- 1.1 Laboratory Clinical Biochemical Assays
- 1.2 Biosensor Technology
- 1.2.1 Biosensor Architecture
- 1.2.2 Probe Attachment to Device Surfaces
- 1.2.3 Devices and Transduction
- 1.3 Biosensors and Measurement of Clinical Targets
- 1.4 Signal Interference and the Non-specific Adsorption Problem
- 1.5 A Look at Surface Chemistries to Solve the NSA Issue
- 1.6 A Final Comment
CHAPTER 1: Biosensor Technology and the Clinical Biochemistry Laboratory – Issue of Signal Interference from the Biological Matrix
-
Published:23 Sep 2013
-
Series: Detection Science Series
M. Thompson, S. Sheikh, C. Blaszykowski, and A. Romaschin, in Detection Challenges in Clinical Diagnostics, ed. P. Vadgama and S. Peteu, The Royal Society of Chemistry, 2013, pp. 1-34.
Download citation file:
This chapter discusses the potential use of biosensor technology in the clinical biochemistry laboratory. Various relevant key aspects of biosensor technology are introduced such as the chemistry of attachment of probes to device surfaces and a summary of the main categories of sensors based on electrochemistry, acoustic-wave physics and optical science. Important performance characteristics of typical clinical measurements are appraised with examples being presented. Following this discussion, the relevant issues of device selectivity, sensitivity, dynamic range and calibration with respect to target concentration, and possibility for label-free operation are evaluated. A critical issue for potential clinical measurement is the mandatory requirement for devices to function in biological fluids and matrices, with avoidance of signal interference caused by nonspecific surface adoption. Solutions for the latter problem are summarized. The chapter closes with a look at the possible features of biosensor technology that could be employed in the clinical biochemistry laboratory.
1.1 Laboratory Clinical Biochemical Assays
Assays for biochemicals of clinical interest in biological fluids, tissues or feces can be arbitrarily divided into tests performed in either the clinical biochemistry laboratory or via a point-of-care device.1 The latter technology often takes the form of a one-use structure that yields a result in a localized environment such as the home, hospital bedside or general practitioner’s facility. The blood glucose and pregnancy assay devices, which involve discardable test strips, will be familiar to many. This type of structure has attracted a high level of interest in recent years, not least from the commercial standpoint. Often quoted advantages are said to lie in convenience, speed of analysis and cost savings. There are of course attractive features in certain cases for procuring a measurement in a particular location. Obviously, biosensor devices might be expected to play a major role in this arena. However, the present discussion concentrates on the potential for use of biosensors in the dedicated central biochemistry laboratory, which is typically located in a major hospital for obvious reasons. Assays performed in this type of facility are very wide ranging in scope with samples originating from the areas of diagnosis of medical conditions, toxicology or monitoring of drug therapy. Sera or plasma dominate the types of samples analyzed but many assays also involve various tissues, feces and other bodily fluids such as urine.
Test technology associated with all the various aspects of medicine is a large and expensive operation. In terms of the cost of performing assays, the Province of Ontario, for example, spends well over 1 billion Canadian (CDN) dollars annually, the tests being conducted in both private laboratories and hospital facilities. A large hospital such as St. Michael’s in Toronto performs a vast array of tests such as those involving immunology, hematology, bacteriology, mycology, cytology, genetics and various aspects of pathology in addition to biochemical measurements. With respect to the latter, some 3 million tests are performed annually with reagent costs alone being in the region of 2 million CDN dollars. Although not intended to be exhaustive, Table 1.1 provides an overview of the type of biochemical and immuno-assays conducted by the clinical biochemistry on an annual basis together with total numbers comprised of both in- and out-patient tests.2 The “winner” is the thyroid stimulating hormone (TSH) test, which alone costs several million dollars annually. The equipment employed in the portfolio of clinical measurements ranges from conventional chromatography to sophisticated mass spectrometry, and combinations thereof. As would be expected given the sheer volume of samples, the laboratory is highly automated and robotized. Typically, a blood sample will be bar-coded then centrifuged to extract serum before entering an automated analytical train. Samples may then be directed automatically to their various analytical stations. A good example of a subsequent analysis would be the prevalent magnetic-bead enzyme-linked immunosorbent assay (ELISA) method for targets where immuno-assay is feasible. Despite the high level of automation incorporated into the analytical train, it should be emphasized that measurements are still performed in a batch-based fashion. Having said that, it is certainly the situation that a number of protocols offer multi-analyte capability, the aforementioned ELISA assay being a case in point.
Overview of the type of biochemical and immuno-assays conducted by a clinical biochemistry laboratory on an annual basis (source: St. Michael’s Hospital, Toronto, Canada).
Type of Test . | Number performed (per year) . |
---|---|
Acetone, quantitative* | 448 |
Albumin, quantitative* | 125 360 |
Alcohol, ethyl-quantitative* | 5010 |
Alcohol, fractionation and quantification* | 5889 |
Amylase* | 21477 |
Barbiturates, quantitative* | 13 898 |
Barbiturates, fractionation and quantification (serum) – includes other drugs requiring similar methodology (e.g. tricyclic antidepressants)* | 704 |
Bilirubin, total* | 72 000 |
Bilirubin, conjugated* | 22 545 |
pH* | 20 |
pCO2, pO2 and pH in combination* | 62 757 |
Carbamazepine, quantitative (Tegretol)* | 504 |
Chlordiazepoxide, quantitative (Librium)* | 10 |
Calcium* | 91 723 |
Calcium ionized* | 9676 |
Catecholamines, fractionated* | 2485 |
Chloride* | 227 057 |
Cholesterol, total* | 34 663 |
Acetaminophen* | 5036 |
Carboxyhemoglobin* | 4 |
CO2 content, CO2 combining power, bicarbonate (measured not calculated)* | 219 783 |
Creatine phosphokinase* | 63 914 |
Creatinine* | 257 139 |
Creatinine clearance* | 4884 |
Target drug testing, urine, qualitative or quantitative* | 37 691 |
Diazepam, quantitative (Valium, Vivol)* | 37 |
Drugs of abuse screen, urine* | 48 |
Broad spectrum toxicology screen, urine – includes confirmatory testing* | 7399 |
Electrophoresis, serum – including total protein | 7556 |
Electrophoresis, other than serum – including total protein* | 599 |
Glycosylated hemoglobin – Hgb A1* | 19 394 |
Flurzepam, quantitative (Dalmane)* | 0 |
Glucose tolerance test in pregnancy* | 2694 |
Glucose tolerance test* | 973 |
Gamma glutamyl transpeptidase* | 7705 |
Glucose, quantitative (not by dipstick)* | 188 863 |
Glucose, semiquantitative (dipstick if read with reflectance meter)* | 46 538 |
High density lipoprotein cholesterol* | 32 179 |
5H1AA quantification – U* | 1554 |
Hemoglobin A2 by chromatography* | 2030 |
Iron, total – with iron binding capacity and percent saturation* | 21 054 |
Lactic acid (lactate)* | 13 733 |
Lactic dehydrogenase (L.D.H), total* | 19 245 |
Lipase | 1475 |
Lipoprotein, electrophoresis* | 3 |
Lipoprotein, ultracentrifugation* | 499 |
Lithium* | 866 |
Lidocaine* | 86 |
Methotrexate (amethopterin)* | 115 |
N-acetylprocainamide* | 74 |
Magnesium* | 84 408 |
Metanephrines, total – U* | 2712 |
Methemoglobin* | 77 |
Myoglobin, quantitative – U* | 176 |
Occult blood* | 4293 |
Osmolality (osmolarity)* | 10 044 |
Oxalic acid (Oxalate) – U* | 3104 |
Phenothiazines, quantitive – U* | 69 |
Phosphatase, alkaline* | 82 037 |
Phosphatase, alkaline fractionation* | 62 |
Phosphorus (inorganic phosphate)* | 86 956 |
Plasma hemoglobin* | 44 |
Potassium* | 235 045 |
Protein, total* | 110 937 |
Primidone, quantitative (Mysoline)* | 420 |
Procainamide* | 71 |
Quinidine* | 18 |
Salicylate, quantitative* | 4875 |
SGOT (AST)* | 88 095 |
SGPT (ALT)* | 80 304 |
Sodium* | 237 209 |
Thiocyanates* | 15 |
Triglycerides* | 33426 |
Urea Nitrogen (B.U.N.)* | 201 706 |
Uric Acid (urate)* | 23 038 |
Urinalysis, routine chemical (any of S.G., pH, protein, sugar, hemoglobin, ketones, urobilinogen, bilirubin, leukocyte esterase, nitrite)* | 37 534 |
Urinalysis microscopic examination of centrifuged specimen* | 11 497 |
Valproic acid (valproate)* | 1154 |
VMA, Vanillylmandelic acid (Vanillylmandelate)* | 1212 |
Biochemical assays not included above* | 7844 |
Prostate specific antigen (PSA), total | 1063 |
Alpha Glycoprotein Subunit | 179 |
Amylase isoenzymes | 535 |
Apolipoproteins A and B | 12 478 |
FK506 (Tacrolimus) | 8577 |
Lipoprotein (a) | 792 |
Prostate specific antigen (PSA), total | 4258 |
Troponin | 44 477 |
Homocysteine | 1585 |
Chylomicrons | 4 |
Oligoclonal banding | 1437 |
Citrate | 2967 |
Sirolimus (Rapamycin) | 526 |
Natriuretic peptide – Brain (BNP) | 3201 |
Foetal fibronectin | 16 |
Bioavailable testosterone | 12 886 |
