Universidad de Antioquia 

Facultado de Medicina 

Corporación de Ciencias Básicas Biomédicas 

Doctorado en Ciencias Básicas Biomédicas con Énfasis en Bioquímica, Farmacología y 

Fisiología 

 

 

 

 

Biosensores electroquímicos para la medición de citocinas:  Primer biosensor electroquímico 

basado en el receptor natural de la Interleucina-5 

 

 

 

Electrochemical biosensors for cytokine measurement: The first electrochemical biosensor 

based on the natural receptor of the Interleukine-5 

 

 

 

David José Pérez Cardona 

 

 

 

Tutor: Edwin Bairon Patiño González, Ph.D. 

Co-tutor: Jahir Orozco Holguín, Ph.D. 

 

 

 

 

Comité Asesor: 

Marcela Manrique Moreno, Ph.D. 

Lucas Hernán Blandón Naranjo, Ph.D. 

 

 

 

 

Tesis Doctoral 

2025



Resumen 

Las citocinas son proteínas que regulan el sistema inmune, aumentando o disminuyendo según el 

estado inmunofisiológico. Si bien la medición de citocinas puede tener valor diagnóstico, sus niveles 

suelen ser variables y difíciles de analizar dado su pleotropismo, su actividad esencialmente 

paracrina y la sinergia entre sí y con otros mediadores inmunes. Adicionalmente, las citocinas se 

producen y destruyen inmediato a la toma de la muestra, planteando un claro reto para su 

monitoreo como biomarcadores del proceso salud-enfermedad. En este contexto, es necesario 

desarrollar dispositivos altamente específicos y sensibles para el monitoreo de citocinas. Los 

biosensores electroquímicos se presentan como alternativa a las técnicas convencionales para la 

determinación de citocinas. Dadas las posibilidades de miniaturización, los bajos costos y sus 

notables virtudes metrológicas, los biosensores electroquímicos han sido propuestos para la 

medición de biomarcadores en los puntos de atención (POCT, sigla en inglés), así como para el 

seguimiento permanente en dispositivos portables. Estos últimos son los dos mejores enfoques 

tecnológicos para evitar los problemas mencionados en la medición de citocinas. Puesto que los 

receptores naturales de las citocinas tienen un único ligando, su especificidad y selectividad es de 

las más altas encontradas en la naturaleza.  

Como hipótesis de trabajo, aquí se explora la posibilidad de usar los receptores naturales de las 

citocinas como biorreceptores para el desarrollo de biosensores electroquímicos destinados a su 

medición. Por lo tanto, el objetivo general fue desarrollar un biosensor electroquímico para la 

medición de interleucina-5 usando como biorreceptor la cadena alfa de su receptor natural (IL-

5Rα). La aproximación metodológica incluyó la revisión sistemática de la biología implicada en la 

biodistribución y el mecanismo de acción general de las citocinas (y otros analitos con un 

comportamiento similar). Así mismo, se contrastaron las principales plataformas electroquímicas 

usadas para el monitoreo citocínico en los formatos POCT y portable (i.e., tiempo real o continuo). 

Se renaturalizaron tanto la IL-5 como una forma recombinante del IL-5Rα con una marca de poli-

glicina y poli-histidina (Gly-His IL-5Rα), a partir de cuerpos de inclusión aislados de Escherichia coli.  

Análisis de la IL-5 por cromatografía de afinidad e infrarrojo con Transformada de Fourier sugirieron 

una adecuada renaturalización. El Gly-His IL-5Rα se inmovilizó sobre un electrodo de trabajo de 

carbón modificado con nanopartículas de óxido de Níquel por electrodeposición. Los análisis 

electroquímicos y espectroscópicos de rayos X (XPS) confirmaron la composición superficial del 

electrodo. Se ensambló una plataforma impedimétrica capaz de detectar IL-5 en buffer fosfato y 

diluidos de suero dopados. El rango lineal dinámico fue de 0.125-2.5 µg/mL y el límite de detección 

de 150 ng/mL. Hasta donde sabemos, este es el primer biosensor electroquímico publicado que usa 

IL-5Rα como biorreceptor resaltando la novedad del trabajo y la contribución al estado del arte. 

Este trabajo sugiere que es posible emplear los receptores naturales de citocinas como 

biorreceptores para mejorar la especificidad y la selectividad de los biosensores electroquímicos, así 

como ampliar su uso al estudio de interacciones proteína-proteína. En perspectiva, se requiere más 

trabajo que permita producir receptores naturales más estables y a escala, así como estrategias de 

inmovilización más eficientes. 



Abstract 

Cytokines are proteins that regulate the immune system, increasing or decreasing according to the 

immunophysiological state. Although measuring cytokines can have diagnostic value, their levels 

are often variable and complex to analyze given their pleiotropism, their essentially paracrine 

activity, and the synergy between them and other immune mediators. Additionally, cytokines are 

produced and destroyed immediately after the sample is taken, posing a clear challenge for their 

monitoring as biomarkers in the health-disease process. In this context, developing highly specific 

and sensitive devices for cytokines monitoring is required.  

Electrochemical biosensors could be an alternative to conventional techniques for determining 

cytokines. Given the possibilities of miniaturization, low costs, and remarkable metrological virtues, 

electrochemical biosensors have been proposed for point-of-care testing (POCT) and permanent 

monitoring in wearable devices. The latter are the two best technological approaches to avoid the 

aforementioned problems in cytokine measurement. Since natural cytokine receptors have a single 

ligand, their specificity and selectivity are among the highest found in nature. 

Here, the possibility of using natural cytokine receptors as bioreceptors to assemble 

electrochemical biosensors for their measurement is explored as a working hypothesis. Then, the 

general objective was to develop an electrochemical biosensor for measuring interleukin-5 using 

the alpha chain of its natural receptor (IL-5Rα) as a bioreceptor. The methodological approach 

included a systematic review of the biology involved in the biodistribution and general mechanism 

of action of cytokines (and other analytes with similar behavior). The leading electrochemical 

platforms used for cytokine monitoring were also contrasted in POCT and wearable formats (i.e., 

real-time or continuous). Both IL-5 and a recombinant form of IL-5Rα with a poly-glycine and poly-

histidine tag (Gly-His IL-5Rα) were refolded from inclusion bodies isolated from Escherichia coli. IL-

5-based affinity chromatography and infrared analysis suggested adequate refolding. Gly-His IL-5Rα 

was immobilized on a carbon working electrode electroplated with nickel oxide nanoparticles. 

Electrochemical and X-ray spectroscopic (XPS) analyses confirmed the surface functionalization of 

the electrode. An impedimetric platform capable of detecting IL-5 in phosphate buffer spiked with 

serum was assembled. The dynamic linear range was 0.125-2.5 µg/mL, and the detection limit was 

150 ng/mL. 

To our knowledge, this is the first published electrochemical biosensor using IL-5Rα as a 

bioreceptor, highlighting the novelty of the work and the contribution to the state-of-the-art. It 

suggests that it is possible to employ natural cytokine receptors as bioreceptors to improve the 

specificity and selectivity of electrochemical biosensors and extend their use to study protein-

protein interactions. From this perspective, further work is required to produce more stable and 

scalable natural receptors and more efficient immobilization strategies. 

 



Table of Contents 

 

1. Introduction and Problem Statement ................................................................................................... 6 

2. Research question ............................................................................................................................... 8 

3. General Hypothesis .............................................................................................................................. 8 

4. Objectives ............................................................................................................................................ 8 

4.1. General Objective .............................................................................................................................. 8 

4.2. Specific Objectives............................................................................................................................. 8 

5. Theoretical Framework ........................................................................................................................ 9 

5.1. Chapter 1. Protein Production by Refolding ....................................................................................... 9 

5.1.1. From 1D to 3D, the in vitro protein folding problem .................................................................... 9 

5.1.2. Proteostasis: In-vivo protein folding dictates the Inclusion Bodies (IBs) formation ................. 11 

5.1.3. Fundamentals on protein refolding ............................................................................................. 13 

5.2. Chapter 2. Electrochemical Nanobiosensors as Point-of-care Testing Solution to Cytokines 

Measurement Limitations ...................................................................................................................... 19 

5.2.1. Introduction .................................................................................................................................. 19 

5.2.1.1. Physiology of Cytokines Behavior: A Biological Limitation for their Measurement............... 20 

5.2.1.2. Technical Limitations of (Pre)Analytical Procedures to Cytokines Measurement ................. 25 

5.2.1.3. Common Technology Used to Measure Cytokine in Clinically Relevant Conditions .............. 26 

5.2.2. Electrochemical Nanobiosensors and their Potential Advantages in Cytokines Determination

 ................................................................................................................................................................ 28 

5.2.2.1. Nanostructures as Enhancers of the Working Electrode Features .......................................... 30 

5.2.2.2. Assembly of Cytokine Biorecognition Platforms for Electrochemical Biosensing .................. 40 

5.2.2.3. Chemistry for Bioreceptor Immobilization onto the Working Electrode ................................ 43 

5.2.2.4. Cytokine Electrochemical Immunosensors .............................................................................. 45 

5.2.2.5. Cytokine Electrochemical Aptasensors .................................................................................... 47 

5.2.2.6. Electrochemical Biosensors Based on Cytokine Cognate Receptors: An Innovative Door to be 

Opened ................................................................................................................................................... 51 

5.2.3. Microfluidic Adapted Electrochemical Nanobiosensors: Towards Automated POC Multiplexed 

Cytokine Measurement .......................................................................................................................... 52 

5.2.4. Considerations for Optimal Sampling to Cytokines Measurement ............................................ 58 

5.2.5. Conclusion and Future Remarks .................................................................................................. 59 

References .............................................................................................................................................. 61 



5.3. Chapter 3. Wearable electrochemical biosensors to measure biomarkers with complex 

blood‑to‑sweat partition, such as proteins and hormones ...................................................................... 73 

5.3.1. Introduction .................................................................................................................................. 73 

5.3.2. Sweat gland physiology and WEB's sweat sampling .................................................................. 76 

5.3.3. Wearable epidermal biosensors (WEBs) design and mode of function for sweat's analysis .... 84 

5.3.4. Wearable epidermal biosensors based on screen‑printed electrodes (WEB‑SPEs) .................. 87 

5.3.5. Wearable epidermal biosensors based on field‑effect transistors (WEB‑FETs) ........................ 93 

5.3.6. Challenges of WEBs design and possible solutions ..................................................................... 97 

5.3.7. Conclusions, remarks, and prospects ........................................................................................ 102 

References ............................................................................................................................................ 105 

6. Results ............................................................................................................................................. 119 

6.1. Chapter 4. IL-5Rα-Based Electrochemical Biosensor: Towards Building Biosensors with Natural 

Receptors ............................................................................................................................................ 119 

6.1.1. Introduction ................................................................................................................................ 120 

6.1.2. Experimental procedures ........................................................................................................... 121 

6.1.3. Results and Discussion ............................................................................................................... 121 

6.1.4. Conclusion .................................................................................................................................. 131 

References ............................................................................................................................................ 131 

6.2. Supplementary Material ............................................................................................................... 137 

6.2.1. Materials ................................................................................................................................. 137 

6.2.2. Methods and results for Gly-His IL-5Rα-based electrochemical biosensor assembly ......... 140 

References of Supplementary Material........................................................................................... 160 

7. General Discussion and Perspectives ................................................................................................ 162 

8. General Conclusions......................................................................................................................... 164 

 

  

 



6 
 

1. Introduction and Problem Statement 
 

Clinical laboratories and other healthcare facilities spend most of their time carrying out vital 

physiologic measurements related to patients, which are necessary for diagnosis and 

prognosis. Despite the expertise of lab scientists in executing and analyzing the tests, there is a 

delay in collecting biological specimens, processing them, and delivering the results. Some 

conventional tests lack specificity and selectivity, contributing to false-positive or false-negative 

outcomes and misdiagnosis. Yet, biosensors may offer some solutions to the commented 

issues.  

