonbound
sample components are washed off of the packing, as shown in Fig. 1
. An
elution buffer is then applied, and the analyte is collected as it
elutes from the column. In some cases, this eluted fraction is analyzed
directly by a second technique, but in most situations the collected
fraction is first dried and reconstituted in a solvent that is
compatible with the method to be used for quantification. If needed,
the collected solute fraction may also be derivatized before it is
examined by other techniques to obtain improved detection or more
appropriate physical properties (e.g., an increase in solute volatility
before separation and analysis by GC).
The most common ligands in affinity extraction are antibodies, with the terms "immunoextraction" or "immunoaffinity extraction" often being used to refer to this particular extraction technique. Sample preparation by off-line immunoextraction has been the subject of several recent reviews (5)(76)(77)(78)(79), most of which have emphasized its applications in the area of drug residue analysis. Examples involving human samples include the use of immunoextraction before reversed-phase liquid chromatography (RPLC) in the determination of albuterol in plasma (80), human chorionic gonadotropin in urine (81), and ochratoxin A in human serum, plasma, or milk (82). Off-line immunoextraction also has been used for sample clean-up before analysis by GC or GCmass spectrometry in the determination of prostaglandins and thromboxanes (83)(84)(85)(86) or alkylated DNA adducts (87)(88) in human urine. The same approaches have been used in several animal studies involving off-line immunoextraction and RPLC or GC for the detection of alkylated DNA adducts in DNA extracts from rats (89), chloramphenicol in urine and tissue samples from pigs (90), dexamethasone and flumethasone in equine urine (91)(92), ivermectin and avermectin in sheep serum (93), and estrogens (94)(95), nortestosterone (96), or trenbolone (97) in bovine urine and bile samples.
Although antibodies are the most popular ligands in off-line affinity extraction, they are not the only ligands used for this approach. For example, sample extraction by an organomercurial agarose column followed by RPLC analysis has been used for the assessment of urinary 2-thioxothiazolidine-4-carboxylic acid, a proposed indicator of environmental exposure to carbon disulfide (98). Off-line boronic acid columns similarly have been used for the reversed-phase analysis of modified nucleosides in patients with gastrointestinal cancer (99) and in the purification of human platelet glycocalicin before analysis by anion-exchange HPLC (100). A method based on a wheat germ agglutinin extraction column combined with high-performance anion-exchange chromatography has been reported as a means to purify and analyze angiotensinase A and aminopeptidase M in human urine and kidney samples (101). Another application of affinity extraction is in the removal of specific interferences from samples. Examples include the use of protein A and anti-mouse immunoglobulin supports for the removal of human anti-mouse antibodies before the analysis of a sample by immunoassay (102) and with the use of anti-human immunoglobulin IAC or protein A supports to selectively adsorb enzyme-immune complexes (i.e., macroenzymes) from patient samples (103).
It should always be kept in mind when using affinity extraction that
many ligands (even antibodies) will show some binding or
cross-reactivity with solutes that are closely related to the desired
analyte in structure. Each affinity extraction method should be
evaluated for such cross-reactivity by the use of binding and
interference studies with any solutes or metabolites that are similar
to the analyte and that may be present in the samples of interest.
However, even if several solutes do bind to the same extraction column,
this does not present a problem as long as the analyte can be resolved
or discriminated from these other compounds by the method that is used
for quantification. In many cases, this can even be used to an
advantage because it allows several species in the same class of
compounds to be determined in a single analytical run. For example, the
ability of antibodies to cross-react with a parent compound and related
agents or metabolites has been used for the development of
immunoextraction methods for 17
- and 17ß-trenbolone
(97), 17
- and 17ß-nortestosterone (96), and
diethylstilbestrol, dienestrol, and hexestrol (94). This
idea can be taken one step further by the use of multiple types of
antibodies in the same column. This has been used in the HPLC analysis
of testosterone, nortestosterone, methyltestosterone, trenbolone,
zeranol, estradiol, diethylstilbestrol, and related compounds in urine,
where samples were extracted off-line with an affinity column that
contained seven different types of immobilized antibodies
(76).