Aldosterone* | 3059 |
Cortisol* | 3470 |
Aminoglycosides (e.g. Gentamicin, Tobramycin)* | 1108 |
Androsternedione* | 1215 |
Digoxin* | 1103 |
ACTH (andrenocorticotrophic hormone)* | 1520 |
Folate, in red cells, to include hematocrite and if requested, serum folate* | 7710 |
Estradiol* | 1914 |
FSH (pituitary gonadotropins)* | 2807 |
Growth hormone* | 2343 |
HCG (human chorionic gonadotropins)* | 9848 |
Hepatitis associated antigen or antibody immuno-assay ‐ per assay (e.g. hepatitis B surface antigen or antibody, hepatitis B core antibody, hepatitis A antibody)* | 11 469 |
Aminophylline (Theophylline)* | 167 |
Anti-DNA* | 1822 |
Diphenylhydantoin (Phenytoin), quantitative (Dilantin)* | 2424 |
Insulin* | 802 |
LH (luteinizing hormone) | 3091 |
Ferritin* | 17 240 |
Parathyroid hormone* | 7256 |
Progesterone* | 908 |
Prolactin* | 3104 |
17‐OH Progesterone* | 308 |
IgE* ‐ not to be billed for RAST test | 1131 |
T4, free ‐ absolute (includes T‐4 total)* | 11 888 |
Testosterone | 12 767 |
TSH (thyroid stimulating hormone)* | 30 406 |
Phenobarbitone* | 363 |
Vitamin B12* | 10 897 |
C-peptide immunoreactivity* | 1084 |
Dehydroepiandrosterone sulfate (DHEAS)* | 664 |
25-hydroxy vitamin D* | 7703 |
T‐3, free* | 6350 |
Thyroglobulin* | 2466 |
Alphafetoprotein | 942 |
Hormone receptors for carcinoma (to include estrogen and/or progesterone assays)* | 793 |
Total | 3 274 901 |
Type of Test . | Number performed (per year) . |
---|---|
Acetone, quantitative* | 448 |
Albumin, quantitative* | 125 360 |
Alcohol, ethyl-quantitative* | 5010 |
Alcohol, fractionation and quantification* | 5889 |
Amylase* | 21477 |
Barbiturates, quantitative* | 13 898 |
Barbiturates, fractionation and quantification (serum) – includes other drugs requiring similar methodology (e.g. tricyclic antidepressants)* | 704 |
Bilirubin, total* | 72 000 |
Bilirubin, conjugated* | 22 545 |
pH* | 20 |
pCO2, pO2 and pH in combination* | 62 757 |
Carbamazepine, quantitative (Tegretol)* | 504 |
Chlordiazepoxide, quantitative (Librium)* | 10 |
Calcium* | 91 723 |
Calcium ionized* | 9676 |
Catecholamines, fractionated* | 2485 |
Chloride* | 227 057 |
Cholesterol, total* | 34 663 |
Acetaminophen* | 5036 |
Carboxyhemoglobin* | 4 |
CO2 content, CO2 combining power, bicarbonate (measured not calculated)* | 219 783 |
Creatine phosphokinase* | 63 914 |
Creatinine* | 257 139 |
Creatinine clearance* | 4884 |
Target drug testing, urine, qualitative or quantitative* | 37 691 |
Diazepam, quantitative (Valium, Vivol)* | 37 |
Drugs of abuse screen, urine* | 48 |
Broad spectrum toxicology screen, urine – includes confirmatory testing* | 7399 |
Electrophoresis, serum – including total protein | 7556 |
Electrophoresis, other than serum – including total protein* | 599 |
Glycosylated hemoglobin – Hgb A1* | 19 394 |
Flurzepam, quantitative (Dalmane)* | 0 |
Glucose tolerance test in pregnancy* | 2694 |
Glucose tolerance test* | 973 |
Gamma glutamyl transpeptidase* | 7705 |
Glucose, quantitative (not by dipstick)* | 188 863 |
Glucose, semiquantitative (dipstick if read with reflectance meter)* | 46 538 |
High density lipoprotein cholesterol* | 32 179 |
5H1AA quantification – U* | 1554 |
Hemoglobin A2 by chromatography* | 2030 |
Iron, total – with iron binding capacity and percent saturation* | 21 054 |
Lactic acid (lactate)* | 13 733 |
Lactic dehydrogenase (L.D.H), total* | 19 245 |
Lipase | 1475 |
Lipoprotein, electrophoresis* | 3 |
Lipoprotein, ultracentrifugation* | 499 |
Lithium* | 866 |
Lidocaine* | 86 |
Methotrexate (amethopterin)* | 115 |
N-acetylprocainamide* | 74 |
Magnesium* | 84 408 |
Metanephrines, total – U* | 2712 |
Methemoglobin* | 77 |
Myoglobin, quantitative – U* | 176 |
Occult blood* | 4293 |
Osmolality (osmolarity)* | 10 044 |
Oxalic acid (Oxalate) – U* | 3104 |
Phenothiazines, quantitive – U* | 69 |
Phosphatase, alkaline* | 82 037 |
Phosphatase, alkaline fractionation* | 62 |
Phosphorus (inorganic phosphate)* | 86 956 |
Plasma hemoglobin* | 44 |
Potassium* | 235 045 |
Protein, total* | 110 937 |
Primidone, quantitative (Mysoline)* | 420 |
Procainamide* | 71 |
Quinidine* | 18 |
Salicylate, quantitative* | 4875 |
SGOT (AST)* | 88 095 |
SGPT (ALT)* | 80 304 |
Sodium* | 237 209 |
Thiocyanates* | 15 |
Triglycerides* | 33426 |
Urea Nitrogen (B.U.N.)* | 201 706 |
Uric Acid (urate)* | 23 038 |
Urinalysis, routine chemical (any of S.G., pH, protein, sugar, hemoglobin, ketones, urobilinogen, bilirubin, leukocyte esterase, nitrite)* | 37 534 |
Urinalysis microscopic examination of centrifuged specimen* | 11 497 |
Valproic acid (valproate)* | 1154 |
VMA, Vanillylmandelic acid (Vanillylmandelate)* | 1212 |
Biochemical assays not included above* | 7844 |
Prostate specific antigen (PSA), total | 1063 |
Alpha Glycoprotein Subunit | 179 |
Amylase isoenzymes | 535 |
Apolipoproteins A and B | 12 478 |
FK506 (Tacrolimus) | 8577 |
Lipoprotein (a) | 792 |
Prostate specific antigen (PSA), total | 4258 |
Troponin | 44 477 |
Homocysteine | 1585 |
Chylomicrons | 4 |
Oligoclonal banding | 1437 |
Citrate | 2967 |
Sirolimus (Rapamycin) | 526 |
Natriuretic peptide – Brain (BNP) | 3201 |
Foetal fibronectin | 16 |
Bioavailable testosterone | 12 886 |
Aldosterone* | 3059 |
Cortisol* | 3470 |
Aminoglycosides (e.g. Gentamicin, Tobramycin)* | 1108 |
Androsternedione* | 1215 |
Digoxin* | 1103 |
ACTH (andrenocorticotrophic hormone)* | 1520 |
Folate, in red cells, to include hematocrite and if requested, serum folate* | 7710 |
Estradiol* | 1914 |
FSH (pituitary gonadotropins)* | 2807 |
Growth hormone* | 2343 |
HCG (human chorionic gonadotropins)* | 9848 |
Hepatitis associated antigen or antibody immuno-assay ‐ per assay (e.g. hepatitis B surface antigen or antibody, hepatitis B core antibody, hepatitis A antibody)* | 11 469 |
Aminophylline (Theophylline)* | 167 |
Anti-DNA* | 1822 |
Diphenylhydantoin (Phenytoin), quantitative (Dilantin)* | 2424 |
Insulin* | 802 |
LH (luteinizing hormone) | 3091 |
Ferritin* | 17 240 |
Parathyroid hormone* | 7256 |
Progesterone* | 908 |
Prolactin* | 3104 |
17‐OH Progesterone* | 308 |
IgE* ‐ not to be billed for RAST test | 1131 |
T4, free ‐ absolute (includes T‐4 total)* | 11 888 |
Testosterone | 12 767 |
TSH (thyroid stimulating hormone)* | 30 406 |
Phenobarbitone* | 363 |
Vitamin B12* | 10 897 |
C-peptide immunoreactivity* | 1084 |
Dehydroepiandrosterone sulfate (DHEAS)* | 664 |
25-hydroxy vitamin D* | 7703 |
T‐3, free* | 6350 |
Thyroglobulin* | 2466 |
Alphafetoprotein | 942 |
Hormone receptors for carcinoma (to include estrogen and/or progesterone assays)* | 793 |
Total | 3 274 901 |
Biosensor technology – in principle – offers rapid, label-free measurements, which can be conducted in highly automatable configurations. Furthermore, the possibility for multiplying tests using generic sensing physics is certainly an attractive strategy. However, the clear reality is that biosensor devices do not figure prominently in the clinical biochemistry laboratory. Use of ion-selective electrode technology for certain cations might represent a contradiction to this statement, but it could justifiably be argued that these devices do not constitute biosensors anyway! The dearth of biosensor devices in the clinical biochemistry laboratory also appears to be matched by the lack of employment of so-called “lab-on-a‐chip”3,4 devices, which are generally characterized by microfluidic sample handling. Time will tell but at present neither of these technologies appears to have fulfilled the promise offered. This chapter attempts to evaluate possible reasons for this including a detailed look at one major problem – that of interference by adsorbed species from the biological matrix. In order to provide a backcloth to discuss the issues at stake, we take a prior concise look at biosensor technology. The chapter has something of an emphasis on label-free methodology.
1.2 Biosensor Technology
1.2.1 Biosensor Architecture
A biosensor is composed of chemical recognition sites (probes) attached to a substrate surface that, in turn, is in close proximity and union with a transducer. Ideally, this configuration responds sensitively and selectively to the presence of (bio)chemicals, we term the target or analyte, usually present in the liquid phase. The technology is based on the recognition of species through selective binding of the target to the probe at the substrate surface of the device with such surface presence being converted into an electrical signal. The overall architecture is depicted in Figure 1.1. Note that probes are invariably composed of biological entities such as cells or biochemicals such as antibodies. The sample itself may or may not be of biological origin. An important aspect of biosensor technology is the nature of the response of the device with respect to time. The field is often considered to include the aforementioned “one-test” disposable systems, where there is no attempt to conduct a measurement over a period of time. This type of device is more often than not at the heart of the point-of-care assay. From the clinical biochemistry point of view with regard to the central laboratory, both time-responsive and one-shot structures might be applicable. This issue will be considered in more detail later in the chapter.
Schematic representation of a biosensor architecture featuring the biosensing interface, transducer and output system.