 

An electrochemical biosensor uses material from a biological entity, called a bioreceptor, in 

direct contact with a transducer platform called a modified electrode. After the bioreceptor 

binds the target analyte, the electric properties of the electrode produce variations correlated 

with changes in the target analyte concentrations. Detected changes over a polarized electrode 

are converted into electrical signals (i.e., current or potential) at the electrolyte-electrode 

interface to determine unknown target concentrations in a sample. Part of the versatility of 

electrochemical biosensors to be miniaturized resides in their relatively simplistic approach to 

measurement. Since the required potentiostat to do electrochemical assessments may be 

portable and require few quantities of samples, they tend to be accessible and affordable for 

most health professionals and even for patients (e.g., glucometer). On the other hand, the 

biorecognition element improves the required accuracy to estimate the real level of the 

analyte. The commented set of elements, accessibility, affordability, and accuracy, was 

proposed in 2003 by the World Health Organization Special Programme for Research and 

Training in Tropical Diseases (WHO/TDR), mainly as mandatory features that must have a 

device to do point-of-care testing (POCT) in diagnostics.  

 

POCT focusing is valuable for analytes affected by their low half-lives in biological fluids. 

Cytokines are a paramount example. Cytokines are proteins that act as the first delivery 

couriers that inform the immune system about the potential risks from the surrounding 

environment. As they are not exclusively produced by immune cells, septic or aseptic 

challenges occurring at any tissue are often accompanied by local cytokine release. Cytokine 

messengers are found at very low concentrations in corporal fluids because they tend to be 

autocrine or paracrine. The systemic circulation of cytokines takes place mainly during 

advanced compromise. Additionally, once a sample is harvested, cytokines may be produced or 

destroyed locally. The last one is one of the reasons why it has not been possible to consolidate 

the cytokines levels as a diagnostic criterion. For instance, the time between sampling and the 

turnaround of the result generated by enzyme-linked immunosorbent assay (ELISA) or flow 

cytometry may reach ~2-4 hours. If the administrative regular journey of data is included, the 



7 
 

release time of results, until they reach the physician, is even 6-7 hours. The commented 

landscape implies that any effort to measure the cytokines as biomarkers must include a highly 

specific and sensitive system that may be multiplexed for fast assessments, ideally on a POCT 

approach.  

 

Electrochemical biosensors are versatile and portable, easy, and rapid to use. Most reported 

cytokine measurement platforms use antibodies and nucleic acids as biorecognition elements. 

Thus, in this research work, we evaluated the possibility of developing an electrochemical 

biosensor to measure cytokines using their natural receptors as biorecognition elements, 

harnessing their high selectivity and specificity properties. Concretely, a Gly-His tagged 

recombinant version of the ectodomain of alpha chain natural receptor of interleukin-5 (Gly-His 

IL-5Rα) was synthesized in-house and immobilized onto a screen-printed carbon electrode 

(SPCE). The built device was tested in its capacity to detect a recombinant version of IL-5, which 

was also synthesized in-house. The first chapter provides a structural biochemical approach 

about the challenges to produce recombinant proteins by refolding. The second one reviews 

various immune-physiological obstacles to measuring cytokine levels and how electrochemical 

biosensors could overcome them using a POCT approach. The third chapter contrasts several 

wearable technologies based on electrochemical biosensing to monitor analytes with an 

intricate blood-to-sweat partition, including cytokines. Finally, we present an original research 

paper where an impedimetric biosensor is developed and proposed as a prototype for further 

improvement in cytokine measurement and protein-protein interaction testing. Given that the 

two first chapters discuss the cytokine´s biology and their measurement and the 

electrochemical biosensors functioning, the presented state-of-the-art relies only on protein 

(re)folding theory, which was necessary for producing IL-5 and Gly-His IL-5Rα. 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 



8 
 

 

2. Research question 
 

Is it possible to utilize cytokine natural receptors as biorecognition elements of electrochemical 

biosensors for cytokine monitoring?  

3. General Hypothesis 
 

The ectodomain of the alpha chain receptor of interleukin-5 (IL-5Rα) can be utilized as a 

biorecognition element to develop an electrochemical biosensor for IL-5 monitoring. 

4. Objectives 
 

4.1. General Objective 

 

To develop an electrochemical biosensor for the measurement of Interleukin-5 concentration 

by the use of the alpha chain of its natural receptor. 

 

4.2. Specific Objectives 

 

• To design an in-house protocol for the optimal production and isolation of IL-5/Gly-His 

IL-5Rα in cultures of E. coli. 

 

• To evaluate the surface chemistry of commercial carbon electrodes modified with the 

produced Gly-His IL-5Rα throughout the assembling of the electrochemical biosensor. 

 

• To examine the metrological features of the assembled electrochemical biosensors for 

IL-5 measurement. 

 

 

 

 

 

 

 

 

 



9 
 

5. Theoretical Framework 

5.1. Chapter 1. Protein Production by Refolding 
 

The lateral chains of amino acids explain the variable physical chemical response to the 

microenvironment changes. The astonishing pleiotropy of some proteins reveals the effects of 

the surrounding environment over the structure-function dynamics, which in turn restricts and 

dictates the 3D folding information found in the 1D primary structure of proteins. The latter 

presents important challenges to any effort for protein production, either for research or even 

therapeutic purposes. This section highlights the fundamentals of protein folding and the 

difficulty of experimentally reproducing it. Then, even while natural receptors may be useful as 

biorecognition components in electrochemical biosensing, optimizing in-house synthesis and 

achieving additional scalability could be extremely challenging. 

 

5.1.1. From 1D to 3D, the in vitro protein folding problem 

 

During the ‘50s and ‘60s decades, the Nobel laureates Stanford Moore, William Howard Stein, 

and Christian Boehmer Anfinsen demonstrated that the enzyme's single chain arrangement has 

a unique amino acid sequence (primary structure) folded over itself (secondary structure), 

resulting in a compact functional shape (tertiary structure) [1] [2] [3] [4]. Anfinsen´s group 

especially highlighted the relationship between the one-dimensional (1D) information 

contained in the sequence of amino acids and the final three-dimensional (3D) structure of 

proteins. Their findings suggest that the simple amino acid sequence is, alone, inadequate to 

carry out any function; however, the functional region of a protein resides in a short region or 

“active center” [5]. In other words, the secondary structure was important only in the active 

center but not necessarily in the entire protein. The observation that the reduction-reoxidation 

experiments (i.e., refolding) recover the activity of some proteins confirmed the 

thermodynamical favorability of tertiary structure imposed by the sequence [6]. 

 

Besides the sequence-structure-function relationship, Anfinsen´s work introduced a general 

refolding workflow from reduced proteins [7]. Kinetics studies directed to elucidate the 

refolding mechanism under the achieved workflow gave rise to the proposal that the in vitro 

process was relatively random, leading to disulfide bridges that may undergone reshuffling 

according to the thermodynamics limits imposed by the primary structure in the aqueous 

oxidizing environment [8]. Then, to improve the refolding yield, relatively low protein 

concentrations are required to achieve the maximum functional protein quantities in the 

shortest possible time. Adding redox couples (e.g., glutathione, β-mercaptoethanol, L-cysteine) 



10 
 

or some microsomal enzymes into the refolding buffer provides evidence that incorrect 

coupling of disulfide bridges may be exchanged, improving the in vitro process [9] [10] [11]. 

Altogether, to explain how a protein 

reaches the biological functional 

geometry in solution, Anfinsen 

proposed the “thermodynamic 

hypothesis.” This hypothesis states 

that protein folding is a 

thermodynamically mediated process 

wherein the conformational 

restrictions to the tertiary structure 

are set by the primary structure solely  

[12] [13] [14]. As a result, when a 

protein interacts with its physiological 

environment (i.e., solvent, pH, 

solutes, ionic strength, redox 

potential, and temperature), unique 

or even very few folding possibilities 

are reached, which in turn have the 

lowest Gibbs free energy and its most 

stable conformation. While most of 

the measurements before the ‘90s 

showed a sudden two-state folding 

mechanism (i.e., unfolded → folded), 

Anfinsen proposed that, in vitro, the 

folding proceeds through a 

cooperative process. 

 

All in vitro folding phenomenon is 

summarized in Figure 1. Briefly, a 

region with a specific local amino acid 

backbone undergoes very 

thermodynamically favored short- 

and medium-range interactions, 

bringing a localized folding nucleus. 

Then, the nucleation is followed by a 

rapid cascade of large-range interactions that immediately collapse the structure in a native-like 

intermediate called the molten globule. The final native shape is achieved after the disulfide 

 

Figure 1. Protein-energy landscape for in vitro folding process. 
Conformational space is imagined as a vast landscape with hills, 
valleys, and funneled abysses. The representation is a map 
showing the potential energy associated with each protein 
conformation. Then, as Anfinsen predicts, the top positions have 
high conformational energy (i.e., high entropy). Inversely, the 
bottom ones at the basement of the funnel-like abyss follow 
Gibbs free energy minimums (i.e., low entropy). When an 
unfolded protein is added to a refolding buffer, smaller 
conformational units, the foldons, initiate the folding (1). Once 
some secondary structures are reached, medium- and long-
range interactions occur, leading to the molten globule 
intermediate (2). As Levinthal predicts, in some cases, the faster 
kinetic folding pathway (yellow line) drives to the native folded 
shape (3), establishing a folded ↔ unfolded equilibrium (yellow 
line with a double arrowhead). Kinetically, other possible 
pathways drive the folding to intermediates or even misfolded 
and unfolded metastable states (orange lines). The apparition of 
alternative pathways to irreversible aggregation (dark-reddish 
line) depends on the protein concentration and the refolding 
buffer conditions. Note that the protein hydration level is an 
inverse function of the folding state (i.e., hydrophobic effect). 
Taken and modified with permission from [15]. 

 



11 
 

bridges are formed in the correct places. Because secondary structures exhibit an intrinsically 

cooperative tendency, some researchers have opted to refer to "foldons," the smaller, 

independently cooperative units (Figure 1, step 1) [16] [17]. The last scenario could explain the 

sudden transition between unfolded and refolded shapes in a very narrow ensemble of 

denaturing conditions (i.e., reducing agent, chaotropic salt, pH, temperature) [18] [13] [14]. 

 

Alternatively to Anfinsen's thesis, Cyrus Levinthal proposed the “kinetic hypothesis” (Figure 1, 

yellow lines). Levinthal argued that any protein's lowest configurationally achieved energy as an 

evolutionary result rather than a necessary fact of its associated physical chemistry [19]. The 

natural protein's structure is merely one of several metastable states where the configurational 

energy is at a local minimum but not always at an absolute minimum [19]. In consequence, 

proteins do not explore their entire conformational space during folding; instead, they follow 

particular easier and faster pathways around any conformational energy barrier until they 

reach their ultimate shape (Figure 1, yellow lines) [19] [20]. Once the final metastable form is 

obtained, the remarked configurational energetic barriers could hamper its equilibrium with 

unfolded states and further aggregation [20].  

 

Folding models based on the kinetic and thermodynamic studies of the chain entropies showed 

an evident correlation between the level of structural compaction, the low-energy 

conformational ensembles, and the smaller quantity of conformations [21]. The last is why the 

protein-folding energy landscapes look like a funnel (Figure 1). Few compact low-energy 

conformations exist (i.e., native-like shapes) compared to opened ones in any protein-folding 

energy landscape, as Afinsen's prediction suggested [21]. Funneled energy landscape advises 

that, for the same protein, more than one microscopic folding pathway may exist to the same 

native functional configuration [21].  