One advantage of off-line affinity extraction is that the samples collected from the extraction column can be derivatized readily or placed into a different solvent between the sample purification and quantification steps. This advantage is particularly important when affinity extraction is combined with GC, where it is desirable to remove any water from the collected sample before injection onto the GC system and solute derivatization is often required to improve solute volatility or detection. Another advantage of off-line affinity extraction is that it is relatively easy to set up once an appropriate ligand preparation has been selected or obtained. The cost of an affinity extraction cartridge is typically much higher than for conventional solid-phase extraction; however, this difference can be minimized by the careful selection of application and elution conditions so that the same affinity cartridge can be used for multiple samples (76).
on-line affinity extraction
The direct coupling of affinity extraction with other analytical
methods is yet another area that has been the subject of increasing
research. The use of immunoextraction columns as part of HPLC systems
has been of particular interest (5)(6). The
relative ease with which immunoaffinity columns can be incorporated
into an HPLC system makes this appealing as a means for automating
immunoextraction methods and for reducing the time required for sample
pretreatment. In addition, the relatively high precision of HPLC pumps
and injection systems provides on-line immunoextraction with better
precision than off-line extraction methods, because the on-line
approach has more tightly controlled sample application and elution
conditions.
Clinical applications of on-line immunoextraction in HPLC have been
developed for such analytes as
1-antitrypsin
(104), cortisol (105), digoxin (106),
estrogens (107)(108), human epidermal growth
factor (109), lysergic acid diethylamide, lysergic acid
diethylamide analogs and metabolites
(110)(111), phenytoin (112),
propranolol (110),
9-tetrahydrocannabinol (113), and
transferrin (104)(114). Additional details on
these methods are provided in Ref. (5). All of these
particular examples have used immunoaffinity columns combined with
standard analytical columns for RPLC; however, there have also been
reports from the field of biotechnology that have described the use of
on-line immunoextraction with size exclusion (115) or
ion-exchange chromatography (116)(117).
One reason for the large number of reports involving the combination of on-line immunoextraction with RPLC undoubtedly has to do with the popularity of RPLC in routine analytical separations. Another, more fundamental, reason arises from the fact that the elution buffer for an immunoaffinity column is an aqueous solvent that generally contains little or no organic modifier, a feature that makes this same buffer act as a weak mobile phase for RPLC. This means that as a solute elutes from an antibody-based column, it will tend to have strong retention on any on-line reversed-phase support, thus leading to analyte reconcentration. This effect is valuable for analytes that desorb slowly from immobilized antibody columns and thus are difficult to analyze by the direct detection mode of affinity chromatography.
One common format for on-line immunoextraction in RPLC (Fig. 8
) involves injecting the sample onto an immunoaffinity
extraction column, with the nonretained components being flushed into a
waste container. The immunoaffinity column is then switched
on-line with a RPLC precolumn, and an elution buffer is applied to the
antibody support to dissociate any retained analyte. As these analytes
elute, they are captured and reconcentrated at the head of the RPLC
precolumn. After all solutes have left the immunoaffinity column, this
column is then switched back off-line and regenerated by washing with
the initial application buffer. Meanwhile, the RPLC precolumn is placed
on-line with a larger analytical RPLC column, and both are developed
with an isocratic or gradient elution scheme involving the application
of a solvent with an increased organic modifier content. This causes
analytes at the head of the RPLC precolumn to move through the
analytical column and to be separated on the basis of their differences
in polarity. As these solutes elute, they are monitored and quantified
through the use of a flow-through detector.
|
Other ligands besides antibodies, particularly boronates, have been shown to be valuable in performing on-line affinity extraction with HPLC. Examples include several separation methods in which boronate columns have been combined with HPLC columns for the clinical analysis of catechol-related compounds such as epinephrine, norepinephrine, and dopamine (118)(119)(120), dihydroxyphenylalanine (121), dihydroxyphenylacetic acid (121)(122), 5-S-cysteinyldopa (123), and vanillylmandelic acid (124). This same approach has been adapted for profiling (125) and quantifying ribonucleotides in urine (126)(127) and serum (127).