Schematic representation of a biosensor architecture featuring the biosensing interface, transducer and output system.
where C and S values represent the concentration of analyte and interferents in proximity to the device, and the response sensitivities, respectively. An enormous number of components would be expected to be involved in this equation in terms of samples such as serum or urine. A sensitive response implies maximization of one S value and, for a selective signal, a minimization of all other S values. Clearly, the sensor signal will be a composite of the chemistry of the attachment process and the physical perturbation caused by the probe–analyte complex. In this respect, it is crucial that, when designing a sensor for particular applications, the physics of the complete transduction process be thoroughly characterized and understood. In fact, some devices simply respond to the presence of the analyte, whereas others detect structural shifts caused by target–probe binding. The distinction between these mechanisms is not always evident.
1.2.2 Probe Attachment to Device Surfaces
A mandatory aspect of biosensor fabrication is the attachment of the biochemical probe to the surface of the transducer of choice. Over the years, a host of methods have been developed with this goal in mind.6,7 A number of criteria are relevant to this component of sensor production:
It will be vital to retain the binding activity of the probe in terms of its ability to bind the target. It is anticipated that certain proteins, for example, may be partially denatured and lose affinity upon attachment to the sensor surface, whereas others such as antibodies may be more robust.
The proper spatial orientation of the probe with respect to the plane of the device surface will be an important element to ensure exposure of binding sites for target capture.
In terms of sensitivity, it is crucial to maximize the density of probe molecules per surface area unit. Simply stated, the more probe available on the device surface, the greater will be the sensor signal.
The spatial characteristics of the probe with regard to the surface plane are a factor that is important but not studied widely. If probes are in too close proximity on the surface, this may result in steric hindrance to target binding. In practice, there exists a maximal (optimal) effective probe loading.
There are other more mundane factors such as the cost of reagents and the potential longevity of the modified sensor surface. These will clearly be critical from a commercial standpoint.
Some examples of methods for probe attachment follow. This short compendium is not intended to be an exhaustive treatment.
1.2.2.1 Noncovalent Attachment
The probe of interest can be chemically or physically adsorbed from solution directly, in most cases, on the substrate surface of the device. Various protocols are employed to achieve such an effect such as dip casting, painting, spraying and spin coating, where these terms will be self-explanatory to the reader. The interaction of the probe with the surface will be characterized primarily by hydrogen bonding and/or hydrophobic interactions. Adsorbed molecules can also act as a linker for eventual biomolecule attachment. A system that is widely employed is avidin/streptavidin/neutravidin chemistry.8 Avidin is a tetrameric protein that contains four binding sites for the ligand, biotin. Interaction of the protein with this molecule results in a particularly strong binding (Kd=10−15 M). Accordingly, various biomolecules can be modified with relative facility by biotin incorporation in order to link them to surface-attached avidin or a sister molecule such as the aglycosylated version, neutravidin. An analogous strategy uses protein A.9 The latter is a polypeptide (MW∼42 kDa) isolated from Staphylococcus Aureus that is capable of binding specifically to the Fc region of various antibody molecules and has thus been employed in immunosensor technology.
Another noncovalent approach is the trapping of the recognition molecule within the three-dimensional structure of a specific chemical matrix. The matrix may simply act as a “holding” moiety or be modified in some way to take part in the transduction process. For the former, polymers such as polyacrylamide have been employed in a number of experimental protocols. However, there are a number of potential disadvantages to this type of approach for placing the probe on the device surface including the possibility of biomolecule leakage and/or denaturation. An additional consideration is the necessity for the target to diffuse into the polymer matrix in order for the biochemical interaction to take place. An analogous procedure uses encapsulation by sol-gel technology,10 wherein a solution of a monomer such as an alkoxide (the sol) is induced to polymerize into a biphasic configuration (the gel) that incorporates both liquid and solid. A typical monomer among many would be tetraethylorthosilicate (TEOS – EtO4Si), which is readily hydrolyzed by water to produce a siloxane-based polymeric structure with a gel consistency. A cavity can be formed around the probe of interest in somewhat the same manner as for the polymers mentioned above.
1.2.2.2 Covalent Binding
Attachment via covalent bonds has been by far the most used approach. A very wide variety of chemistries have been employed with modest success in terms of the criteria outlined above.11,12 Many functional groups, whether directly present on the device substrate or obtained by modification, have been utilized to form a probe partial monolayer. Groups are available on biomolecules to instigate the surface link and examples of these for proteins are presented in Table 1.2.
Target functional groups for biosensor immobilization of proteins.
Functional Group . | Amino Acid . |
---|---|
![]() | Cysteine |
![]() | Lysine (side chain) |
![]() | Tyrosine |
![]() | Histidine |
![]() | Tryptophan |
![]() | Arginine |
![]() | Glutamic acid, Aspartic acid |
Functional Group . | Amino Acid . |
---|---|
![]() | Cysteine |
![]() | Lysine (side chain) |
![]() | Tyrosine |
![]() | Histidine |
![]() | Tryptophan |
![]() | Arginine |
![]() | Glutamic acid, Aspartic acid |
A common protocol to bind enzymes, antibodies or molecular receptors to the substrate is to initially functionalize the surface, which is then followed by a second reaction of activation. In essence, this process simply allows a convenient, highly reactive “linker” moiety to be introduced to the system. The resulting interface is then normally allowed to react in turn with one of the protein nucleophilic groups mentioned above. The literature is replete with many examples of this sort of approach and, in Figure 1.2, we provide a schematic of one strategy. If groups already evident on the device surface are not used directly, functionalization of surfaces is often achieved with species such as aminopropyltriethoxysilane (APTES), which reacts with interfacial hydroxyl groups on whatever substrate they are present.
Schematic representation of a common strategy (EDC/NHS chemistry) to covalently biofunctionalize surfaces via a preactivation process. (Reprinted with kind permission of the Royal Society of Chemistry.)
Schematic representation of a common strategy (EDC/NHS chemistry) to covalently biofunctionalize surfaces via a preactivation process. (Reprinted with kind permission of the Royal Society of Chemistry.)
1.2.2.3 Self-Assembled Monolayer Chemistry
The introduction of close-packed monolayers has been used widely in biosensor development. Two very different strategies are employed the first being the Langmuir–Blodgett film technique.13 In this experiment, close-packed monolayers are imposed on a surface by transferring lipid or lipid-like (amphiphilic) films from the Langmuir trough, under the correct surface pressure conditions, by a dipping process. In principle, several layers can be deposited in sequence using this dipping approach. The very important advantage offered by this technique is the possibility to combine artificial lipid membrane configurations directly and in situ with integral membrane proteins (IMPs). Such membrane systems are, of course, widely prevalent as signaling devices in biology.
An important chemistry, which has been significantly developed in recent years, is the self-assembled monolayer (SAM). This approach relies on the use of linking molecules that are engineered to spontaneously form ordered molecular assemblies on solid substrates. In this case, the most common strategy has been the assembly of relatively long chain thiols on the surfaces of clean gold substrates,14,15 as shown in Figure 1.3. The distal end of bifunctional thiols can then be employed for conventional covalent binding of proteins and oligonucleotides as described above (Figure 1.2).16 An alternative technique is the use of trichlorosilanes rather than thiols, which can be bound to surfaces that have been functionalized with or naturally possess –OH groups.17 Examples of substrates in this case would be silicon dioxide (e.g. quartz – Figure 1.4)18 and indium-tin oxide (ITO).19
Long-chain alkyl thiol deposition onto a gold substrate to generate a self-assembled monolayer. (Reprinted with kind permission of the Royal Society of Chemistry.)
Long-chain alkyl thiol deposition onto a gold substrate to generate a self-assembled monolayer. (Reprinted with kind permission of the Royal Society of Chemistry.)
Preparation of an organosilane adlayer on quartz and subsequent covalent functionalization. (Reprinted with kind permission of the American Chemical Society.)
Preparation of an organosilane adlayer on quartz and subsequent covalent functionalization. (Reprinted with kind permission of the American Chemical Society.)
1.2.3 Devices and Transduction
A very large amount of work has been performed on a variety of biosensor structures. A bibliography detailing this effort up to the end of 2012 is provided herein.20 Three distinct areas of physics have been employed with respect to the transduction process, these being electrochemistry, acoustic wave technology and electromagnetic radiation. There is considerable variety in terms of the nature of measurement, for example, possible label-free operation and secondly dipstick or response with time methodology. Moreover, a number of adjunct techniques have been employed in order to enhance the level of information obtained from the sensor determination.
1.2.3.1 Electrochemical Devices
A number of different approaches are possible via electrochemistry,21,22 which involves the transfer of electrons or charge at solid/liquid or liquid/liquid interfaces. Those involving solid structures unsurprisingly dominate the field, which is summarized in Table 1.3. A particularly useful feature of many of these techniques is that they can be combined “naturally” with various forms of integrated electronic circuitry. Indeed, the necessary chemistry can be imposed directly on the surface of such an electronic device (see below).
Electrochemical techniques available for application in biosensor technology.
Technique . | Methodology . |
---|---|
Amperometry | Measures current versus concentration; potential is controlled |
Conductometry | Measures conductance (=1/resistance) at controlled concentration |
Coulometry | Measures change over time with controlled potential |
Potentiometry | Measures electrochemical potential at current=0 |
Voltammetry | Measures current at applied potential; controlled concentration of electroactive species |
Technique . | Methodology . |
---|---|
Amperometry | Measures current versus concentration; potential is controlled |
Conductometry | Measures conductance (=1/resistance) at controlled concentration |
Coulometry | Measures change over time with controlled potential |
Potentiometry | Measures electrochemical potential at current=0 |
Voltammetry | Measures current at applied potential; controlled concentration of electroactive species |
Historically, one of the earliest devices generated from the world of electrochemistry was the sensor based on potentiometry. In this technique, an indicating electrode for the analyte of interest is incorporated in an electrochemical cell (with reference electrode) for which minimal or no current is passed. In this category, the ion-selective (indicating) electrode (ISE) forms the basis of systems, which are capable of detecting ions and other species of biochemical interest.23,24 By far the most common in use is the glass electrode, which is sensitive to changes in hydronium-ion concentration. Other systems available for selective ion detection include electrodes that employ crystalline matrices, e.g. LaF3 for F−, or incorporate liquid ion exchangers in polymer matrices such as systems for sensing Ca2+. The ISE has also been the basic structure used to develop enzyme-based molecular sensors where analytes are allowed to interact with immobilized enzymes and transform into products detected by the electrode.