 

 

5.1.2. Proteostasis: In-vivo protein folding dictates the Inclusion Bodies (IBs) formation  

 

The intrinsic nature of proteins exposed to their physiological environment imposes 

conformational energetic barriers to achieving their metastable functional shape [17]. The last 

is hacked by cell machinery, forcing the synthesized proteins to adopt unique steric dispositions 

through predefined folding pathways (Figure 2) [17] [22] [15] [23]. Briefly, it occurs as follows: 

1) Co-translational folding occurs in the exit channel of the large ribosome subunit; it provides 

enough space only for helices or small tertiary structure elements [15] [24]. 2) The hydrophobic 

effect occurs after the protein leaves the ribosome channel; the surrounding water molecules 

promote the burial and clustering of hydrophobic amino acids in the protein's interior [25]. 3) 

Ribosome-binding chaperones (RBC) called Trigger Factor (TF) in bacteria and the Nascent 



12 
 

Polypeptide-Associated Complex (NAC) in eukaryotes bind to the nascent polypeptide chain to 

relocate the improperly exposed hydrophobic amino acids to the aqueous environment [26]. 4) 

Chaperones (i.e., Hsp 70 in eukaryotes and Dnak in bacteria) recruit the partially folded 

intermediates that leave the ribosomes, inducing a conformational shift to their native shape 

[22] [26]. 5) Chaperonin folding machinery called "GroEL-GroES" complex in bacteria, or 

"TRiC/CCT" complex in eukaryotes, both cotranslationally and post-translationally also helps in 

protein folding [22]. 6) Prolyl isomerases, often referred to as peptidylprolyl isomerase or 

PPIase (e.g., cyclophilins, parvulin, FK506-binding proteins), post-translationally isomerize 

regions with X-prolyl peptide if the chaperones described above are unable to do it [27]. 7) 

Disulfide bond formation by protein disulfide isomerases (PDIs) helps to stabilize the protein’s 

tertiary and quaternary structures [28]. 

 

The conformational space that any protein may explore during and after folding is 

circumvented and constrained by all listed mechanisms (Figure 2) [15]. It avoids its aggregation 

and even serves as a rescue tool for an eventual unfolding. Other post-translational 

modifications, such as glycosylation, may also be required to guarantee protein stability [15]. 

The organizational principles of cytosolic chaperone pathways are highly conserved. For most 

proteins too large to fold in association with the ribosome, chaperones delay the chain 

compaction and prevent misfolding [22] [23]. The last increases the folding efficiency but 

typically does not alter (or only slightly) the folding rate constants (Figure 2, see the enlarged 

area, yellow line). The abovementioned folding pathway stabilizes the folded proteins in 

kinetics traps at metastable states by a vast unfolding barrier (e.g., >26 kcal/mol). For several 

proteins (and mainly multidomain ones), the metastable native shape is not necessarily under 

the Gibbs free energy minimum, and the kinetic trapping is isolated by a small unfolding barrier 

[29] [30]. The folded proteins are continuously shielded from abnormal interactions and early 

cytosolic breakdown. The intricate web of biological mechanisms that regulate the synthesis, 

folding, trafficking, and destruction of proteins has been referred to as proteostasis or protein 

homeostasis [22]. It guarantees that the cell´s proteins maintain balance, structure, and 

function. 

 

As it may be deduced, the protein production rate during the artificial expression of 

recombinant proteins in organisms such as E. coli surpasses the proteostasis capacity (Figure 2) 

[31]. Accordingly, the inclusion bodies (IBs) are formed essentially because synthesized proteins 

are not folded (Figure 2). Intermolecular interactions and further aggregation are caused by 

both the high local concentration and the exposure of hydrophobic areas [32]. A remarkable 

feature of IBs is their capacity to aggregate structurally similar proteins. Based on the structural 

features of aggregated protein IBs may be classified as classic and non-classic. Classic IBs are 

formed during high-level expression and comprise wholly unfolded proteins that form 



13 
 

amorphous insoluble aggregates. If the structure is disordered, the classic IBs are amorphous; if 

it is organized, they are amyloid [32]. On the other hand, non-classical IBs are induced at low 

temperatures (e.g., 18-25 °C), are smaller, and contain partially or even wholly folded proteins 

[33]. Unlike classical, the non-classical IBs require a lower molar concentration of chaotropic 

salts for their complete solubilization [33]. 

 

 
Figure 2. In-vivo protein folding landscape. As is discussed in the main text, all cell machinery related to the 

translation process is directed to reach the native metastable conformation in the fastest and easiest possible 

pathway. The folding machinery is surpassed, the inclusion bodies are formed, and irreversible unfolding in natural 

conditions is induced. Taken and modified with permission from [15]. 

 

 

5.1.3. Fundamentals on protein refolding  

 

Active protein isolation helps the study of the biological functions of genes and the 

development of therapeutic drugs and biomaterials in the biotechnology industry. DNA 

recombinant technology commonly overexpresses the protein of interest to be synthesized at 

scale. Several organisms, such as bacteria, yeast, and mammalian cells, are used to do that. In 

the original report of this thesis, Escherichia coli had been used. This bacterium is a great choice 

because of its cheap price, time-saving, and scalability of production [32]. However, it is typical 

for inclusion bodies (IB) to occur after the induction of expression. IBs consist of protein 

aggregates with non-native conformations. IBs are at local minimums of Gibbs free energy that 

confers greater stability than the native ones. 

 

As inclusion bodies contain relatively pure and intact unfolded proteins, they may be used as a 

source of massive protein production in an active form. Commonly, four steps are required. The 

first is the bacteria lysis by chemical and/or physical methods followed by centrifugation. In this 



14 
 

work, the induced cells were resuspended into a lysis solution for their sonication. The function 

of the lysis solution is to enhance the lysis process while preserving the integrity of IBs. It is 

prepared by the addition of several components as a buffer (i.e., Tris) at alkaline pH to maintain 

the aggregation state of IBs; a chelating agent (i.e., EDTA) to neutralize the action of proteases 

dependents of some metals; a highly concentrated solute (i.e., NaCl, sucrose) to reduce the 

viscosity induced by the release of DNA and the membrane disruption. The next step is the IBs 

solubilization. Chaotropic salts were historically used to dissolve them (e.g., urea, guanidinium 

chloride). 

 

Some simulations suggest that urea at high concentrations alters the structure and dynamics of 

water, distorting the hydration sphere around the proteins [34]. Consequently, the hydrophobic 

effect is ameliorated, exposing the inner regions [35]. Urea may interact with polarized peptide 

groups by hydrogen bonding and stabilizing the unfolded configuration [36]. Most of the 

unfolding curves in urea show a sudden solvation between 7 and 9 M. Thus, a concentrated 

acidic solution of 8 M is commonly used. The acidic environment improves the unfolding effect 

because of the protonation of the side chains of amino acids, altering the intramolecular ionic 

and hydrogen bonds. Protein refolding is the third step once the unfolded purified protein is 

obtained. The hydrophobic areas must be kept apart during the process, as was previously 

discussed.  

 

High quantities of unfolded protein are used to achieve the refolded one to hack the yield of 

naturally occurring folding pathways artificially, following two ways. The first one is to adsorb 

the protein onto an affinity column (e.g., cobalt or nickel for his-tagged proteins). On-column 

refolding may be faster, but it may also be challenging regarding temperature maintenance 

(commonly 4 °C) and the exchange of refolding buffer. The second form is to prepare a 

refolding buffer containing customized additives and use it to dilute the unfolded protein. For 

instance, in this thesis, the dilution was made from a more potent chaotropic solution of 

guanidinium chloride to a weaker refolding buffer made with urea. In general, the following 

steps to dilution may vary depending on the complexity of the protein of interest. A unique 

dilution step may be applied if the protein has one or very few simple domains. The unfolded 

protein is directly added to the final buffer without denaturing agents. In contrast, a stepwise 

dialysis is required for complex proteins with several domains or many disulfide bonds. During 

the process, the unfolded protein is dialyzed against refolding buffers with decreasing 

concentrations of denaturing agents to avoid the kinetic trapping of improper metastable 

intermediates or their aggregation [37]. All mentioned refolding procedures are time-

consuming, and often, recovering yields of active proteins are low, as has been reviewed 

everywhere [38] [39] [40]. Additionally, a trial-and-error process is required to customize the 

conditions for refolding buffer preparation [38] [39] [40].  



15 
 

The original report of this thesis may be used to illustrate the last point. The expensive redox 

couple of reduced/oxidized glutathione was used to prepare the IL-5 refolding buffer. As IL-5 

has a relatively simple dimeric structure stabilized by two interchain disulfide bonds, one-step 

dialysis through overnight incubation is enough. Alternatively, the cheaper redox couple 

CuSO4/L-cysteine was used during the stepwise dialysis refolding applied to Gly-His IL-5Rα. 

Given at least two disulfide bridges and the wide required folded patches to interact with IL-5, 

the refolding process of IL-5Rα required four overnight incubations in different dialysis buffers. 

 

To conclude, gel filtration chromatography is the last step of protein production by refolding 

[39] [40]. Several works are improving automation to encounter general protocols to find the 

optimal refolding conditions [32] [41]. The difficulty of refolding any protein depends on the 

size and the number of contiguous hydrophobic residues [42] [43]. Proteins with high molecular 

weights and more than one domain (i.e., functional units commonly rich in hydrophobic 

residues) tend to have more intermediates, predisposing them to aggregate, as was discussed 

in the last two sections [42] [43]. 

 

 

References 

 

[1]  C. H. Hirs, S. Moore and W. H. Stein, "The Sequence of the Amino Acid Residues in 

Performic Acid-oxidized Ribonuclease," J Biol Chem, vol. 235, pp. 633-647, 1960.  

[2]  C. Anfinsen, M. Flavin and J. Farnsworth, "Preliminary studies on ribonuclease structure," 

Biochim Biophys Acta, vol. 9, no. 4, pp. 468-469., 1952.  

[3]  C. B. Anfinsen, R. R. Redfield, W. L. Choate, J. Page and W. R. Carroll, "Studies on the gross 

structure, cross-linkages, and terminal sequences in ribonuclease," J Biol Chem, vol. 207, 

no. 1, pp. 201-210, 1954.  

[4]  R. R. Redfield and C. B. Anfinsen, "The structure of ribonuclease. II. The preparation, 

separation, and relative alignment of large enzymatically produced fragments," JBC, vol. 

221, no. 1, pp. 385-404, 1956.  

[5]  C. B. Anfinsen, "The inactivation of ribonuclease by restricted pepsin digestion," Biochim 

Biophys Acta, vol. 17, no. 4, pp. 593-594, 1955.  

[6]  C. B. Anfinsen, M. Sela and J. P. Cooke, "The reversible reduction of disulfide bonds in 

polyalanyl ribonuclease," J Biol Chem, vol. 237, pp. 1825-1831, 1962.  



16 
 

[7]  C. B. Anfinse and E. Haber, "Studies on the reduction and re-formation of protein disulfide 

bonds," J Biol Chem, vol. 236, pp. 1361-1363, 1961.  

[8]  C. B. Anfinsen, E. Haber, M. Sela and F. H. White Jr, "The kinetics of formation of native 

ribonuclease during oxidation of the reduced polypeptide chain," Proc Natl Acad Sci U S A, 

vol. 47, no. 9, pp. 1309-1314, 1961.  

[9]  D. Givol, R. F. Goldberger and C. B. Anfinsen, "Oxidation and Disulfide Interchange in the 

Reactivation of Reduced Ribonuclease," J Biol Chem, vol. 239, pp. PC3114-3116, 1964.  

[10]  E. Haber and C. B. Anfinsen, "Side-chain interactions governing the pairing of half-cystine 

residues in ribonuclease," J Biol Chem, vol. 237, pp. 1839-1844, 1962.  

[11]  D. Govil, F. Delorenzo, R. F. Goldberger and C. B. Anfinsen, "Disulfide Interchange and the 

Three-Dimensional Structure of Proteins," Proc Natl Acad Sci U S A, vol. 53, no. 3, pp. 676-

684, 1965.  

[12]  C. B. Anfinsen, "Some observations on the basic principles of design in protein molecules," 

Comp Biochem Physiol A, vol. 4, no. 2-4, pp. 229-240, 1962.  

[13]  C. B. Anfinsen, "The formation and stabilization of protein structure," Biochem J, vol. 128, 

no. 4, pp. 737-49, 1972.  

[14]  C. B. Anfinsen, "Principles that govern the folding of protein chains," Science, vol. 181, no. 

4096, pp. 223-230, 1973.  

[15]  I. Sorokina and A. Mushegian, "Modeling protein folding in vivo," Biol Direct, vol. 13, no. 1, 

p. 13, 2018.  