Although not as common as on-line extraction in HPLC, there has been some work investigating the use of affinity extraction coupled directly with GC for the determination of ß-19-nortestosterone and related steroids in urine (128). In this case, a RPLC precolumn was again used to capture and reconcentrate retained analytes as they eluted from an immunoaffinity extraction column. However, this RPLC precolumn also now served to remove any water from the analytes and to place them into a volatile organic solvent (ethyl acetate, which was used as the elution mobile phase). A portion of the analytes that eluted from the RPLC precolumn was then passed into the injection gap of a GC system. Once the solute/organic solvent plug had entered the GC system, a temperature program was initiated for solute separation. One advantage of this approach (and also of immunoextraction/HPLC) is that large volumes of sample can be applied to the immunoaffinity column, thus providing low detection limits. The main disadvantage of on-line immunoextraction in GC is the greater complexity of this method vs off-line immunoextraction or on-line immunoextraction/HPLC.
Several recent studies have considered the additional possibility of
combining on-line immunoextraction with CE. For example,
immunoextraction based on immobilized Fab fragments was used to extract
and concentrate tear samples for the CE analysis of cyclosporin
A and its metabolites in samples from corneal transplant
patients (Fig. 9
) (129). In another study, antibodies were covalently
immobilized in microcapillary bundles or in laser-drilled glass rods
that were then connected to a CE capillary for the on-line
immunoextraction and detection of immunoglobulin E in serum
(130). Finally, a capillary packed with a protein G
chromatographic support has been used to adsorb antibodies for the
extraction and concentration of insulin from serum before
quantification by CE (131).
|
| Postcolumn Affinity Detection |
|---|
|
|
|---|
The direct detection mode of affinity chromatography represents the simplest approach for postcolumn quantification of an analyte, provided that the solute is capable of generating a sufficiently strong signal for detection. One example of this approach involved the use of size exclusion chromatography and postcolumn immunodetection for the analysis of acetylcholinesterase (AChE) in amniotic fluid (134). The method in this report used an immunoaffinity column containing anti-AChE antibodies to capture AChE as it eluted from the analytical column. After the AChE was adsorbed to the immunoaffinity column, a substrate solution for AChE was passed through the column, and the resulting colored product was detected by an on-line absorbance detector.
Other formats also are possible for postcolumn immunodetection, including techniques based on competitive binding immunoassays (133)(135) and sandwich immunoassays (136). However, the one-site immunometric assay is the most common format for immunodetection, and it is the only additional approach that has been used previously in clinical applications. The basic operation of this format involves taking the eluate from the HPLC analytical column and combining this with a solution of labeled antibodies or Fab fragments that bind to the analyte of interest. The mixture of column eluate and antibody or Fab fragments is then allowed to react in a mixing coil and passed through an immunodetection column that contains an immobilized analog of the analyte. The antibodies or Fab fragments that are bound to the analyte will pass through this column and onto the detector, where they will provide a signal that is proportional to the amount of bound analyte. If desired, the immunodetection column can be washed later with an eluting solvent to dissociate the retained antibodies or Fab fragments; but a sufficiently high binding capacity is generally used so that a reasonably large amount of analytical column eluate can be analyzed before the immunodetection column must be regenerated.
One-site immunometric detection originally was used to quantify digoxin and digoxigenin as they eluted from a standard RPLC column by the use of fluorescein-labeled Fab fragments (raised against digoxigenin) and an immobilized digoxin support in the postcolumn detection system (137). This method was then used to successfully monitor both digoxin and its metabolites in plasma and urine samples (137). The same general system was later used along with a restricted-access RPLC column to monitor digoxin, digoxigenin, and related metabolites in serum samples (138).