An alternative transducer to the conventional electrode outlined above is the field-effect transistor (FET), which was introduced by Bergveld25 in the 1980s for ion sensing. Figure 1.5 shows the structure of a typical n-channel enhancement mode insulated-gate FET (IGFET). A conductive pathway, termed a channel, can be instigated between n-type regions. One of the n-type contacts is called the source and the other the drain; a potential is applied that drives current in the channel. The contact on the surface of the center part of the insulating layer is called the gate. It is in this location that the device is employed to detect charge changes connected to the chemistry conducted at the gate/solution interface (Figure 1.6). Note that certain regions of the device must be encapsulated and that the system as mentioned above needs an adjunct reference electrode. The device has been employed in various formats including the detection of ion concentrations (ISFET)26 or immunochemical interactions (IMMUNOFET).27
The n-channel enhancement mode insulated-gate field effect transistor – IGFET. (Reprinted with kind permission of the Royal Society of Chemistry.)
The n-channel enhancement mode insulated-gate field effect transistor – IGFET. (Reprinted with kind permission of the Royal Society of Chemistry.)
IGFET typical immunosensor set-up. Note the insulating protection of source and drain contacts. (Reprinted with kind permission of the Royal Society of Chemistry.)
IGFET typical immunosensor set-up. Note the insulating protection of source and drain contacts. (Reprinted with kind permission of the Royal Society of Chemistry.)
Techniques that are based on measurement of the resulting current versus applied potential are called voltammetric methods. Anodic stripping and cyclic voltammetry would be two examples of such techniques. Amperometry has many definitions in the literature but is more often than not simply regarded to involve the measurement of current associated with an oxidation–reduction reaction at an electrode. It is worth noting that the genesis of biosensor technology was the oxygen-mediated electrode for the amperometric assay of glucose. Over the years, the electrochemical detection of this analyte, given its importance, has been the subject of literally thousands of studies, the history of which has been reviewed by Wang.28,29 One of the most important advances in the technology was the introduction of organometallic mediators such as ferrocene. This iron π‐arene complex composed of a cyclopentadiene sandwich of Fe acts as a convenient electron acceptor for glucose oxidase, thus avoiding the dependence on the role of O2.
The detection of nucleic acid interactions and duplex formation at an electrode surface by an amperometric protocol has been the subject of intensive research for some time.30 One example among several is the use of the redox properties of, for example, ruthenium complexes to enhance the electroactivity of nucleic acid species (the intrinsic redox behavior of nucleic acid moieties is insufficiently sensitive for analytical purposes). The electrochemistry community often describes this protocol as a “label-free” detection strategy, which is obviously nonsense given the necessary use of the Ru adjunct agent.
Finally, we mention electrochemical impedance spectroscopy (EIS). In this technique, a sinusoidally-varying voltage is applied to an electrochemical system and the resulting current is measured, usually over a range of angular frequencies. Recording of these parameters can be employed to compute the real and imaginary components of the electrical impedance (Z). Primarily, the system has been used to detect biomolecular recognition events taking place at an electrode much as the situation with the other electrochemical arrangements outlined above.31
1.2.3.2 Acoustic Wave-Based Biosensor Technology
Acoustic wave sensors are generally very much associated with piezoelectric physics. (It should be noted, however, that a number of available structures do not employ piezoelectric components in a direct sense.) Piezoelectricity is the electric polarization produced by mechanical strain in crystals belonging to certain classes, the polarization being proportional to the strain and changing sign with it.32 Of course, we now recognize that the reverse is true, that is, a specific crystal can be mechanically strained when subjected to an electric polarization. The devices used in biosensing have in common the deformation caused on a piezoelectric crystal by the application of an electric field. The origin of this effect lies in the interaction between electric charge and elastic restoring forces in the crystal. All importantly, the phenomenon cannot take place in a crystal possessing central symmetry. In order to maximize the coupling of electrical and mechanical effects, it is conventional for practitioners to use particular slices of crystals or cuts. For example, with respect to the predominant piezoelectric material, quartz, these are AT‐, ST‐ and BT-cuts. Movement of particles in a piezoelectric crystal caused by an oscillating electric field-induced stress, restored by elastic forces, leads to standing or travelling waves in the material and the phenomenon of piezoelectric resonance.
Several types of acoustic wave devices have been used as biosensors, but the simple thickness–shear mode (TSM) has been by the most employed for detection purposes. The TSM is composed of an electroded (often gold) piezoelectric wafer (usually AT-cut quartz), as depicted in Figure 1.7. An oscillating electrical potential is applied via the electrodes in order to drive mechanical motion in the device. A resonant acoustic shear wave is generated, which travels through the piezoelectric material with little energy dissipation. The wave is reflected at the device/surroundings interface in order to maintain a standing wave (Figure 1.7). In chemical and biosensor applications, material is generally added to the device surface, for example, species involved in biomolecular interaction. On a simple theoretical basis, this scenario was the subject of very early work by Sauerbrey,33 who showed that when material is deposited on the device surface the resonant frequency is changed according to his famous equation that is expressed for quartz as the piezoelectric material (Eq. (1.2)):
where Δf is the frequency change, f0 is the primary resonant frequency of the sensor, A is the effective surface area associated with the piezoelectric process, ρq is the density of quartz, μq is the shear modulus of the particular cut of quartz employed and Δm is the change in mass. The idea of mass detection has spawned the ubiquitous term, quartz crystal microbalance or QCM. Unfortunately, this argument has been widely extended to operation of the device in the liquid phase and has become something of a dogma, especially in the community of electrochemists. In such a medium however, it is a reality that acoustic energy is transferred to the surrounding liquid resulting in a damped wave via the effects of bulk phase viscosity (Figure 1.7).
Working principle of the thickness-shear mode (TSM) acoustic wave device featuring the standing wave in the substrate with propagation of a damped acoustic shear wave into the liquid. (Reprinted with kind permission of the Royal Society of Chemistry.)
Working principle of the thickness-shear mode (TSM) acoustic wave device featuring the standing wave in the substrate with propagation of a damped acoustic shear wave into the liquid. (Reprinted with kind permission of the Royal Society of Chemistry.)
Experimentally, a number of approaches have been employed in order to measure the resonant frequency and other acoustic parameters, in some case in a static fashion, and in others, in flowing liquid through a flow-injection configuration. The most rigorous approach is that provided by what is often termed acoustic network analysis. In this method, the magnitude and phase of impedance of the sensor is determined at a set of frequencies under resonance conditions. A network analyzer is employed to record the Butterworth–Van Dyke equivalent circuit, which essentially relates the physical properties of the device to electrical parameters. In terms of applications, recent years have seen a rapid increase in the development of TSM technology, which has been employed to detect surface-induced protein conformational changes,34 immunochemical interactions,35 nucleic acid hybridization36 and cell-surface attachment.37
A recent acoustic wave device development is the introduction of the electromagnetic piezoelectric acoustic sensor (EMPAS) structure.38 In this technology, acoustic waves are instigated in an electrodeless quartz wafer by an electromagnetic field produced in close proximity to the wafer by a flat spiral coil (Figure 1.8). The secondary electric field associated with the coil drives the piezoelectric effect in the device. The configuration possesses a number of important advantages:
It is not necessary to operate the device with contact metal electrodes in place and with electrical connections. Acoustic resonance is driven remotely. This renders advantages in terms of flow-through design and surface chemistry can be studied directly in an unhindered fashion.
Crucially, it is possible to operate the sensor at ultra-high frequencies, e.g. 1 GHz, via bulk acoustic wave overtones. This leads to higher analytical sensitivity.
It is possible to tune the device with ease to specific frequencies, which could potentially lead to important interfacial chemical information.
The surface chemistry for biomolecule attachment involves SiO2, which is a more developed area of chemistry than is the case for binding to metals such as gold.
Working principle of the electromagnetic piezoelectric acoustic wave sensor (EMPAS): acoustic resonance is remotely driven in the electrode-free quartz substrate by a secondary electric field induced by the electromagnetic field associated with an AC-powered flat spiral coil. (Reprinted with kind permission of the Royal Society of Chemistry.)
Working principle of the electromagnetic piezoelectric acoustic wave sensor (EMPAS): acoustic resonance is remotely driven in the electrode-free quartz substrate by a secondary electric field induced by the electromagnetic field associated with an AC-powered flat spiral coil. (Reprinted with kind permission of the Royal Society of Chemistry.)
Surface-launched wave devices have also been employed in biosensing, where waves are generated in a piezoelectric substrate by transducers that are placed on the surface of the material. Particle movement, which is generally detected by separate transducer(s) imposed on the same surface, is often restricted to the “near” surface of the substrate, unlike the case for the TSM discussed previously. For an introduction to several of these devices, please refer to reference 39. The surface acoustic or Rayleigh wave (SAW) sensor has an interdigital transducer (IDT) fabricated on the piezoelectric substrate (e.g. ST-cut quartz) as shown in Figure 1.9. The chemistry related to the sensing processes is conducted on the area between the IDTs and it is generally the case that the electrodes have to be isolated from the liquid if the device is operated in that medium, in a similar fashion to that described above for FETs. An example of another device employed via this technology is also portrayed in Figure 1.9. This is the surface transverse wave (STW).40 The use of these biosensors, which all involve very similar biomolecule immobilization strategies to those outlined above, have been reviewed by Rapp and coworkers.41
Upper: schematics of a surface acoustic wave device. Center: Rayleigh waves contain both shear and compressional components. Lower: schematics of a surface transverse sensor. (Reprinted with kind permission of the Royal Society of Chemistry.)
Upper: schematics of a surface acoustic wave device. Center: Rayleigh waves contain both shear and compressional components. Lower: schematics of a surface transverse sensor. (Reprinted with kind permission of the Royal Society of Chemistry.)