[16]  M. M. K. M. M. L. E. S. Maity H, "Protein folding: the stepwise assembly of foldon units," 

Proc Natl Acad Sci U S A, vol. 102, no. 13, pp. 4741-4746, 2005.  

[17]  K. S. Hingorani and L. M. Gierasch, "Comparing protein folding in vitro and in vivo: 

foldability meets the fitness challenge," Curr Opin Struct Biol, vol. 24, pp. 81-90, 2014.  

[18]  J. London, C. Skrzynia and M. E. Goldberg, "Renaturation of Escherichia coli tryptophanase 

after exposure to 8 M urea. Evidence for the existence of nucleation centers," Eur J 

Biochem, vol. 47, no. 2, pp. 409-415, 1974.  

[19]  C. Levinthal, "Are there pathways for protein folding?," J Chim Phys, vol. 65, pp. 44-45, 

1968.  



17 
 

[20]  C. B. Anfinsen and H. A. Scheraga, "Experimental and theoretical aspects of protein 

folding," Adv Protein Chem, vol. 29, pp. 205-300, 1975.  

[21]  K. A. Dill and J. L. MacCallum, "The protein-folding problem, 50 years on," Science, vol. 338, 

no. 6110, pp. 1042-1046, 2012.  

[22]  D. Balchin, M. Hayer-Hartl and F. U. Hartl, "In vivo aspects of protein folding and quality 

control," Science, vol. 353, no. 6294, p. aac4354, 2016.  

[23]  C. A. Waudby, Dobson, C. M and J. Christodoulou, "Nature and Regulation of Protein 

Folding on the Ribosome," Trends Biochem Sci, vol. 44, no. 11, pp. 914-926, 2019.  

[24]  F. Wruck, P. Tian, R. Kudva, R. B. Best, G. von Heijne, S. J. Tans and A. Katranidis, "The 

ribosome modulates folding inside the ribosomal exit tunnel," Commun Biol, vol. 4, no. 1, 

p. 523, 2021.  

[25]  R. L. Baldwin and G. D. Rose, "How the hydrophobic factor drives protein folding," Proc 

Natl Acad Sci U S A, vol. 113, no. 44, pp. 12462-12466, 2016.  

[26]  E. Deuerling, M. Gamerdinger and S. G. Kreft, "Chaperone Interactions at the Ribosome," 

Cold Spring Harb Perspect Biol, vol. 11, no. 11, p. a033977, 2019.  

[27]  P. A. Schmidpeter and F. X. Schmid, "Prolyl isomerization and its catalysis in protein folding 

and protein function," J Mol Biol, vol. 427, no. 7, pp. 1609-1631, 2015.  

[28]  M. Narayan, "The Structure-Forming Juncture in Oxidative Protein Folding: What Happens 

in the ER?," Adv Exp Med Biol, vol. 966, pp. 163-179, 2017.  

[29]  K. Pauwels, I. Van Molle, J. Tommassen and P. Van Gelder, "Chaperoning Anfinsen: the 

steric foldases," Mol Microbiol, vol. 64, no. 4, pp. 9917-22, 2007.  

[30]  A. E. Varela, K. A. England and S. Cavagnero, "Kinetic trapping in protein folding," Protein 

Eng Des Sel, vol. 32, no. 2, pp. 103-108, 2019.  

[31]  M. S. Hipp and F. U. Hartl, "Interplay of Proteostasis Capacity and Protein Aggregation: 

Implications for Cellular Function and Disease," J Mol Biol, vol. 436, no. 14, p. 168615, 

2024.  

[32]  A. Bhatwa, W. Wang, Y. I. Hassan, N. Abraham, X. Z. Li and T. Zhou, "Challenges Associated 

With the Formation of Recombinant Protein Inclusion Bodies in Escherichia coli and 

Strategies to Address Them for Industrial Applications," Front Bioeng Biotechnol, vol. 9, p. 



18 
 

630551, 2021.  

[33]  A. K. Upadhyay, A. Murmu, A. Singh and A. K. Panda, "Kinetics of inclusion body formation 

and its correlation with the characteristics of protein aggregates in Escherichia coli," PLoS 

One, vol. 7, no. 3, p. e33951, 2012.  

[34]  Z. Su and C. L. Dias, "Molecular interactions accounting for protein denaturation by urea," 

J. Mol. Liq, vol. 228, pp. 168-175, 2016.  

[35]  J. Almarza, L. Rincon, A. Bahsas and F. Brito, "Molecular mechanism for the denaturation of 

proteins by urea," Biochemistry, vol. 48, no. 32, pp. 7608-7613, 2009.  

[36]  B. J. Bennion and V. Daggett, "The molecular basis for the chemical denaturation of 

proteins by urea," Proc Natl Acad Sci U S A, vol. 100, no. 9, pp. 5142-5147, 2003.  

[37]  A. E. Varela, J. F. Lang, Y. Wu, M. D. Dalphin, A. J. Stangl, Y. Okuno and S. Cavagnero, 

"Kinetic Trapping of Folded Proteins Relative to Aggregates under Physiologically Relevant 

Conditions," J Phys Chem B, vol. 122, no. 31, pp. 7682-7698, 2018.  

[38]  H. Yamaguchi and M. Miyazaki, "Refolding techniques for recovering biologically active 

recombinant proteins from inclusion bodies," Biomolecules, vol. 4, no. 1, pp. 235-251, 

2014.  

[39]  A. Singh, V. Upadhyay, A. K. Upadhyay, S. M. Singh and A. K. Panda, "Protein recovery from 

inclusion bodies of Escherichia coli using mild solubilization process," Microb Cell Fact, vol. 

14, p. 41, 2015.  

[40]  A. Nabiel, Y. Yosua, S. Sriwidodo and I. P. Maksum, "Overview of refolding methods on 

misfolded recombinant proteins from Escherichia coli inclusion bodies," J App Biol Biotech, 

vol. 11, no. 3, pp. 47-52, 2023.  

[41]  Y. Wang, N. v. Oosterwijk, A. M. Ali, A. Adawy, A. L. Anindya, A. S. S. . Dömling and M. R. 

Groves, "A Systematic Protein Refolding Screen Method using the DGR Approach Reveals 

that Time and Secondary TSA are Essential Variables," Sci Rep, vol. 7, p. 9355, 2017.  

[42]  C. M. Dobson, "Principles of protein folding, misfolding and aggregation," Semin Cell Dev 

Biol, vol. 15, no. 1, pp. 3-16, 2004.  

[43]  K. A. Markossian and B. I. Kurganov, "Protein folding, misfolding, and aggregation. 

Formation of inclusion bodies and aggresomes," Biochemistry (Mosc), vol. 69, no. 9, pp. 

971-984, 2004.  



19 
 

 

5.2. Chapter 2. Electrochemical Nanobiosensors as Point-of-care Testing Solution 

to Cytokines Measurement Limitations 

 

Cite: Perez D, Patiño E, Orozco J. Electroanalysis. 2022; 34(2): 184-211. DOI: 

10.1002/elan.202100237 

Abstract: Most cytokines are present in reduced amounts in body fluids due to their biological 

features of production, release, and action mechanisms. The required time between sampling 

and measurement is critical for diagnosis and treatment. Electrochemical nanobiosensors offer 

the possibility to be tailor-made and cost-affordable, producing direct and rapid readouts with 

low sample volume, explaining their feasibility in timely measurements and potential in 

designing unique and multiplexed Point-Of-Care (POC) testing platforms. This review 

summarizes and discusses the measurement limitations of the standard methods and the 

recent progress on electrochemical nanobiosensors as a plausible alternative to measuring 

them. 

 

5.2.1. Introduction 

 

The primary response to any physiological insult from a microorganism or another pathological 

condition relies on the abrupt release of cytokines [1]. The severe effects of the sudden and 

potent activity of massively released cytokines in the clinical context are critical complications 

that may drive a demise. The required time between sampling and measurement is critical to 

apply any treatment. Point-of-care (POC) systems are medical assays done in place of patient 

care without a clinical lab infrastructure. Also known as the bedside, such devices provide 

prompt results, improving the diagnosis and treatment of secondary care. They may integrate 

electrochemical nanobiosensors to have an output signal in a concentration-dependent manner 

with high sensitivity and specificity. Then, POC systems based on electrochemical 

nanobiosensors offer a potential solution to cytokine measurement limitations. 

Cytokines are small glycoproteins (~ 6–70 kDa) that control immune response orchestration and 

cell communication. These macromolecules are soluble in body fluids such as serum, plasma, 

urine, saliva, sweat, and tears. Some of them modulate the most important nonspecific innate 

response, called inflammation, by both its induction (e.g., IL-1, IL-6, TNFα) or its blockade (e.g., 

IL-10, TGFβ). Other cytokines regulate the adaptive immune response, either mediating the 

specification of immunity during antigen presentation (e.g., IL-2, IL-4, IL-12, IL-23) or as 



20 
 

signature cytokines released by the different secretion profiles of T helper cells (e.g., Th1, Th2, 

Th17). Historically, cytokine classification has been too complicated because no systematic 

evolutionary criterion has been applied [2]. Classical textbooks dispose them into six categories: 

interleukin-1 (e.g., IL-1, IL-1Ra, IL- 18, IL-33), hematopoietin (class I cytokines e.g., IL-2, IL- 4, IL-

5, IL-6, IL-13, IL-23), interferon (class II cytokines, e.g., IFN-α, IFN-β, IFNγ, IL-10), tumor necrosis 

factor (TNF, e.g., TNF-α, TNF-β), interleukin-17 (e.g., IL-17A, IL-17B) and chemokine families 

(e.g., CXCL2, IL-8/CXCL8, CCL2, CCL5). Homeostatic cytokines are produced all the time, having 

basal levels in their soluble form even when there is no stimulus; meanwhile, those that 

potentiate or inhibit the inflammatory response are called pro- or anti-inflammatory cytokines, 

respectively. Most attempts to develop POC systems based on electrochemical nanobiosensors 

have been for pro-inflammatory cytokines, given their role as diagnostic and prognostic 

biomarkers. In principle, they may be used as devices to follow a chronic condition, such as the 

glucose biosensor for diabetic patients. 

The commercial methodologies applied in cytokines testing are not harmonized but are also 

intrinsically limited by cytokines’ physiology (see below). Recent advances in electrochemical 

nanobiosensors in POC-testing formats put them in a promising place as possible solutions to 

those physiological limitations [3]. Two recently published reviews discussed several 

electrochemical biosensors’ elements for assessing pro-inflammatory cytokine levels during the 

cytokine storm produced by SARS-CoV-2 or other non-infectious diseases, such as 

neurodegeneration and cancer [4] [5]. Yet, in this review, we discuss how the hindrance 

imposed by the biology of cytokines' behavior, together with technical and methodological 

limitations for their analysis, may be overcome by developing electrochemical nanobiosensor-

based POC systems, emphasizing the screen-printed electrode (SPE) platforms, but detailing 

some distinct electrochemical formats. Additionally, some preanalytical considerations are 

presented to stimulate the discussion about the urgent requirements for harmonized cytokine 

measurement and the opportunities that offer POC testing in the context of ASSURED criteria 

(i.e., Affordable, Sensitive, Specific, User-friendly, Rapid and robust, Equipment-free, and 

Deliverable to end-users). 

 

5.2.1.1. Physiology of Cytokines Behavior: A Biological Limitation for their Measurement 

 

Any cytokine must reach its cognate receptor at the appropriate stage to complete the 

associated biological function. To do that, cytokines “exploit” almost all cell communication 

strategies. However, the pivotal issue about their physiological behavior as communication 

mediators is related to the fact that their effects tend to be restricted to a given 

microenvironment, acting as intra-, yuxta-, auto-, and paracrine factors compared with their 



21 
 

endocrine effects (Figure 1). The 

advantages, challenges, and technical 

implications of cytokine measurement may 

be deduced from brief examples. 