| Affinity-based Chiral Separations |
|---|
|
|
|---|
As will be seen later, various naturally occurring proteins and
carbohydrates have been used as ligands for chiral separations of
clinical analytes (142)(143)(144)(145)(146)(147)(148)(149)(150)(151)(152)(153)(154)(155)(156)(157)(158)(159)(160)(161). Other, synthetic ligands that
have also been used for chiral separations with clinical samples, such
as derivatives of amylose or cellulose and Pirkle-type
stationary phases (162)(163)(164)(165)(166)(167)(168)(169)(170)(171)(172)(173)(174)(175)(176)(177)(178)(179)(180)(181)(182)(183)(184)(185)(186)(187)(188)(189), but these other ligands will
not be considered in this present review. Most clinical separations
that will be discussed were performed by routine liquid-liquid or
solid-phase extraction of the sample, with the content of this extract
later being injected onto the chiral column of interest. However, this
is not the only approach that can be used. In some cases, a chiral
column was first used to resolve the enantiomers of a particular
solute, followed by collection of these fractions and their on-line or
off-line injection onto a second, achiral column for further separation
and quantification (Fig. 10
) (150)(156). Alternatively, an achiral
column, such as a reversed- or normal-phase support, was sometimes used
to isolate the compounds from the sample, and a chiral column was then
used on-line or off-line to resolve the enantiomers in each peak of
interest (151)(152).
|
protein-based stationary phases
Proteins are one group of affinity ligands that have received some
attention as chiral HPLC stationary phases. Although all proteins are
chiral, only one [
1-acid glycoprotein (AGP)]
has seen any significant use in the analysis of chiral drugs in
clinical samples. AGP (also known as AAG or orosomucoid) is a human
serum protein involved in the transport of many small solutes
throughout the body. AGP differs from HSA (another drug-binding protein
in serum) in that AGP has a lower isoelectric point and contains
carbohydrate residues as part of its structure. The lower isoelectric
point makes AGP more useful than serum albumin in binding cationic
compounds, whereas the carbohydrate residues may play a role in
determining the stereoselectivity of the binding properties of AGP
(141). There are many drugs and related solutes that have
been separated by AGP in human urine, serum, or plasma. Examples of
clinical interest include bunolol (142), citalopram
(143), fenoprofen (144), flurbiprofen
(145), ibuprofen (144)(146), ketamine
(147), ketoprofen (144), methadone
(148)(149)(150), norketamine (147),
norverapamil (151), pindolol (152), thiopentone
(153), vamicamide (154), and verapamil
(151)(155).
Other proteins that have received some attention in clinical applications of chiral HPLC are bovine serum albumin (BSA) and ovomucoid. Ovomucoid is a glycoprotein obtained from egg whites that has been shown to be useful in the separation of cationic solutes (141). BSA is a member of the serum albumin family, which constitute most of the protein content of serum in mammals and are involved in the transport of a wide range of small organic and inorganic compounds throughout the body, including many pharmaceutical agents (190)(191). BSA, and the related protein HSA, tend to bind best to neutral or anionic compounds, thus making these proteins complementary to AGP and ovomucoid in their applications (139)(141). In clinical work, BSA has been used for the chiral separation of leucovorin in plasma (156), and ovomucoid has been used for separating the individual forms of pentazocine in serum samples (157).
carbohydrate-based stationary phases
One class of natural carbohydrates that can be used as
stereoselective ligands in HPLC are the cyclodextrins
(158)(159)(160)(161). These are circular polymers of
-1,4-D-glucose that are produced through the
degradation of starch by the microorganism Bacillus
macerans. The most common forms of these polymers are
-, ß-,
and
-cyclodextrin, which contain six, seven, or eight glucose units,
respectively (139)(141). The cone-shaped
structure and hydrophobic interior cavity of cyclodextrins give them
the ability to form inclusion complexes with numerous small, aromatic
solutes. Furthermore, the well-defined arrangement of hydroxyl groups
about the upper and lower faces of the cyclodextrins provide these
agents with the ability to discriminate between various chiral
compounds. Examples of some clinical applications for cyclodextrins in
HPLC include methods reported for chlorpheniramine (158),
citalopram, desmethylcitalopram, and didesmethylcitalopram
(159), hexobarbital (158), the M1 and M2
metabolites of moguisteine (160), and propranolol
(161).