1.2.3.3 Electromagnetic Radiation and Optical Biosensor Technology
Virtually, the full gamut of physics offered by optical science has been employed over the years for the detection of fundamental biophysical processes, biochemical binding events or species of bioanalytical interest such as biomarkers for disease. Techniques include those based on measurements of absorption, luminescence, interference, reflectance, scattering (including Raman spectroscopy) and refractive-index phenomena. The wide variety of techniques employed in optical sensing have been nicely reviewed.42 One of the main features of this area has been the concentration on label-free detection.43
One of the earlier devices was based on optical waveguide technology and, in particular, the optical-fiber sensor. In essence, light transmission along a fiber is produced via a guided wave through an integral process of internal reflection. Light can transmit along a fiber with complete internal reflection or via some “loss” of electromagnetic energy through the effect of reflection/refraction at the interface where materials of different refractive index are involved. Both these processes have been employed extensively in the development of biosensors and in this context the term “optrode” has appeared, as obviously spawned by the older concept of an electrode in the field of electrochemistry. In the first methodology, the fiber is simply used as a delivery system for light interaction with an optical configuration placed at the distal end of the structure,44 as shown in Figure 1.10a. In the second scenario called the intrinsic structure, some light energy finds its way under certain conditions into the medium outside the fiber in the form of a penetrative evanescent wave (Figure 1.10b). Importantly, the interaction of surface chemistry at this interface with the evanescent wave can result in perturbation of the light phase, intensity and polarization. There are many examples of such an arrangement in biosensor detection.45 This type of device has been employed successfully in the assay of “real” samples such as for the detection of pathogens.46
(a) Schematics of extrinsic fiber-optic delivery of radiation to a cell for absorption measurement. (b) Evanescent radiation penetrating to the exterior of a fiber in an intrinsic configuration. (Reprinted with kind permission of the Royal Society of Chemistry.)
(a) Schematics of extrinsic fiber-optic delivery of radiation to a cell for absorption measurement. (b) Evanescent radiation penetrating to the exterior of a fiber in an intrinsic configuration. (Reprinted with kind permission of the Royal Society of Chemistry.)
Surface plasmon resonance (SPR) has become one of the leading technologies with respect to detection in bioanalytical and biophysical chemistry. The physics effect is based on the fact that valence electrons of metals such as gold and silver exhibit oscillations in their density. If these oscillations, in the form of surface waves, are present at the interface of the metal with a material of a different dielectric constant, they can be excited by the introduction of electromagnetic radiation. The classical approach to the study of this phenomenon was initially introduced by Kretchmann47 and is shown schematically in Figure 1.11a. The incident radiation is reflected at the metal/dielectric boundary resulting in an evanescent wave in the metal, much as described above for optical fibers. At the correct (resonant) angle, this wave can couple in a resonant fashion with the frequency of the incoming radiation, resulting in the excitation of electron density oscillation mentioned above, leading to the term surface plasmon resonance. Experimentally, in sensor operation, biochemical binding events are conducted at the metal surface and detected through measurement of shifts in the SPR angle, observed wavelength of absorption or change in the position of reflectivity. A resonant transfer of electromagnetic energy into the surface plasmon wave is observed through a minimum in the plot of incident angle versus reflected intensity (Figure 1.11b).
(a) Kretschmann configuration for surface plasmon resonance (SPR). (b) Incidence angle versus reflected intensity of radiation as a result of SPR. (Reprinted with kind permission of the Royal Society of Chemistry.)
(a) Kretschmann configuration for surface plasmon resonance (SPR). (b) Incidence angle versus reflected intensity of radiation as a result of SPR. (Reprinted with kind permission of the Royal Society of Chemistry.)
A simplified schematic of a typical SPR experiment is shown in Figure 1.12. This type of instrument designed for work in the biosensor arena is generally capable of analyzing several channels (gold surfaces) using microfluidic sample introduction in a real-time, label-free fashion through flow-injection technology. The system generates a response plot, which is often referred to in the field as a “sensorgram”, presumably in the light of chromatograms and the like! With respect to applications, the SPR technique has been employed in a very wide variety of cases such as adsorption of proteins, cells and nucleic acids on gold and modified metal surfaces, detection of biomolecular interactions such as found in immunochemistry and nucleic acid duplex formation.48–53 The method has proven to be particularly helpful for epitope determination in the former area.
Typical SPR experiment involving flow injection of target analyte into the flow cell. (Reprinted with kind permission of the Royal Society of Chemistry.)
Typical SPR experiment involving flow injection of target analyte into the flow cell. (Reprinted with kind permission of the Royal Society of Chemistry.)
Finally in this section, we mention the technique of interferometry. The latter is based on the superposition of, usually, electromagnetic waves that yields information concerning the original nature of the waves. In biosensing technology, several types of device designs have been employed in order to attempt the detection of biochemical species, examples being Mach–Zehnder’s, Young’s and Hartman’s sensors (these are depicted in Figures 1.13a–c). The first of these involves two different light pathways, as is typical in interferometry, with one being subjected to passage through the sample. The laser radiation is then combined with a reference beam resulting in the interference-based signal. The analytical arm incorporates an evanescent field, which interacts with, in this application, a biochemical moiety. This technology was employed in the earlier years for the detection of protein chemistry, where analytical detection limits of around 50 pM were found.54
(a–c) Mach–Zehnder’s, Young’s (multichannel) and Hartman’s interferometer sensors, respectively. (Reprinted with kind permission of Elsevier.)
(a–c) Mach–Zehnder’s, Young’s (multichannel) and Hartman’s interferometer sensors, respectively. (Reprinted with kind permission of Elsevier.)
An additional interference-based structure is Young’s device.55 The overall principle is similar to the sensor mechanism employed in Mach–Zehnder’s device, the detection mode is different, however. In this case, the output from the sample and reference channels is combined to form interference fringes on a CCD screen. The method involves Fourier transform of the spatial intensity measure at the detector screen. In Hartman’s interferometer,56 electromagnetic radiation is coupled into a waveguide using gratings, which allows the interference phenomenon to occur between reference and sample strips. As with the other two devices outlined above, protein chemistry was detected as imposed on the waveguide strip.
1.3 Biosensors and Measurement of Clinical Targets
It is very evident from Table 1.1 that clinical biochemical detection and measurements involve a vast array of targets in biological matrices. Although it is not the aim of this chapter to detail the fundamental requirements of the laboratory – these are provided elsewhere57 – we do summarize some of the main necessary criteria. These are then compared in the context of what is offered by biosensor technology. Essential properties and protocols of clinical determinations are:
High throughput. As outlined above, the clinical biochemistry laboratory faces literally thousands of samples on a daily basis, which may have one particular analyte determination required or perhaps several on the same sample. The sheer numbers involved mandate a very high level of automation and robotic technology. This means that considerable technical effort is required to keep equipment in operation. For the most part, modern-day clinical analysis incorporates the capability to handle samples in a batch-wise process.
Speed of analysis. Obviously, the requirements in this criterion will vary enormously. However, it would be desirable to possess the capability to perform rapid analysis, for example, as requested by a physician in an emergency situation.
Generic methodology. Where possible, it is advantageous to possess the possibility to perform the determination of different targets employing the same detection physical chemistry. It goes without saying that this aspect, by the very nature of target analyses required, will be somewhat limited. However, there exist several technologies possessing this attribute, such as magnetic-bead ELISA.
Validation of dose–response and limit-of-detection (LOD) characteristics. The former parameter refers to the meaningfulness of the signal obtained from the target in a biological matrix in terms of its concentration. This would be more familiar to the analytical chemist as the “calibration curve” for the particular method involved. The LOD would also be well known to the analyst being the minimum concentration value that a method can detect (in analytical chemistry, generally 3 times the standard deviation of the background noise).
Cost per analysis. Given the great expense involved in the operation of a central clinical laboratory, it is hardly surprising that the calculated cost per assay will always be a factor under consideration. Factored into this are also the costs associated with salaries of operators, reagents, equipment and its maintenance.
Functionality in complex media. It is clearly crucial for a particular method to function in samples such as blood, serum, plasma or urine. We have left this aspect to the end for reasons that will become apparent.
We now turn to the “ideal” biosensor. The relevance of these criteria to the clinical chemistry operation will be apparent to the reader.
High sensitivity with resulting low value of limit-of-detection. This parameter will be governed by the physics of the detection technique employed – the S values outlined above.
High accuracy in terms of concentration measurement for the analyte. Signaling conducted with high reproducibility – a precision issue.
The dynamic range with respect to analyte response should be as wide as possible. For certain targets, indeed, clinical values may vary considerably.
Signal approach – dipstick/batch or real-time sensing. It would be advantageous for a particular device to be capable of operation in the dipstick/batch manner or as a real-time sensor. The speed of response will obviously be important in both situations although for different reasons. Real-time operation offers the potentially attractive possibility to incorporate the biosensor into an automated flow-injection scenario. This is not the well-known segmented situation, but a technique based on analyte dispersions allowed to flow past the device surface. The flow-injection analysis (FIA) will require that the biosensor signal be reversible or the target can be washed from the surface within the flow system.
Device response calibration in terms of concentration. This is mandatory and especially important for the real-time sensor. This feature has constituted an extremely difficult problem when it comes to, for example, the operation of a corporeal implantable structure. Early solutions to this issue may well lie in a strategy of device self-referencing.
Selectivity or even specificity (see reference 5 for an excellent definition of these parameters) with respect to response to the analyte. In this respect with specific regard to clinical chemistry, it is essential that the device function selectively in media that contain very high concentration of various biomolecules such as proteins and even cells. These species will tend to adsorb to the sensor surface potentially interfering with the required signal. This is one of the reasons why the legendary closed-loop implantable glucose device is not available despite many years of research. The issue of nonspecific adsorption (NSA) represents a true Achilles Heel for application of biosensor technology in clinical biochemistry.