Counterintuitively, some cytokines as 

alarmins (e.g., IL-1α, IL-33, HMGB1), are 

maintained intracellularly to accomplish 

essential functions once they are produced 

(intracrine in Figure 1). For instance, IL-33 

is retro-transported and cumulated into 

the nucleus [6], where it mediates the 

chromatin compaction by the promotion 

of the nucleosome-nucleosome 

interactions [7] and the transcription 

modulation [6] [8] [9], a skill that remains 

unconfirmed [10]. After an exogenous 

insult, IL-33 alarmin is released to the 

extracellular space and complexed with 

histones, which confers it a tunable effect, 

controlling the free IL-33 availability after 

necrosis or synergizing (i.e., additive 

effects) to initiate its proinflammatory 

effect [11] [12]. In summary, some 

cytokines’ basal distribution, like IL-33, 

suggests an intracellular role, functioning 

as intracrine factors, while they are not 

generally present in any body fluid. 

Similarly, some cytokines mediate 

physiological processes in a juxtacrine 

manner, in which either the ligand or the 

receptor must be held to the membrane 

of two interacting cells and, instead, be 

released (Figure 1). A notable example is 

the tumor necrosis factor (TNF), a 

pleiotropic cytokine (i.e., acts over more 

than one cell target) that can exert a 

myriad of roles, according to the context 

Figure 1. Physiological cytokine behavior determines their 
levels in any corporal fluid, with a special emphasis on 
blood. After sampling, it may be processed by 
conventional tests (e.g., ELISA, flow cytometry-based 
methods, Microspot) or electrochemical techniques. 
Some features are highlighted on the right side. In 
parentheses, examples do not exactly represent the 
drawn situation. See the text for more details. Some 
images were taken from free smart service medical art. 

 



22 
 

and held receptors (i.e., TNFR1 and TNFR2). TNF may exist held to a membrane (mTNF, 26 kDa) 

[13] [14], or can be actively cleaved to a plasma soluble form (sTNF, 17 kDa) [13]. The mTNF 

tends to be expressed by the monocyte/macrophage system and lymphocytes, interacting 

primarily with TNFR2 and mediating processes such as lymphocyte proliferation and activation 

[15] [16]. On the other hand, sTNF acts mainly by the union to TNFR1 and mediates either 

proinflammatory or cell death-related effects (e.g., apoptosis or necroptosis) [17]. Hence, the 

observation that cytokines’ physiological flexibility includes the juxtacrine form of 

communication highlights that some functional forms held to the membrane cannot be 

measured at a clinical laboratory with the available technology, unlike electrochemical 

techniques coupled to microfluidic systems (see below). 

Assuming a random antigen searching of immune cells into a confined space (e.g., lymph node) 

or a liquid or semiliquid diffusion medium, the extracellular cytokine concentration is 

approximately proportional to the inverse of the square of the distance from the releasing cell 

[18]. Consequently, the cytokine spread rate depends on released cytokine production and 

consumption competing processes [19]. This fact modulates the immune response by both cell 

orientation during migration and gene expression induction (e.g., polarization). The 

extracellular matrix adsorption and swarming must also be considered because they force a 

short action radius of cytokine activity. For instance, once a potential immunogen is 

phagocytosed at a given tissue (e.g., skin), the antigen-presenting cell (APC) inclines to use the 

lymph as the flow medium to migrate. This liquid circulates into lymph vessels until it reaches 

its nodes. The spatial disposition of lymph increases the probability of encountering two 

leukocytes to form an immune synapse (IS) (e.g., between classical dendritic cells, cDC, and T 

cells, LT) (Figure 1). In essence, the actively secreted chemokines serve either as a stimulus to 

extravasate (e.g., CC-chemokine ligand 19 and 21, CCL19 and CCL-21), when passing 

tangentially to cells of high endothelial venules (HEVs) or during intranodal cell migration (e.g., 

of CCL21, CXCL13) [20] [21]. 

The biological sense that explains cytokines’ close action radius is probably a quantitative 

threshold so that a massive focused secretion of a specific type of cytokines polarizes the 

signaling of the target cell towards a precise effect but, at the same time, the other 

simultaneous signals from surrounding microenvironment are lessened or ignored [19]. As a 

way of illustration, it has been proposed that the strong activation of macrophages by the union 

of high-avidity ligands (e.g., immune complexes) to the immunoreceptor tyrosine-based 

activation motif (ITAM)-coupled receptors induces transient calcium signals that block the 

cytokine receptors and synergize with Toll-like receptors (TLR) signaling. In contrast, the low-

avidity ligands (e.g., monomeric IgG) of ITAM-coupled receptors cause the inverse effect [22]. 

Another benchmark case occurs in T lymphocyte activation during IS. Here, the T cell receptor 

(TCR) binds to a peptide-charged major histocompatibility complex (pMHC), on the surface of 



23 
 

an APC, and, at the same time, costimulatory receptors (e.g., CD40-CD40L, CD80/ CD86-CD28) 

must come together yuxtacrinally (Figure 1). The pair TCR-pMHC and costimulatory complexes 

are surrounded and sealed by other molecules, allowing a cytokine-directed secretion to T cells 

into the IS, in a paracrine manner that acts as the third required signal to promote its proper 

activation and polarization during antigen presentation [23] [24] [25]. In the end, some 

cytokines do not spread significantly beyond the two interacting cells (e.g., IL-2 and IFNγ), while 

others are multi-directionally secreted (e.g., TNF, CCL3) [26]. Furthermore, evidence suggests a 

controlled release process by exo- and ectosomes during IS exchange [27]. Thus, it is important 

to investigate if the cytokines can also be included in vesicles but are not freely released [28], 

decreasing their levels in the body fluids. 

After CD4+ T cells polarize into helper specialized phenotypes (e.g., Th1, Th2, Th9, Th17), they 

are inclined to release a close group of cytokines. A tangible case is a Th2 profile. The IL-4 sign 

from the primed T cell itself (i.e., autocrine) and surrounding cells, together with the activation 

signals provided by dendritic cells (e.g., IL-25, IL-33), synergizes in a paracrine manner [29], 

leading to the polarization of the Th2 cell, while Th1 and Th17 cell responses are suppressed 

[30]. Once differentiated, Th2 synthesizes its signature cytokines (i.e., IL-4, IL-5, and IL-13). The 

IL-4, together with IL-13, induces the class switching to IgE and IgG1 over B lymphocytes [31] 

and promotes the alternative activation of macrophages [32] [33]. For its part, the wrench-like 

form by which the IL-5 receptor alpha chain (IL-5Rα) binds its ligand (i.e., IL-5) seems to trigger 

conformational changes that initiate the signaling activity of typical β chain (IL-5Rβc), up-

regulating the eosinophil metabolism and its effector mechanisms (e.g., degranulation) [34] 

[35]. 

The cytokine release also occurs in a nonspecific manner when it is induced by activation of 

innate mechanisms (e.g., TLR signaling). In this scenario, both immune and non-immune cells 

can serve as a source. Initially, these cells’ cytokine discharge acts autocrine and paracrine 

manner. However, when the immune system is significantly challenged, several cytokines, 

especially the endogenous pyrogens (e.g., IL-1, IL-6, TNFα), can act far away from being 

released in an endocrine form. 

These synergistic molecules may travel, even at a subnanomolar concentration, to 

hypothalamic thermal regulator centers, resulting in vasoconstriction, shivering, and brown 

adipose tissue activation, which in turn leads to a systemic thermal increase (i.e., fever) [36] 

[37]. In the classical endocrine pathway, the IL-6 circulates until it reaches the liver, inducing 

the hepatic production of positive acute-phase proteins (e.g., c reactive protein, CRP; serum 

amyloid A protein, SAA), but alternatively, it may act held to its soluble receptor (sIL-6R), 

leading to systemic pro-inflammatory effects [38] [39] [40] [41] [42]. Finally, there is evidence 

that TNF also has notable systemic effects, especially on the cardiovascular (e.g., bradycardia) 

and muscle systems (e.g., myocyte apoptosis) [43]. 



24 
 

Notice that the initial source may be the local tissular cells like macrophages, but during severe 

diseases, when cytokine serum levels are important, the endocrine effect synergizes with the 

local autocrine release [44]. Because of their physiological role, it may be expected that the 

cytokines with an endocrine behavior, like pyrogens, have basal monitorable levels at a steady 

state and that their systemic noxious effects could be valued as a function of their serum 

concentration increase in extreme situations (e.g., systemic inflammation). This feature 

differentiates these cytokines from hormones, being incomparably more active than the latter 

because they act at sub-pM levels [45]. Besides, some of the endocrine cytokines undergo a 

neuroendocrine control, showing greater levels in the morning (e.g., IL-1, IL-6, TNF-α, IFNγ), 

which correlates inversely with nocturnal melatonin rhythm [46] but proportionally to awake 

cortisol peak (Figure 1) [47]. Cytokines also may fluctuate according to the lifestyle, affecting, in 

turn, the circadian clock rhythm by the induction/repression of clock proteins [48] [49]. For 

example, it has been proposed that TNF-α and IL-6 may serve as mediators by which a fat diet 

(i.e., high-fat and saturated fatty acids diet) induces inflammation, feeds back, and modulates 

fundamental circadian properties of peripheral clocks [50]. There is also evidence that IL-6 is 

produced by the muscle throughout exercise, acting in an endocrine manner over the liver to 

induce hepatic glucose output and lipolysis [51] [52]. Finally, there is evidence that a fraction of 

circulating cytokines is held to α2-macroglobulin carrier protein and undergoes degradation by 

extracellular proteases (Figure 1), which decreases its availability to be tested, masking its 

actual levels [53] [54]. After physiological troubleshooting associated with immune challenges, 

the cytokine production is shut down, reaching homeostasis again. The above implies that 

cytokine measurement is useful when they accumulate, surpassing the threshold to be 

significantly functional or pathological (i.e., when cytokine levels > effective concentration for 

the half-maximum response, EC50). 

Hence, in principle, it is possible to correlate a cytokine production fingerprint to different 

phases and types of disease progression during treatment [55] [56] [57]. Nonetheless, it is 

important to consider that most of them have extremely low basal levels or even they not have 

any. Meanwhile, cytokines with endocrine roles (pyrogens mainly) have baseline ranges that 

oscillate due to multiple stimuli (e.g., stress, diet, exercise). In the first situation, detectable 

serum or plasma levels are not necessarily associated with an endocrine function but instead 

with a basal leakage or a surpass from its local action threshold, serving as disease severity 

biomarkers. Similarly, when pyrogenic cytokines reach high serum levels, a hyperinflammatory, 

harmful state is installed in the host, called a cytokine storm. Three criteria have been proposed 

to determine the latter state: (i) acute systemic inflammatory signs (i.e., fever, fatigue, 

anorexia, headache, rash, diarrhea, arthralgia, myalgia, and neuropsychiatric findings), (ii) 

secondary organ dysfunction (e.g., renal, hepatic, or pulmonary), and (iii) elevated circulating 

cytokine levels [58]. Furthermore, since their pleiotropy, most cytokines are not ideal 

biomarkers alone because an individual cytokine level in patients cannot specify a unique 



25 
 

disease (or prognosis) in comparison with levels in a control group of healthy subjects or 

treated patients [21] [59] [60]. The most reasonable way to include these molecules as 

biomarkers in situations like cytokine storms is their simultaneous multiple detections (i.e., 

multiplexing). 