| Characterization of Drug- and Hormone-Protein Interactions |
|---|
|
|
|---|
Drug- and hormone-protein binding has been examined in affinity chromatography by the use of both immobilized drugs and immobilized proteins, but protein-based columns currently are more common (195). One advantage of using an immobilized protein column for binding studies is the ability to reuse the same ligand preparation for multiple experiments (e.g., up to 500-1000 injections per column in some HPLC studies) (203)(204)(205). It is important in such experiments to first consider and evaluate how effectively the immobilized protein models the behavior of the same protein in its soluble form. Fortunately, there is growing evidence that at least some immobilized proteins, particularly HSA, can be used quite successfully for the study of drug-protein interactions. For example, it has been shown that association constants measured by equilibrium dialysis for soluble HSA with R- and S-warfarin or L-tryptophan (i.e., solutes that interact with one of the two major binding regions of HSA) are in close agreement with values determined using immobilized HSA columns (206)(207)(208). It has also been found that displacement phenomena and allosteric interactions observed for HSA columns are representative of behavior noted for HSA in solution (208)(209)(210)(211)(212)(213).
zonal elution studies
The method of zonal elution is the technique that has been used
most frequently to study the binding of drugs and other solutes on
immobilized protein columns (193)(214). This
generally is done by injecting a small sample of the drug or solute of
interest into the presence of buffer only or a fixed concentration of a
competing agent in the mobile phase. Analysis of the results is
performed by determining how the elution time, or retention factor
(k', also known as the capacity factor) of the injected
solute changes as a function of the concentration of the competing
agent (Fig. 11
). Alternatively, similar experiments can be used to examine
how various solvent conditions affect drug-protein interactions
(141)(207)(215)(216)(217)(218)(219)(220)(221) or to develop
quantitative structure-retention relationships that describe these
binding processes (222)(223)(224).
|
The most common application of zonal elution and HPLC-based affinity chromatography in drug- and hormone-protein studies has been in the examination of the displacement of drugs and hormones from proteins by other solutes (195)(225). Examples include the use of zonal elution to examine the displacement of D,L-thyronine and D,L-tryptophan from HSA by bilirubin or caprylate (226); the competition of R/S-warfarin with racemic oxazepam, lorazepam, and their hemisuccinate derivatives on an HSA column (211); the direct or allosteric competition of octanoic acid on immobilized HSA for the binding sites of R/S-warfarin, phenylbutazone, tolbutamide, R/S-oxazepam hemisuccinate, ketoprofen A/B, and suprofen A/B (213); the competition of R-warfarin and L-tryptophan with D-tryptophan (207) or L-thyroxine and related thyronine compounds on HSA (203)(208); and the displacement of R- and S-ibuprofen by one another at their binding regions on HSA (227). The same technique has been used to characterize the binding sites of nonsteroidal antiinflammatory drugs on HSA (228) and the displacement of nonsteroidal antiinflammatory drugs and benzodiazepines by phenylbutazone, R/S-ibuprofen, or 2,3,5-triiodobenzoic acid from serum albumin columns (229). This type of work can provide not only qualitative information on binding and displacement, but also quantitative information on the equilibrium constants for these processes and retention mechanisms (203)(208)(213)(222)(224)(227)(228)(229). Information on the kinetics of these solute-protein interactions can also be obtained if appropriate data are collected on the width and retention for solute peaks under various flow-rate conditions, as demonstrated recently for R- and S-warfarin (204) and D,L-tryptophan (205) on HSA columns.
frontal analysis studies
The technique of "frontal analysis" or "breakthrough
analysis" is a second method that can be used in affinity
chromatography to study biological interactions (195). In
this method, a solution containing a known concentration of the solute
to be studied is applied continuously to an affinity column. As the
solute binds to the immobilized ligand, the ligand becomes saturated
and the amount of solute eluting from the column gradually increases,
forming a characteristic breakthrough curve. If fast association and
dissociation kinetics are present in the system, then the mean
positions of the breakthrough curves can be related to the
concentration of applied solute, the amount of ligand in the column,
and the association equilibrium constants for solute-ligand binding.
Regarding systems of clinical interest, frontal analysis and affinity
chromatography have been used to investigate the binding of HSA to
R- or S-warfarin
(206)(208) and D- or
L-tryptophan
(205)(207)(208)(213); to
determine the binding capacities of monomeric vs dimeric HSA for
salicylic acid, warfarin, phenylbutazone, mefenamic acid,
sulfamethizole, and sulfonylureas (230); and to examine the
competition of sulfamethizole with salicylic acid for HSA binding
regions (231). This same approach recently was used to
characterize the binding of chemically modified HSA to various
site-specific probe compounds (232<