Finally, it is worth noting that there have been several attempts to assess the capability of biosensors to detect molecules of clinical interest. Table 1.4 shows an example where the authors place performance criteria in the context of classical analytical chemistry.58
Analyte detected (matrix) . | Sensor type . | Linear range . | LOD . | Sensitivity (slope) . | Additional information . |
---|---|---|---|---|---|
Glucose (buffer) | CNT enzyme-based biosensor | Up to 10 mM | 0.25 μM | 30.2±0.5 μA/mM | r2=0.999; n=10 |
Glucose (blood serum) | Amperometric enzyme-based biosensor | 1.0 μM–0.8 mM | 0.5 μMb | 2.3 mA/M | r2=0.999; n=7 |
CV(%)=4.8 | |||||
Glucose (human serum) | Amperometric enzyme-based biosensor | 1.0 μM–1.6 mM | 0.69 μMb | 69.26 mA/M.cm2 | r2=0.999; n=25 |
CV(%)=1.1 | |||||
Glucose (buffer) | Amperometric CNT-based biosensor | 0.005–0.3 mM | 3 μMb | 80±4 mA/M.cm2 | r2=0.996; n=5 |
CV(%) <5% | |||||
Glucose (buffer) | CNT enzyme-based biosensor | Up to 2 mM | 4 μM | 0.33 μA/mM | r2=0.998 |
Glucose (buffer) | CNT enzyme-based biosensor | 15.0 μM–6.0 mM | 7 μMb | r2=0.992; n=8 | |
CV(%)=1 .6 | |||||
Glucose (buffer) | Amperometric enzyme-based biosensor | 0.02–5.7 mM | 8.2 μMb | 8.8±0.2 mA/M.cm2 | CV(%)=2.1 |
Cholesterol (human serum) | Amperometric CNT-based biosensor | 0.004–0.7 mM | 1.0 pMb | 1.55 μA/mM | r2=0.9954; n=8 |
CV(%)=4.2% | |||||
Cholesterol (human serum) | CNT enzyme-based biosensor | 0.004–0.10 mM | 1.4 μMc | y=14.23x+0.2602 | |
r2=0.999; n=11 | |||||
CV(%)=2.5 | |||||
Cholesterol (buffer) | Amperometric enzyme-based biosensor | 1.2 μM–1 mM | 0.12 mMd | r2=0.9998; n=11 | |
CV(%)<4.0 | |||||
Cholesterol (buffer) | Amperometric enzyme-based biosensor | Up to 12 mM | 0.18 mM | 1.61 nA/mM | |
Cholesterol (buffer) | CNT amperometric biosensor | 0.2–6.0 mM | 0.2 mMb | 0.559 μA/mM | |
Cholesterol (buffer) | Impedance enzyme-based biosensor | 50–400 mg/dL | 25 mg/dL | r2=0.9946 | |
Cholesterol (buffer) | Impedance enzyme-based biosensor | 25–400 mg/dL | 25 mg/dL | 7.76×105 Abs/mg.dL | y=0.03546x+7.76×10−5 |
Cholesterol (human serum) | SPR enzyme-based biosensor | 50–500 mg/dL | 50 mg/dL | 1.04 m°/mg.dL | |
Cholesterol (human blood) | Voltammetric enzyme-based biosensor | 0.01–0.06 mM | 0.85 nA/μM | r2=0.99 | |
y=0.6054x+0.2411 | |||||
Glutamate (buffer) | CNT enzyme-based biosensor | 0.2–250 μM | 0.01 μMb | 433 μA/mM.cm2 | r2=0.9984 |
CV(%)=4.8 | |||||
Glutamate (buffer) | Amperometric enzyme-based biosensor | 0.5 μM–500 mM | 0.01 μMb | 100±10 mA/M.cm2 | r2=0.991 |
Glutamate (buffer) | Amperometric enzyme-based biosensor | 0.25 mM | 2.0 μMb | 80±10 nA/μM.cm2 | n=5 |
Glutamate (buffer) | Amperomctric enzyme-based biosensor | 10 μM–1.5 mM | 5 μM | 88.8 nA/mM | y=88.832x |
r2=0.9945; n=3 | |||||
Glutamate (buffer) | Amperometric enzyme-based biosensor | 5 μM–0.5 mM | 5 μM | 0.62 mA/mM.cm2 | r2=0.9968 |
Glutamate (buffer) | Amperometric enzyme-based biosensor | 20 μM–0.75 mM | 20 μM | y=5.0145x+0.1638 | |
r2=0.9950; n=3 | |||||
CV(%)=9.6 | |||||
Glutamatc (buffer) | CNT enzyme-based biosensor | 25 μM | 25 μM | 1.29±0.18 nA/μM.cm2 | y=1.29x+0.088 |
CEA (human serum) | Amperomctric immunosensor | 0.5–5.0 ng/mL | 0.2 ng/mLb | ||
CEA (human serum) | SPR immunosensor | 0.5 ng/mL | S/N=2.7 | ||
CEA (human serum) | Potentiometric immunosensor | 4.4–85.7 ng/mL | 1.2 ng/mLd | y=24.36 log x−10.37 | |
r=0.997; n=5 | |||||
CV(%)=2.2 | |||||
CEA (human serum) | Electrochemical immunosensor | 2–14 U/mL | 1.73 U/mLb | 3.2±0.1 mL/U | r2=0.9975 |
Nitric oxide (buffer) | Amperometric CNT-based sensor | 0.2–150 μM | 0.08 pM | ||
Nitric oxide (buffer) | Hemoglobin-based electrochemical biosensor | 1–50 μM | 0.003 pMb | 0.0424 μA/μM | y=0.0424x+2.441 r=0.9978 |
Nitric oxide (buffer) | Hemoglobin-based amperometric biosensor | 0.01–1.0 μM | 5.0 pMb | y=1.356x+1.168×10−2 | |
r2=0.9989; n=8 | |||||
CV(%)=4.2 | |||||
Nitric oxide (buffer) | Hemoglobin-based amperometric biosensor | 0.04 nM–5.0 μM | 20 pMb | ||
C-Reactive Protein (human blood) | Magnetic immunosensor | 0.025–2.5 μg/mL (28.5 pM–2.9 nM) | 25.0 ng/mLe | y=0.2108x+72.31 r2=0.989 | |
C-Reactive Protein (human blood) | SPR Immunosensor | 2–5 μg/mL (2.3–5.75 nM) | 1 μg/mLf | y=10.675x–16.995 r2=0.9867 | |
C-Reactive Protein (human blood) | Capacitive immunosensor | 0.025–1 μg/mL | y=2.78x+1.29 r2=0.9255CV(%)=1.72 | ||
C-Reactive Protein (human blood) | SPR immunosensor | 5–25 μg/mL | |||
Catecholamines: DA, EPI, NEPI (buffer) | Amperometric enzyme-based biosensor | DA: up to 40 μM | 0.2 μMb | 75 nA/μM | CV(%)=5.3 |
EPI: up to 55 μM | 0.4 μMb | 45 nA/μM | CV(%)=6.0 | ||
NEPI: up to 55 μM | 0.3 μMb | 60 nA/μM | CV(%)=6.1 | ||
Catecholamines: DA, EPI, NEPI (human urine and plasma) | Optical fiber enzyme-based biosensor | 5–125 pg/mL | DA: 2.1 pg/mLd | y=0.344x+0.4; r2=0.9998 | |
EPI: 2.6 pg/mLd | y=0.140x+0.04; r2=0.9996 | ||||
NEPI: 3.4 pg/mLd | y=0.252x+0.38; r2=0.9997 | ||||
EPI (buffer) | Fiber-optic enzyme-based biosensor | 0.2–0.9 μM | y=−0.016533x−1.30882 r2=0.988 | ||
NEPI (buffer) | DNA-based biosensor | 0.5–80 μM | 5 nM | y=0.85x+8.87; r2=0.993 | |
DA (buffer) | Electrochemical enzyme-based CNT biosensor | 1.0–30.0 μM | 400 nMb | y=0.61x−1.19 | |
DA (buffer) | Amperometric enzyme-based biosensor | 5–120 μM | 68.6 mA/mM.cm2 | r2=0.999 | |
Acetylcholine (buffer) | CNT-ISFET sensor | 10–10 000 μM | 0.01 μM | 378.2 μA/decade | |
Acetylcholine (buffer) | Amperometric enzyme-based biosensor | 1–1500 μM | 0.6 μMb | ||
Acetylcholine (buffer) | Carbon fiber-based biosensor | 1–100 μM | 1 μMb | r=0.998 | |
Mycobacterium tuberculosis (simulation sample) | Bulk acoustic wave impedance biosensor | 2×103–3×10 7 cells/mL | 2×103 cells/mL | log y=11.23457−0.00763x r2=0.9458; n=5 | |
Mycobacterium tuberculosis (saliva) | QCM immunosensor | 105–108 cells/mL | 105 cells/mL | ||
Mycobacterium tuberculosis (buffer) | QCM sensor | 102–107 cfu/mL | 102 cfu/mL | y=−2.90x−64.29 r2=0.997 | |
Human ferritin (serum) | QCM immunosensor | 0.1–100 ng/mL | y=89.52 log x+117.68 r2=0.9944 | ||
CV(%): 4.97–20.4 | |||||
Human ferritin (serum) | SPR immunosensor | 0.2–200 ng/mL | y=152.13 log x+168.16 | ||
r2=0.9997 | |||||
Choline (buffer) | Carbon fiber-based biosensor | 1–100 μM | 1 μMg | r2=0.9837 | |
Diphtheria antigen (buffer) | Potentiometric CNT-based immunosensor | 24–1280 ng/mL | 7.8 ng/mL | y=74 log x−98.6 r=0.9978; n=13CV(%)=2.3 | |
hCG (human serum) | Electrochemical CNT-based immunosensor | 0.5–5.0 mlU/mL | 0.3 mlU/mLd | 5.38±0.10 mlU/mL | r=0.9998 |
hCG (human serum and urine) | QCM immunosensor | 2.5–500 mlU/mL | 2.5 mlU/mLb | CV(%)<5.0 | |
Insulin (serum) | SPR immunosensor | 6–200 ng/mL | 6 ng/mL | ||
Vibrio cholerae (buffer) | SPR immunosensor | 3×105–3×109 cells/mL | 105 cells/mL | r=0.9984 |
Analyte detected (matrix) . | Sensor type . | Linear range . | LOD . | Sensitivity (slope) . | Additional information . |
---|---|---|---|---|---|