 

5.2.1.2. Technical Limitations of (Pre)Analytical Procedures to Cytokines Measurement 

 

Overall, no standardized baseline cytokine levels may be associated with technical issues such 

as the absence of harmonized preanalytical and analytical protocols [61]. An important 

observation is related to the method used for sampling. At least three important phenomena 

may increase the cytokines level after sampling (Figure 1), namely, (i) their synthesis induced by 

the material of the sampling tube (e.g., polyvinyl chloride induces a marginal increase of IL-1β, 

IL-6, IL-10, and TNF-α, but markedly IL-8 and MCP-1) [62], (ii) a degranulation process during 

clotting and the further cytokine release to the serum [63], (iii) the presence of pyrogenic 

endotoxins in sampling tubes that stimulates a synthesis de novo (e.g., IL-1β, IL-6, TNF-α) [64]. It 

has been demonstrated that indwelling venous catheters also stimulate local cytokine 

production (e.g., IL-6), leading to a misinterpretation of the actual systemic levels in the patient 

[65] [66]. In contrast, both the proteases-mediated degradation, after two hours of storage, as 

well as the matrix adsorption (e.g., fibrin, α2-macroglobulin, albumin, anti-cytokine antibodies) 

can partially explain the decrease in serum compared to plasma  [67] [68]. Some studies have 

shown that the (defibrinated) plasma levels do not correlate with serum cytokine levels. Even if 

the plasma samples are obtained with different anticoagulants (e.g., ACD, heparin, EDTA), the 

cytokine levels vary significantly, regardless of whether unique or multiplexed kits are used [59] 

[62] [63] [64] [67] [69] [70]. Remarkably, measured cytokine secretion profiles were closer 

between ethylenediaminetetraacetic acid (EDTA) and acid-citrate-dextrose (ADC) used as an 

anticoagulant [59]. 

Along with the challenges associated with the cytokines’ physiological behavior, the 

commented technical limitations can partially explain the enormous interindividual variability 

reported elsewhere [63] [71]. In this line, several cytokine levels may increase (e.g., IL-1β, IL-7, 

IL-12p70) or decrease (e.g., Flt-3 Ligand) between age groups [59]. Similarly, although they did 

not report differences associable to sex, independently of sample conditions, other studies 

have shown both increases (e.g., CCL9, XCL1, CXCL11) and decreases (e.g., CXCL10, CXCL12, 

CXCL16) during ovulation, which may be interpreted as a prelude of uterus NK cell homing [72]. 

 

 



26 
 

5.2.1.3. Common Technology Used to Measure Cytokine in Clinically Relevant Conditions 

 

Many techniques oriented to measure cytokines massively have been reviewed [60] [73]. 

Outstandingly, the latter papers coincide with the potential of ultrasensitive nanobiosensors 

(i.e., fM, aM) and call about their issues in analyzing complex matrices (e.g., nonspecific binding 

(NSB) and fouling). Despite plenty of available ways to measure cytokines, testing them 

generally relies on antibody-based capture immunoassays. Enzyme-linked immunosorbent 

assay (ELISA) is the gold standard method for clinical protein measurements [74]. ELISA's most 

critical limitation is that it detects one analyte per kit. Additionally, ELISA is expensive, requires 

extensive analysis time (3–8 h), needs a relatively large sample size (10–100 μL), has a narrow 

dynamic range, and is difficult to adapt to POC use. As mentioned above, tens of cytokines 

govern the immune response, so multiplexing formats are highly desirable. Several cytokines 

are simultaneously tested in a single assay using a single sample [71] [75]. The most used 

multiplexed way to measure cytokines is the cytometric bead array (CBA) and Luminex xMAP 

technology (x=analyte, Multi-Analyte Profiling) [76] [77] [78]. Polystyrene magnetic microbeads 

(~ 6 μm) are stained with internal dyes with variable intensities, making it possible to 

discriminate them by the fluorescence (FL2) and size (FL3) parameters. CBA assays enable 

measuring ~ 30 proteins simultaneously, while xMAP can discriminate up to 100 types of 

different analytes because it uses different ratios of red and near-infrared fluorophores, 

increasing its multiplexing degree [75] [76]. Nevertheless, while CBA requires any flow 

cytometer, Luminex uses a particular machine for its specific purpose [77]. Both technologies 

use a specific protein-capturing antibody conjugated onto the bead surface. A secondary 

antibody with a specific fluorescence intensity is used for analyte detection [78]. The required 

volumes oscillate between 25 to 50 μL of serum for all analytes, compared with 50 to 200 μL of 

serum per analyte needed for a conventional ELISA. Luminex-based kits are not optimal at sub-

pg/mL concentrations but may have a greater dynamic range (1–1000 pg/mL) than ELISA [63] 

[71]. Some examples illustrate the usefulness and limitations of these multiplexed technologies 

in diverse clinical situations. 

For example, Drop assay (DA) on beads in conjunction with multiplexed Luminex technology has 

been done with small volumes of tears (5 μL) to improve the ocular disease diagnosis. In a 

similar study, the levels of the proinflammatory cytokine were measured with a very low limit 

of detection (LOD ~ 1 pg/mL) and spread detection ranges (e.g., IL-1β, 2.41–2502 pg/mL; IL-6, 

2.47–2497 pg/ mL; TNF-α, 4.98–16024 pg/mL; IFNγ 10.06–37464 pg/mL) [79]. Interestingly, in 

this study, just IL-1β was not detected in all of the 1000 healthy participants, suggesting very 

low or not basal levels of it in tears, which correlates with other studies on the serum from 

healthy individuals [80] [81] but contrasts with some exposure to industrial pollutants [82]. For 

instance, a study that used four types of tests, two based on the Luminex 100 system and two 



27 
 

based on CBA, indicated that IL-1β was detected in almost all supernatants of isolated white cell 

cultures from healthy Spain and Mozambique participants (endemic for malaria), that were 

exposed to Plasmodium falciparum antigens [83]. This observation agrees with the proposed 

early role in immunity response for IL-1β and suggests that its levels in the blood can be 

correlated with disease severity [84], but raises the question whether the sensitivity of the used 

techniques was sufficient to detect it. Bacterial infection severity also has been correlated with 

IL-1β and other pro-inflammatory cytokines (i.e., IL-6, IL-7, IL-8, IL-10, IL-13, TNF-α, IFN-α, MCP-

1). It suggests that their levels significantly increased in septic shock in comparison with severe 

sepsis when they were measured with a 17-multiplexed kit (Bio-Rad), which in turn showed a 

good correlation with ELISA measurements (r= 0.815; P < 0.001) [85]. There are also approaches 

to characterise blood-related diseases with a chronic course, accordingly to the cytokine levels 

[57]. For instance, by the use of a 12-plexed assay, a study showed that it was possible to 

discriminate the similar symptoms of patients with secondary polycythemia (SP) and 

polycythemia vera (PV) from the fact that SP patients showed decreased plasma levels of the IL-

17A, IFNγ, IL-12p70 and TNF-α oncoinflammatory factors, in comparison to PV patients [86]. In 

contrast, a 25-plexed bead array done with blister fluids from patients with complex regional 

pain syndrome type 1 failed to detect several cytokines (e.g., IL-1β, IL-2, IL-5, IL-7, IL-15, IFNγ) 

that were detected by more sensitive ELISA kits [87]. Immulite® is another commercial bead 

multiplexed technology for determining cytokine levels that relies on chemiluminescence. It has 

shown high run precision, short incubation times, and calibration stability (e.g., two weeks), but 

it needs relatively high volumes (e.g., 350 μL), which should not be ideal, mainly for pediatric 

samples [88]. The microspot array is another method based on printing many spots on 96-well 

plates made with different materials (e.g., polylysine and aminopropyl silane, epoxy silane-

treated glass, polyvinylidene difluoride, and nitrocellulose- or nylon-based membranes) [75] 

[89]. This technology enables relatively easy in-house tailoring and its multiple cytokine 

measurement reliability in fluids such as plasma, serum, urine, conditioned medium, tissue, and 

cell culture lysates [90] [91] [92] [93]. Besides, microspoting maintains ELISA specificity, has a 

high throughput, and permits the same antibody label versatility (e.g., chemiluminescence, 

fluorescence). In contrast, compared with bead-based methodologies, it requires high volumes 

and processing time and shows variations of ~ 10 % when made at home [92]. For instance, a 

microspot-based platform showed an 810-multiplexing degree but did require 200 μL of the 

sample, > 3 h to be processed, and its coefficient of variation (CV) oscillated between 5–15 % 

[94]. 

 

 

 



28 
 

5.2.2. Electrochemical Nanobiosensors and their Potential Advantages in Cytokines 

Determination 

 

Most multiplexed commercial tests for cytokine determinations are based on beads and multi-

well plates, as those commented before, and have been used for massive studies (e.g., clinical 

trials, epidemiologic studies) [59] [63] [69] [81] [83]. Nevertheless, nanobiosensors can be easily 

used in a unique or multiplexed shot, 

potentially being implemented in POC 

testing. It allows one to estimate the 

disease severity in a patient or follow the 

behavior after therapy and intervention. 

Like other biosensors, electrochemical 

nanobiosensors function through (i)  

recognition of the analyte, (ii) signal 

transduction, and (iii) measurable signal 

readout. The first process relies on the 

bioreceptor-analyte interaction (Figure 2). 

The bioreceptor is any biological entity, 

usually live-derived (e.g., proteins as 

antibodies or cognate receptors, glycans, 

nucleic acids, whole cells), linked to an 

electrode previously modified with a 

nanostructured material or composite. 

When the analyte is recognized, 

physicochemical changes occur on the 

transducer‘s sensing interface, interpreted 

as a measurable signal in a concentration-

dependent manner [95]. Electrochemical 

biosensors measure changes in the 

transducers’ electrochemical properties, 

including electron transfer and charge 

accumulation. The most common 

arrangement to measure such changes is 

the typical three-electrode cell, a frequent 

format in commercial SPEs. It uses a central 

working electrode (WE), where the 

biorecognition element is linked, and the 

Figure 2. General components and strategies for 
assembling electrochemical nanobiosensors based on 
screen-printed electrodes (SPEs). See the text for a 
detailed discussion. CE: Counter electrode; WE: Working 
electrode; RE: Reference electrode; NPs: Nanoparticles; 
SAMs: Self-assembled monolayers; AF: Anntibiofouling; 
NSB: Nonspecific binding; POC: Point-Of-Care; SWV: 
Square wave voltammetry; DPV: Differential pulse 
voltammetry; CA: Chronoamperometry; EIS: 
Electrochemical impedance spectroscopy. Some images 
were taken from free smart medical art. 



29 
 

bioreceptor-analyte interaction occurs. Ideally, the WE is made with a conductive (or 

semiconductive) material (e.g., Pt, Au, C, semiconducting polymers), modified with 

nanostructures in the case of nanobiosensors (e.g., nanoparticles (NP), graphene, carbon 

nanotubes). The reference electrode (RE) has a very stable equilibrium potential (e.g., AgCl/Ag) 

to control the WE’s potential. This explains why the applied voltage is typically reported 

compared to a specific RE (Figure 2). 

The auxiliary or counter electrode (CE), also built with inert materials, aims to complete the last 

half-reaction in a controlled manner, closing the circuit. After an electrode arrangement is 

immersed in a supporting electrolyte solution, the power source (i.e., 

potentiostatic/galvanostatic electrochemical workstation) is adjusted depending on the 

potential window of reactants and the desired measurement type. A sufficiently concentrated 

electrolyte is required to abolish migration (i.e., ionic solute movements by the action of an 

electric field). Relative to the analyte concentration in solution, high electrolyte concentration 

ensures that it is statistically more probable that the electrolyte migrates to the electrode 

surface for charge balance. The applied potential (E) induces a heterogeneous electron transfer 

(hET) from the WE to the CE, known as cathodic/anodic current (i). The electron flow provides 

the energy to oxidize or reduce electroactive molecules onto WE and CE. The WE redox 

phenomena are the only ones studied because the CE and RE events are so controlled. A tag is 

usually needed to produce a measurable hET, also known as faradaic current. Yet, free-label 

platforms are based on the fact that the bioreceptor-analyte union changes the transducer 

platform‘s electrical properties. Electrochemical biosensors can be classified depending on the 

signal measured on the electrode from the biorecognition event, which may be amperometric, 

potentiometric, conductometric, or impedimetric [96]. Among all interfacial electrochemical 

techniques, those controlling the potential while the current is measured (i ≠ 0) are spread and 

applied to tagged systems. Because the three-electrode conformation is the most used 

arrangement, transduction for the cytokine biorecognition event onto its WE surface is usually 

determined by voltammetric techniques such as chronoamperometry, differential pulse 

voltammetry (DPV), or square wave voltammetry (SWV). Potentiometry and impedance-based 

techniques (e.g., electrochemical impedance spectroscopy, EIS; alternating current 

voltammetry, ACV) are mainly used in two-electrode label-free platforms. Even though the 

latter techniques tend to be less applied with SPEs, they showed their usefulness in cytokines 

detection quickly and in real-time, as will be discussed. 

The SPE is one of the most common transducer platforms used to construct electrochemical 

nanobiosensors. They may be commercially acquired and offer the possibility to bespoke it in-

home relatively cheaply and straightforwardly [97]. The manufacturing process uses a mesh 

screen mask with the desired pattern over the printing substrate, such as ceramic, glass, 

transparent flexible plastic, or paper. The latter substrate is attractive because its porosity 



30 
 

improves both the usable surface and the liquid‘s wicking rate, depending on its pore size and 

thickness. The Whatman grade 1 chromatographic filter paper is widely used because it is 

almost totally composed of α-cellulose (> 98 %), is ashless, and has a uniform pore size with an 

ideal low thickness. Commented features put aside significant amounts of required 

components (e.g., ink, wax) and make it easy to penetrate, optimally outlining the biosensor's 

hydrophilic electrode zone and surrounding hydrophobic zone. The electrodes relying on paper 

can also improve the wettability either in the 2D (stacked) and 3D (folded, origami-like) 

formats, showing advantages to eliminate interferents and separate the sampling from 

electrochemical measurements, respectively [98]. Important features and principal formats of 

paper-based SPEs for cytokine measurement were recently reviewed [99]. When the substrate 

material is selected, a rubber squeegee is passed over the mesh several times to force a liquid 

paste (i.e., ink) into contact. Mandatory components included in the electrode ink are (i) a 

powdered metallic (e.g., Ag, Au, Pt) or nonmetallic (e.g., graphite) conductor, (ii) a solvent (e.g., 

terpineol, 2-ethoxyethanol, cyclohexanone, ethylene glycol) that furnish an apt printing 

viscosity as well as improve the volatility for thermal curing, and (iii) a substance that refines 

the mechanical and binding properties (e.g., glass powder, resins, cellulose acetate) to the 

substrate [100]. Dielectric and non-conductive inks are also printed as layers between 

electrodes to eliminate the interference. Ultimately, the deposited ink is cured by heat or UV 

light [101]. Even though patterns obtained by screen-printing resolve 30–100 μm, it is less than 

inkjet (15–100 μm) and aerosol-jet (10 μm), which are more amenable for miniaturization. In 

short, SPEs are widely used because they are cost-effective and can be scaled up for mass 

production, maintaining their electrochemical properties. The manufacturing process may be 

tailored to include other techniques (e.g., microfluidics, modification with nanostructures), 

refining features such as resolution and sensitivity, decisive features in cytokine interrogation, 

and reviewed in the following sections. Moreover, the required modules to connect commercial 

or homemade SPEs, for data acquisition and the corresponding potentiostats/galvanostats are 

commercially available in portable formats to be used with a computer or a smartphone 

interface, facilitating (and accelerating) the POC testing approach (Figure 2). 

 

5.2.2.1. Nanostructures as Enhancers of the Working Electrode Features 

 

Understanding structures at the nanoscale (< 100 nm) allows designing and synthesizing 

platforms over the working electrodes to enhance their performance. The optical, electrical, 

and magnetic properties may vary depending on the nanoscale‘s size, shape, and surface 

chemistry [102]. For instance, by the variation of gold nanoparticle (AuNP) shape (e.g., 

nanospheres, nanostars, nanocubes, nanorods), a report proved that they differently absorb 

light in the presence of several bacteria (i.e., E. faecalis, E. coli, P. aeruginosa, and S. aureus). 



31 
 

Then,  the last was used as a criterion for linking light to the presence of a specific bacterium in 

the sample [103]. Yet, the interest falls over magnetic and electric properties for nano-

electrochemical devices. NPs can increase the available surface for biomolecule immobilization 

and the associated electroactive area of WE, leading to lower “overpotentials” and higher 

current densities in the resultant electrochemical biosensors. A large area also intensifies the 

scale of the redox conversion signal and the associated sensitivity, so the LOD is also improved 

if it is larger than the noise increase. The higher resolution power conferred by the 

nanostructures is explained by their rapid hET rate capacity, featured by improved signals [104]. 

Regarding the shape, spherical NPs are virtually the only ones applied for cytokine biosensing 

platforms described in this review (Table 1) and others recently published [4] [105]. Here, it is 

interesting to notice that multiplexed affinity biosensors (i.e., aptameric or immune sensors) 

based on wire-like micromotors may be used to accelerate the required testing time and to 

improve the LOD and LDR of the assays, as proven to other analytes [106] [107]. Properties of 

magnetic NPs were applied mainly during analyte pre-concentration steps in electrochemical 

biosensors for cytokine sensing, with the consequent elimination or amelioration of the matrix 

effect (i.e., fouling and NSB) discussed below. 

There are three main ways to modify the SPEs with a nanoparticulated system [108]: (i) drop-

casting is the easiest because it requires drop addition followed by drying selected NPs onto the 

WE. Carbonaceous nano-materials functionalized with NPs (or nanocomposites) may be 

obtained ex-situ to control the final size better, avoiding agglomeration during drying. (ii) 

Electrodeposition induces the NPs formation by the application of a fixed potential 

(potentiostatic) or a constant negative current (galvanostatic), which reduces the precursor 

reagent (usually a metallic salt) up to achieving a zero valence. The tuning of potential or 

current is made accordingly to the used material and the desired size and shape of NPs. (iii) Ink 

mixing and printing imply the combination of NP precursors with ink before curing. Since the 

ink-mixing manufacturing procedure implies all steps associated with the platform, it also 

requires a very sensitive control of variables to ensure its reproducibility (e.g., curing T°, mixing 

recipe, mixing methodology) to avoid NPs agglomeration. Therefore, like most cytokine 

detection platforms, generally preferred methodologies are drop-casting and electrodeposition 

after electrode aeration. As part of cytokine assessment, the synthesis of NPs includes sources 

such as noble metals (e.g., Ag, Au, Pt), Ruthenium (Ru), Nickel (Ni), Copper (Cu), TiO2 (anatase) 

mesocrystal nanoarchitectures, and polypyrrole nanoparticles. Most of them are based on Au 

and Carbon (C) derivatives like nanotubes in their conductive forms as either single-, dual or 

multiwalled forms (i.e., SWCNT, DWCNT, MWCNT), nanocomposites (e.g., AuNP/ 

MWCNTs/chitosan, fullerene (C60)/CNTs) and reduced graphene oxide nanoparticles [105]. Back 

to the acceleration of the testing required time, graphene oxide rolled-up tubes with magnetic 

and catalytic movement [109] may improve the metrological features of cytokine‘s 

electrochemical nanobiosensors, mainly as preconcentration and anti-biofouling strategies. 



32 
 

Table 1. Outstanding electrochemical biosensor platforms for the measurement of cytokine levels in a POC testing manner. 

Cytokine Platform Method Metrology Ref. 

IL-3 

Type: Immunosensor 

Electrode: Gold AT (WE), Ag (RE), Gold AT (CE). 

Nanostructure: 2.7  µm Dynabeads M-270. 

Functionalization: NH- from Ab1 reacts with 

epoxy groups over Dynabeads. 

Detection: Biotinylated anti-IL-3 Ab2 reacts with 

streptavidin-HRP over WE. 

Sample, Time, T°: Serum and whole blood 

(100  µL); ~1  h at 20 °C. 

Equipment: Custom-designed potentiostat. 

Supporting solution: pH  7.0 PBS. 

Reactants: H2O2 (substratum); Ultra-TMB 

(mediator). 

Technique: Chronoampe-rometry, 100  mV 

vs Ag/AgCl  

LOD: 5  pg · mL -1 

LDR: 5-104  pg · mL -1 
[116] 

IL-6 

Other: PSA, 

PSMA, PF-4. 

Type: Microfluidic-based Immunosensor 

Electrode: Pyrolytic graphite (WE); 0.14  cm2. 

Nanostructure: 

1) SWCNT forest over a thin layer of iron oxide-

Nafion layer. 2) 5  nm GSH-AuNPs over a 

positively PDDA charged surface. 

Functionalization: Imidization by EDC and Sulfo-

NHS. 

Detection: Capture biotinylated anti-IL-6  Ab1; 

Ab2-streptavidin-HRP.  

Sample, Time, T°: Serum (5-10  µL); 

~3  h at 22±2 °C. 

Equipment: Eight-electrode CHI 1030. 

Supporting solution: pH  7.0 PBS. 

Reactants: 0.4  M H2O2 (substratum); 1  

mM hydroquinone (mediator). 

Technique: Chronoamperometry, 

 -0.3  V vs SCE. 

LDR: 

1)SWCNT:  40-150  pg · mL -1.  

2) AuNPs:  10-4000  pg · mL -1.  

[158] 

[160] 

[192] 

IL-6 

Type: Aptasensor. 

Electrode: Au (WE; 1  mm ∅), Ag/AgCl  (RE), Pt 

(CE).  

Nanostructure: AuNPs (2-3  nm). 

Sample, Time, T°: Commercial artificial  

sweat spiked with IL-6; ~1  h at 22±2 °C. 

Equipment: Gamry 600 (Warminster, PA, 

USA). 

Supporting solution: 0.1  M PBS (pH  7.40) 

LOD: 0.02  pg · mL -1. 

LDR: 0.02-20  pg · mL -1. 
[136] 



33 
 

Functionalization: Functionalized aptamer with 

alkanethiol HS-(CH2)11(OCH2CH2)3OH. 

Detection: Aptamer against IL-6. 

containing 5  mM K3Fe(CN)6. 

Reactants: None. 

Technique: EIS, 1  MHz to 1.0  Hz, 5.0  mV 

amplitude. 

IL-6 

Type: Aptasensor 

Electrode: GCE (WE). 

Nanostructure: AuNPs 

Functionalization: pATP anchored by the amide 

bond on an electrochemically grafted pABA SAM. 

The COOH- was activated by NHS+EDC chemistry. 

Detection: Anti-IL-6 aptamer. 

Sample, Time, T°: Serum from patients 

with colorectal cancer; ~1  h at °TRoom. 

Equipment: 

Supporting solution: 0.1  M KCl containing 

10  mM  [Fe(CN)6]3 -/4 -.  

Reactants: None. 

Technique: EIS, 1  Hz to 10  kHz, with 10  

mV amplitude. 

LOD: 1.6  pg · mL -1. 

LDR: 5-105  pg · mL -1 
[137] 

IL-8 

Type: Immunosensor 

Electrode: Reduced graphene oxide (WE), Pt  

(CE), Ag/AgCl (RE). 

Nanostructure: Cysteine-capped Au NPs. 

Functionalization: Imidization by EDC and Sulfo-

NHS. 

Detection: Anti-IL-8 label-free. 

Sample, Time, T°: PBS (Proof of concept) 

 at °TRoom.  

Equipment: Not provided. 

Supporting solution: PBS. 

Reactants: None. 

Voltage: 0.2  V vs Ag/AgCl 

Technique: Chronoamperometry. 

LOD: 0.589  pg · mL -1 

LDR: 1-12  pg · mL -1 
[129] 



34 
 

IL-6 

IL-8 

Other: PSA, 

VEGF-C 

Type: Microfluidic-based immunosensor 

Electrode: 1) Home-made SPE of Gold  (WE; 0.42  

mm2 ∅), wire Ag/AgCl (RE) and wire Pt (CE). 2) 

Carbon (WE), Ag/AgCl (RE). 

Nanostructure: AuNPs (5  nm) onto the WE with 

a poly(diallyl dimethylammonium-chloride) 

(PDDA) layer. 

Functionalization: Imidization by EDC and Sulfo-

NHS. 

Detection: Ab1 captures the IL-6/IL-8 form 

sample onto WE, which in turn serves as attach 

points for the anti-IL-6/IL-8 Ab2-MgB-HRP 

bioconjugates. 

Sample, Time, T°: Serum (5  µL); 

50  min at ~22±2 °C. 