Glucose (buffer) | CNT enzyme-based biosensor | Up to 10 mM | 0.25 μM | 30.2±0.5 μA/mM | r2=0.999; n=10 |
Glucose (blood serum) | Amperometric enzyme-based biosensor | 1.0 μM–0.8 mM | 0.5 μMb | 2.3 mA/M | r2=0.999; n=7 |
CV(%)=4.8 | |||||
Glucose (human serum) | Amperometric enzyme-based biosensor | 1.0 μM–1.6 mM | 0.69 μMb | 69.26 mA/M.cm2 | r2=0.999; n=25 |
CV(%)=1.1 | |||||
Glucose (buffer) | Amperometric CNT-based biosensor | 0.005–0.3 mM | 3 μMb | 80±4 mA/M.cm2 | r2=0.996; n=5 |
CV(%) <5% | |||||
Glucose (buffer) | CNT enzyme-based biosensor | Up to 2 mM | 4 μM | 0.33 μA/mM | r2=0.998 |
Glucose (buffer) | CNT enzyme-based biosensor | 15.0 μM–6.0 mM | 7 μMb | r2=0.992; n=8 | |
CV(%)=1 .6 | |||||
Glucose (buffer) | Amperometric enzyme-based biosensor | 0.02–5.7 mM | 8.2 μMb | 8.8±0.2 mA/M.cm2 | CV(%)=2.1 |
Cholesterol (human serum) | Amperometric CNT-based biosensor | 0.004–0.7 mM | 1.0 pMb | 1.55 μA/mM | r2=0.9954; n=8 |
CV(%)=4.2% | |||||
Cholesterol (human serum) | CNT enzyme-based biosensor | 0.004–0.10 mM | 1.4 μMc | y=14.23x+0.2602 | |
r2=0.999; n=11 | |||||
CV(%)=2.5 | |||||
Cholesterol (buffer) | Amperometric enzyme-based biosensor | 1.2 μM–1 mM | 0.12 mMd | r2=0.9998; n=11 | |
CV(%)<4.0 | |||||
Cholesterol (buffer) | Amperometric enzyme-based biosensor | Up to 12 mM | 0.18 mM | 1.61 nA/mM | |
Cholesterol (buffer) | CNT amperometric biosensor | 0.2–6.0 mM | 0.2 mMb | 0.559 μA/mM | |
Cholesterol (buffer) | Impedance enzyme-based biosensor | 50–400 mg/dL | 25 mg/dL | r2=0.9946 | |
Cholesterol (buffer) | Impedance enzyme-based biosensor | 25–400 mg/dL | 25 mg/dL | 7.76×105 Abs/mg.dL | y=0.03546x+7.76×10−5 |
Cholesterol (human serum) | SPR enzyme-based biosensor | 50–500 mg/dL | 50 mg/dL | 1.04 m°/mg.dL | |
Cholesterol (human blood) | Voltammetric enzyme-based biosensor | 0.01–0.06 mM | 0.85 nA/μM | r2=0.99 | |
y=0.6054x+0.2411 | |||||
Glutamate (buffer) | CNT enzyme-based biosensor | 0.2–250 μM | 0.01 μMb | 433 μA/mM.cm2 | r2=0.9984 |
CV(%)=4.8 | |||||
Glutamate (buffer) | Amperometric enzyme-based biosensor | 0.5 μM–500 mM | 0.01 μMb | 100±10 mA/M.cm2 | r2=0.991 |
Glutamate (buffer) | Amperometric enzyme-based biosensor | 0.25 mM | 2.0 μMb | 80±10 nA/μM.cm2 | n=5 |
Glutamate (buffer) | Amperomctric enzyme-based biosensor | 10 μM–1.5 mM | 5 μM | 88.8 nA/mM | y=88.832x |
r2=0.9945; n=3 | |||||
Glutamate (buffer) | Amperometric enzyme-based biosensor | 5 μM–0.5 mM | 5 μM | 0.62 mA/mM.cm2 | r2=0.9968 |
Glutamate (buffer) | Amperometric enzyme-based biosensor | 20 μM–0.75 mM | 20 μM | y=5.0145x+0.1638 | |
r2=0.9950; n=3 | |||||
CV(%)=9.6 | |||||
Glutamatc (buffer) | CNT enzyme-based biosensor | 25 μM | 25 μM | 1.29±0.18 nA/μM.cm2 | y=1.29x+0.088 |
CEA (human serum) | Amperomctric immunosensor | 0.5–5.0 ng/mL | 0.2 ng/mLb | ||
CEA (human serum) | SPR immunosensor | 0.5 ng/mL | S/N=2.7 | ||
CEA (human serum) | Potentiometric immunosensor | 4.4–85.7 ng/mL | 1.2 ng/mLd | y=24.36 log x−10.37 | |
r=0.997; n=5 | |||||
CV(%)=2.2 | |||||
CEA (human serum) | Electrochemical immunosensor | 2–14 U/mL | 1.73 U/mLb | 3.2±0.1 mL/U | r2=0.9975 |
Nitric oxide (buffer) | Amperometric CNT-based sensor | 0.2–150 μM | 0.08 pM | ||
Nitric oxide (buffer) | Hemoglobin-based electrochemical biosensor | 1–50 μM | 0.003 pMb | 0.0424 μA/μM | y=0.0424x+2.441 r=0.9978 |
Nitric oxide (buffer) | Hemoglobin-based amperometric biosensor | 0.01–1.0 μM | 5.0 pMb | y=1.356x+1.168×10−2 | |
r2=0.9989; n=8 | |||||
CV(%)=4.2 | |||||
Nitric oxide (buffer) | Hemoglobin-based amperometric biosensor | 0.04 nM–5.0 μM | 20 pMb | ||
C-Reactive Protein (human blood) | Magnetic immunosensor | 0.025–2.5 μg/mL (28.5 pM–2.9 nM) | 25.0 ng/mLe | y=0.2108x+72.31 r2=0.989 | |
C-Reactive Protein (human blood) | SPR Immunosensor | 2–5 μg/mL (2.3–5.75 nM) | 1 μg/mLf | y=10.675x–16.995 r2=0.9867 | |
C-Reactive Protein (human blood) | Capacitive immunosensor | 0.025–1 μg/mL | y=2.78x+1.29 r2=0.9255CV(%)=1.72 | ||
C-Reactive Protein (human blood) | SPR immunosensor | 5–25 μg/mL | |||
Catecholamines: DA, EPI, NEPI (buffer) | Amperometric enzyme-based biosensor | DA: up to 40 μM | 0.2 μMb | 75 nA/μM | CV(%)=5.3 |
EPI: up to 55 μM | 0.4 μMb | 45 nA/μM | CV(%)=6.0 | ||
NEPI: up to 55 μM | 0.3 μMb | 60 nA/μM | CV(%)=6.1 | ||
Catecholamines: DA, EPI, NEPI (human urine and plasma) | Optical fiber enzyme-based biosensor | 5–125 pg/mL | DA: 2.1 pg/mLd | y=0.344x+0.4; r2=0.9998 | |
EPI: 2.6 pg/mLd | y=0.140x+0.04; r2=0.9996 | ||||
NEPI: 3.4 pg/mLd | y=0.252x+0.38; r2=0.9997 | ||||
EPI (buffer) | Fiber-optic enzyme-based biosensor | 0.2–0.9 μM | y=−0.016533x−1.30882 r2=0.988 | ||
NEPI (buffer) | DNA-based biosensor | 0.5–80 μM | 5 nM | y=0.85x+8.87; r2=0.993 | |
DA (buffer) | Electrochemical enzyme-based CNT biosensor | 1.0–30.0 μM | 400 nMb | y=0.61x−1.19 | |
DA (buffer) | Amperometric enzyme-based biosensor | 5–120 μM | 68.6 mA/mM.cm2 | r2=0.999 | |
Acetylcholine (buffer) | CNT-ISFET sensor | 10–10 000 μM | 0.01 μM | 378.2 μA/decade | |
Acetylcholine (buffer) | Amperometric enzyme-based biosensor | 1–1500 μM | 0.6 μMb | ||
Acetylcholine (buffer) | Carbon fiber-based biosensor | 1–100 μM | 1 μMb | r=0.998 | |
Mycobacterium tuberculosis (simulation sample) | Bulk acoustic wave impedance biosensor | 2×103–3×10 7 cells/mL | 2×103 cells/mL | log y=11.23457−0.00763x r2=0.9458; n=5 | |
Mycobacterium tuberculosis (saliva) | QCM immunosensor | 105–108 cells/mL | 105 cells/mL | ||
Mycobacterium tuberculosis (buffer) | QCM sensor | 102–107 cfu/mL | 102 cfu/mL | y=−2.90x−64.29 r2=0.997 | |
Human ferritin (serum) | QCM immunosensor | 0.1–100 ng/mL | y=89.52 log x+117.68 r2=0.9944 | ||
CV(%): 4.97–20.4 | |||||
Human ferritin (serum) | SPR immunosensor | 0.2–200 ng/mL | y=152.13 log x+168.16 | ||
r2=0.9997 | |||||
Choline (buffer) | Carbon fiber-based biosensor | 1–100 μM | 1 μMg | r2=0.9837 | |
Diphtheria antigen (buffer) | Potentiometric CNT-based immunosensor | 24–1280 ng/mL | 7.8 ng/mL | y=74 log x−98.6 r=0.9978; n=13CV(%)=2.3 | |
hCG (human serum) | Electrochemical CNT-based immunosensor | 0.5–5.0 mlU/mL | 0.3 mlU/mLd | 5.38±0.10 mlU/mL | r=0.9998 |
hCG (human serum and urine) | QCM immunosensor | 2.5–500 mlU/mL | 2.5 mlU/mLb | CV(%)<5.0 | |
Insulin (serum) | SPR immunosensor | 6–200 ng/mL | 6 ng/mL | ||
Vibrio cholerae (buffer) | SPR immunosensor | 3×105–3×109 cells/mL | 105 cells/mL | r=0.9984 |
CNT, Carbon nanotube; CEA, Carcinoembryonic antigen; DA, Dopamine; EPI, Epinephrine; LOD, Limit of detection; NEPI, Norepinephrine; QCM, Quartz crystal microbalance; SPR, Surface plasmon resonance; hCG, Human chorionic gonadotropin.
LOD is three times the signal-to-noise ratio.
LOD is 3s/k with s the standard deviation of 11 determinations and k the slope of calibration plot.
LOD is three times the residual standard deviation.
LOD is 10 times the background noise.
LOD is higher than four times the background (0.5 AU, Arbitrary Units).
LOD is approximately twice the noise level.
1.4 Signal Interference and the Non-specific Adsorption Problem
In biosensor technology, the undesired “non-specific adsorption” (NSA) of adversary species (sometimes also referred to as “non-specific binding”) – as opposed to the “specific adsorption” of target analytes – is a serious and prevailing concern for many detection platforms intended to perform analysis in extremely challenging biological samples, more often than not blood serum or plasma. In fact, even cleared of the cellular components of blood, these biological matrices still consist of highly complex mixtures of potentially interfering biomolecules – in particular various types of proteins at high concentration (60–80 g/L)59 – that have the propensity to adsorb nonspecifically to the sensing surface of devices thereby preventing the detection, not to mention the quantification, of target analytes present at considerably lower concentration (down to ng/L or a difference of nine orders of magnitude).60 Indeed, during the biorecognition phase (wherein analytes are expected to site-specifically bind to complementary receptors immobilized on the sensing platform), non-specifically adsorbing species also generate physicochemical stimuli that are indiscriminately detected by the biosensor and interfere with the specific response of the target analyte (Figure 1.14).
The recurring problem in biosensor technology: the non-specific adsorption (NSA) of adversary species interfering with the specific target analyte response.
The recurring problem in biosensor technology: the non-specific adsorption (NSA) of adversary species interfering with the specific target analyte response.
The most unfortunate immediate consequence of NSA – besides altering the performance of the biosensor, may it even be operational in biological samples in the first place – is the occurrence of “false positives”, i.e. a response incorrectly interpreted as a genuine binding event, which understandably renders these biosensors irrelevant for real-world applications.61 Arguably, NSA is the single most important reason why biosensors still have not found a prominent place as alternative diagnostic tools in clinical analysis at the beginning of the 21st century despite tremendous promise notably in terms of cost/ease of operation and miniaturization for “point-of-care” applications – the ultimate aim of biosensor technology.62 More generally, NSA generates a high, often overwhelming background signal that may lower the sensitivity of the biosensor (poor signal-to-noise ratio) to clinically irrelevant levels.60,63 Admittedly, the conventional “enzyme-linked immunosorbent assay” (ELISA) – which displays remarkable sensitivity down to the low pM range64 but remains a non-domestic test to be performed in a laboratory setting – still is promised a bright future.
To reliably alleviate the detrimental effect of NSA, it is necessary to better understand the complex mechanism(s) and dynamics of interaction involved once the biosensing surface comes into contact with the biological environment and the different types of potentially interfering species present within (e.g. proteins). In essence, the process of non-specific protein adsorption is quite complex and governed by several different factors, which includes: (i) the nature of the proteins contained in the biological fluid notably with respect to their structure (e.g. globular), size (molecular weight) or distribution of charge/polarity (isoelectric point); (ii) the physicochemical properties of the biosensing surface (e.g. charge, topography and morphology, surface energy); and (iii) the environmental conditions (i.e. pH, ionic strength and temperature).65 To add to the complexity of the situation, proteins are flexible entities that can assume a variety of different conformational states, which results in the possibility for them to unfold and adopt the adequate geometry to fit the underlying surface, decrease the energy of interaction and irreversibly adsorb.66 Furthermore, the energetics of adsorption itself is typically composed of diverse molecular forces (hydrophobic/electrostatic interactions, hydrogen bonding) that interplay to various degrees in a complex overall manner.65 In multicomponent samples, proteins compete for surface binding sites, which triggers a series of collisions and results in a cascade of adsorption–desorption/exchange processes,66 governed by the aforementioned conformational phenomena that occur at the more fundamental molecular level. Empirically, it has long been observed that protein adsorption from mixtures follows a temporal pattern, wherein higher mobility, more abundant proteins first adsorb transiently before being gradually replaced by less motile, scarcer ones with higher surface affinity.66 This general, well-established phenomenon of sequential protein adsorption is known as the “Vroman effect”.66
Undoubtedly, many factors affect the interactions of proteins with materials (of which the NSA of adversary species to biosensing surfaces is a particular case) and excellent reviews – pertaining to both the theoretical and experimental efforts made to model/probe the complex molecular events at play – can be found elsewhere in the literature.66,67 Nevertheless, despite the mechanistic complexity of (non-specific) protein adsorption, some general rules relating the properties of proteins with their ability to adsorb to surfaces have been devised. For instance, the larger a protein, the more likely it is to possess multiple adhesion sites and readily adsorb to surfaces.65 The same holds true when considering the flexibility of a protein (taken here as its ability to unfold), especially when (more) binding sites are exposed during conformational remodelling. This can occur in particular when a hydrophobic pocket (may it be a large inner protein core or simply a more limited surface domain) – previously hidden from the surrounding aqueous medium (wherein folded proteins reside fully hydrated) to minimize energetically unfavored polar/nonpolar interactions – is revealed upon adsorption.66
NSA is a particular case of a more general, ubiquitous phenomenon – surface fouling – that spontaneously occurs when exogenous biomaterials are exposed to biological environments. For many situations plagued by fouling – that extend well beyond the field of biosensors – tremendous efforts have been devoted over several decades to engineer antifouling surfaces, most traditionally through the imposition of organic films able to efficaciously resist protein adsorption.68 As will be discussed in some more detail in the following section, numerous types of such coatings have been reported in countless literature publications and ultralow fouling (<5 ng/cm2) has now been achieved – for biotechnologically irrelevant, single-protein buffered solutions.68 Regrettably, few coatings present/retain in actuality such a remarkable level of performance when exposed to highly complex biological media such as blood serum or plasma (even diluted), these being otherwise more demanding.68 Fortunately, recent times have witnessed an increase in studies entirely focused on minimizing – or ideally eliminate entirely – fouling for real-world biological samples.68 Assuredly, this trend should benefit the field of biosensor technology and help eradicate the NSA plague.
1.5 A Look at Surface Chemistries to Solve the NSA Issue
In Section 1.1, we have seen that the actual surface presented by a given biosensor structure to a biological fluid can take many forms. This will include inorganic material such as silica or indium-tin oxide in addition to polymers and organic linker molecules. It is a reality that the NSA or fouling discussed in the previous section can occur on all these materials to varying degrees. Accordingly, it is clear that there will be no panacea in terms of avoiding or even reducing the adsorption phenomenon. Although there has been a level of research aimed at the NSA issue with biosensors, it is certainly the case that most relevant work has been published on the study of material–biological fluid interactions. This topic has recently been reviewed in detail by the present authors.68 The majority of the research described therein deals with adsorption of simple protein solutions and the like, with a much more modest effort devoted to biological fluids. A brief scan at fouling follows with some emphasis on our own efforts to avoid the phenomenon.
By far the majority of past related work describes the use of a great variety of organic coatings such as peptides, polyethylene glycols (PEG), zwitterionic sulfo‐ and carboxybetaines, methacrylates, acrylamide and biomimetic surface chemistries.68 A vast literature on such coatings, often relatively thick in dimension, deals with the minimization, as mentioned above, of protein adsorption from simple buffer solutions. Whether in the latter type of solutions or more complex media, the PEG molecular family constitutes one of the most studied, but despite the high level of interest in this system the mechanism that lies behind the “PEG effect” remains obscure. One prevalent theory is that a kosmotropic water “barrier” to biological macromolecule adsorption is instigated by the polymer.68
In our research, we have focused on attempts to combine antifouling behavior with crosslinkers that are capable of immobilizing biochemical probes.17 Covalently-bound adlayers based on silanization methodology, much as described in Section 1.2, have been employed in conjunction with the ultra-high frequency (GHz) EMPAS sensor. This is in order to study the surface effects caused when devices are subjected to a dispersion of goat serum.69 The adlayer-forming molecules are custom-designed with several properties in mind. The molecules contain the ether linkage as present in PEG; they form ultrathin films (∼5‐Å thick) and are capable of merging with sister crosslinking molecules that form similar surface covalent bonds. The latter contain a distal functionalizable moiety that can be used to attach probes in a subsequent, preactivation-free step.17 The behavior of MEG-OH was compared directly with related molecules with respect to fouling by the components of serum (Figure 1.15).69 The remarkable result from this work is that a single ether oxygen atom in the monoethylene moiety of MEG-OH very significantly alters the surface behavior of the sensor in terms of interfacial adsorption from this matrix. On a speculative level, we believe that an intercalated water-based structure stabilized by the distal ‐OH groups is responsible for this effect. Further studies are underway both from a fundamental point of view and also with employment of the antifouling effect in biosensor experiments.
Modification of quartz with various structurally related organosilane surface modifiers.(Reprinted with kind permission of the Royal Society of Chemistry.)
Modification of quartz with various structurally related organosilane surface modifiers.(Reprinted with kind permission of the Royal Society of Chemistry.)
1.6 A Final Comment
A glance into a clinical biochemistry laboratory in a major hospital – such as St. Michael’s located in Toronto, Canada – reveals groups of large instruments connected together with sizeable trains composed of robotic equipment. Much of this is orchestrated around batch-based assays and technology. The set-up is very far removed from biochip, biosensor and lab-on-a‐chip devices, all of which have been touted for potential application in clinical assays. As outlined herein, the main reasons for this appear to lie in speed of analysis, convenience and saving with respect to reagent cost, etc. The latter appears to be a non-issue since such cost constitutes a minor component for the laboratory budget. In view of the requirements for the clinical laboratory, one may wonder how biosensor technology could realistically contribute to the operation of this sort of facility. In this respect, biosensor devices can function particularly well in a flow-injection scenario, especially those systems based on label-free detection involving acoustic wave and SPR physics. This approach can lead to the avoidance of batch measurements and lengthy trains of machines. Another advantage would be the possibility to perform generic detection, that is, the sensing of different targets with the same aforementioned physics/technologies. However, there exist two main problems that remain to be solved. The first would be to render a particular sensing platform reusable in the FIA configuration. A second would be to find a solution to the all-prevalent fouling problem encountered with biological fluids as discussed in some detail above.
Lastly, we would like to mention a final word about microfluidics and lab-on-a‐chip devices, which appear to be ideal for combination with biosensor detection. These technologies seem to be more suited for point-of-care systems rather than the central clinical laboratory. However, despite the frequently emphasized promises offered by these devices, there is no doubt that NSA still remains a serious obstacle to be tackled. For example, NSA will not only occur at the detector employed in the device but also in channels of the microfluidic set-up.70
The authors are grateful to the Collaborative Health Research Program of BSERC and CIGR for support of their research. SS thanks the Province of Ontario for the award of an Ontario Graduate Scholarship.