Equipment: Eight-electrode CHI 1030. 

Supporting solution: pH  7.0 PBS. 

Reactants: H2O2 (substratum); 

Hydroquinone (mediator). 

Technique: Chronoampe-rometry, 

 -0.3  V vs Ag/AgCl. 

LDR: 

1)  10-1300  fg · mL -1 for IL-6. 

2)  5-50  fg · mL -1 for IL-6; 10-

50  fg · mL -1 for IL-8. 

 

[164] 

[165] 

IFN-γ 

Type: Aptasensor 

Electrode: Gold (WE; 1.6  mm2 ∅), Ag/AgCl (3  M 

KCl, RE), Pt (wire, CE). 

Nanostructure: None. 

Functionalization: Aptamer with a 5´-terminus 

C6-disulfide [HO(CH2)6-S-S-(CH2)6-linker is 

reduced with TCEP, before WE Au-S bond 

is induced overnight. 

Detection: Aptamer against IFN-γ labeled with 

MB. 

Sample, Time, T°: RPMI medium 

supplemented with bovine serum; 

~20  min at °TRoom. 

Equipment: CHI 842B (CHInstruments,  

Austin,  TX). 

Supporting solution: Liquid phase of 

sample. 

Reactants: hET between WE and MB-

tagged aptamer. 

Technique: SWV,  -0.1 to -0.5  V vs 

Ag/AgCl. 

LOD: 1  pg · mL -1 

LDR: 1-160  pg · mL -1 
[138] 



35 
 

IFN-γ 

Type: Immunosensor 

Electrode: Home-made SPE of graphene (WE), 

Ag/AgCl (RE), C (CE). 

Nanostructure: Graphene - aniline. 

Functionalization: Imidization by EDC and Sulfo-

NHS. 

Detection: Anti-hIFN-γ label-free. 

Sample, Time, T°: 2  μL of TCA treated 

serum; ~50  min at °TRoom. 

Equipment: SP200 BioLogic (Biologic 

Science Instrument, France), Autolab 

Electrochemical Analyzer (Ecochemie, 

Netherlands). 

Supporting Solution: 0.1  M KCl containing 

5  mM  [Fe(CN)6]3 -/4 -. 

Reactants: None. 

Technique: EIS, 101-105  Hz; with 10  mV 

amplitude. 

LOD: 3.4  pg · mL -1 

LDR: 5-1000  pg · mL -1 
[130] 

IFN-γ 

Type: Immunosensor 

Electrode: ITO modified with graphite-chitosan 

film (WE; 3  mm2 ∅), Pt wire (CE), SCE (RE). 

Nanostructure: GHS-AuNPs. 

Functionalization: Imidization by EDC and Sulfo-

NHS. 

Detection: Anti-hIFN-γ label-free. 

Sample, Time, T°: Serum; 2  h at 35  °T. 

Equipment: CHI-832 (Chenhua Instruments 

Co, Shanghai, China). 

Supporting solution: 0.1  M PBS - 

[Fe(CN)6]4 -/3 -,  pH  7.0 

Reactants: None. 

Technique: DPV,  -100 to 400  mV vs. SCE, 

at 100  mV/s. 

LOD: 0.5  pg/mL 

LDR: 5-4000  pg/mL 
[132] 

IFN-γ 

Type: 

1) 16-plexed immunosensor coupled  to 

microfluidic  (commercially available).  

2) 8-plexed continuous in-line system. 

Electrode: Au (WE), Au (CE), Ag/AgCl (RE). 

Nanostructure: None. 

Sample, Time, T°: 

1) Plasma, serum (40  µL); ~1  h at °TRoom. 

2) Plasma, serum (40  µL); ~8  min at 

°TRoom. 

Equipment: Custom-made electronic 

board. 

LOD: 

1)  ~10  pg · mL -1.  

2)  40  pg · mL -1  

LDR: 

1)  10-1000  pg · mL -1. 

2)  16-2048  pg · mL -1. 

[176] 

[177] 



36 
 

Functionalization: Cystein-labeled Fab', anti-IFN-

γ. 

Detection: Biotinylated anti-IFN-γ Ab1  reacts 

with  streptavidin-HRP, then the formed complex 

binds to the captured IFN-γ over WE. 

Supporting solution: PBS and plasma. 

Reactants: H2O2 (substratum); TMB  

(mediator).  

Technique: Chronoamperometry,  -300  

mV vs Ag/AgCl. 

 

IFN-γ 

Type: Immunosensor 

Electrode: C (WE; 4  mm2 ∅), C (CE), Ag (pseudo-

RE). 

Nanostructure: None. 

Functionalization: Electrografting of p-ABA 

followed by EDC/sulfo-NHS activation. 

Detection: Biotinylated anti-IFN-γ Ab1 reacts with 

streptavidin-HRP, then the formed complex binds 

to the captured IFN-γ over WE. 

Sample, Time, T°: Undiluted saliva 

(5  μL).; ~2  h 30  min at °TRoom. 

Equipment: PGSTAT 101 for p-ABA 

grafting (Autolab), Autolab type III for 

EIS and CHI 1030B (CH Instruments) 

for amperometry. 

Supporting solution: pH  6.0 PBS. 

Reactants: H2O2 (substratum); 

Hydroquinone (mediator). 

Technique: Chronoampe-rometry, 

 -0.2  V vs Ag pseudo-RE, for 5000  s. 

LOD: 1.6  pg · mL -1. 

LDR: 2.5-2000  pg · mL -1. 
[131] 

TNFα 

Type: Aptasensor 

Electrode: Gold (WE; 1.6  mm2 ∅), Ag/AgCl (3  M 

KCl, RE), Pt (wire, CE). 

Nanostructure: None. 

Functionalization: Aptamer with a 5´-terminus 

C6-disulfide [HO(CH2)6-S-S-(CH2)6-linker is 

reduced with TCEP, before WE Au-S bond is 

induced overnight. 

Sample, Time, T°: Drop of whole blood; 

~20  min at °TRoom. 

Equipment: CHI 842B (CHInstruments,  

Austin,  TX). 

Supporting solution: Liquid phase 

of the sample. 

Reactants: hET between WE and 

MB-tagged aptamer. 

LOD: 10  ng · mL -1 

LDR: 10-100  ng · mL -1 
[139] 



37 
 

Detection: Aptamer against TNFα labeled with 

MB. 

Technique: SWV,  -0.1 to 

 -0.5  V vs Ag/AgCl 

TNFα 

Type: Immunosensor 

Electrode: ITO (WE; 0.24  cm ∅), Pt (CE),  

Ag/AgCl (3.0 KCl; RE). 

Nanostructure: None. 

Functionalization: Electro-grafting in Ar 

atmosphere of PPC-PBA aryldiazonium 

salts followed by EDC/sulfo-NHS induces the 

attachment of the anti-TNFα (Ab1). 

Detection: anti-TNFα (Ab2) conjugated with HRP 

binds to the captured TNFα over WE. 

Sample, Time, T°: Whole blood 

(40  μL).; ~1  h at °TRoom. 

Equipment: Autolab poten-tiostat 

(Metrohm Autolab) for electrografting, 

CV and CA; Solartron SI 1287 with an SI 

1260 frequency response analyser 

(Hampshire, England) for EIS. 

Supporting solution: pH  7.4 PBS. 

Reactants: H2O2 (substratum); 

Ferrocene methanol (mediator). 

Technique: Chronoampero-metry, 

 -0.05  V vs Ag/AgCl. 

LOD: 10  pg · mL -1. 

LDR: 0.01-500  ng · mL -1. 
[112] 

TNFα 

Type: Immunosensor 

Electrode: CSGM (WE; 0.24  cm2 ∅),  Au (CE), Au 

(pseudo-RE).  

Nanostructure: Carboxylated magnetic 

nanoparticles (Sphereotech Inc, USA) sensitized 

with anti-Albumin and Anti-IgG for antifouling 

pretreatment step (i.e., NSB). 

Functionalization: Imidization by EDC and Sulfo-

NHS. 

Detection: Anti-TNFα. 

Sample, Time, T°: Undiluted serum 

(50  μL); ~2  h 30  min at °TRoom. 

Equipment: Reference 600TM (Gamry  

Instrument,  USA). 

Supporting Solution: Redox probe 5  mM  

[Fe(CN)6]3 -/4 - in 10  mM PBS, pH  7.4. 

Reactants: None. 

Technique: EIS, 0.5-500  KHz, 

with 25  mV amplitude. 

LOD: 1  pg · mL -1. 

LDR: 1-1000  ng · mL -1. 
[115] 



38 
 

TNFα 

Type: Immunosensor 

Electrode: Au (WE; 7.07×10 -2  cm2 ∅), Pt (CE), 

Ag/AgCl (3.0 KCl; RE). 

Nanostructure: AuNPs-RGO nanocomposites. 

Functionalization: 4-aminophenyl held the 

AuNPs-RGO nanocomposites to WE. 

Electrografting of PPC-PBA aryldiazonium salts, 

followed by EDC/sulfo-NHS activation, induces 

the attachment of the anti-TNFα (Ab1). 

Detection: anti-TNFα-ph-Fc-GO. 

Sample, Time, T°: Cell culture supernatants 

(10  μL).; ~20  min at °TRoom. 

Equipment: CHI660E (CHI Instrument, 

Shanghai). 

Supporting Solution: Redox probe 5  mM  

[Fe(CN)6]3 -/4 -  in 10  mM PBS, pH  7.4. 

Reactants: None. 

Technique: SWV,  -0.2  V to +0.6  V vs. 

Ag/AgCl, at 100  mV/s 

LOD: 0.1  pg · mL -1. 

LDR: 1-150  pg · mL -1. 
[113] 

TNFα 

Type: Aptasensor 

Electrode: C (WE; 12. 57  mm2 ∅), C (CE), Ag 

(pseudo-RE). 

Nanostructure: AuNPs. 

Functionalization: AuHCF-film-modified surface 

serves as AuNPs deposition via cyanide ions with 

the further thiolated aptamer reaction. 

Detection: Aptamer against IL-6. 

Sample, Time, T°: Serum (50  μL).; 

~20  min at °TRoom. 

Equipment: μStat DropSens potentiostat/ 

galvanostat (Drop Instrument; Spain). 

Supporting Solution: Redox probe 1  mM  

[Fe(CN)6]3 -/4 -  in 100  mM KCl, pH  7.4. 

Reactants: None. 

Technique: DPV,  -0.1 to 

0.3  V vs. Ag (pseudo-RE), at 10  mV/s. 

LOD: 5.5  pg · mL -1. 

LDR: 10-40  pg · mL -1. 
[140] 

IL-1β 

IL-6 

TNFα 

Type: Immunosensor 

Electrode: GC (WE; 0.071  cm2 ∅), Pt (CE), SCE 

(3.0 KCl; RE). 

Nanostructure: None. 

Functionalization: Electrografting of PPC-PBA 

aryldiazonium salts, followed by EDC/sulfo-NHS 

Sample, Time, T°: Whole mouse serum 

(10  μL); ~30  min at °TRoom. 

Equipment: CHI660E (CHI Instrument, 

Shanghai). 

Supporting Solution: Redox probe 1  mM  

[Fe(CN)6]3 - in PBS buffer, pH  7.4. 

LOD: 5  pg · mL -1. 

LDR: a)  IL-1β:  5-200  pg · mL -

1; b)  IL-6:  5-150  pg · mL -1;  c)  

TNFα:  5-200  pg · mL.  

[114] 



39 
 

activation. 

Detection: Anti-IL-1β, anti-IL-6 and anti-TNFα 

secondary antibodies (Ab2) labeled  with GO-MB,  

GO-NB and GO-Fc, respectively. 

Reactants: hET induced by the redox 

reaction between WE, tags and Supporting 

electrolyte. 

Technique: SWV,  -0.33 to -0.13  V for anti-

IL-1β Ab2-GO-MB; -0.58 to  -0.3  V for anti-

IL-6 Ab2-GO-NB; 0.0 to 0.35  V for anti-

TNFα 

Ab2-GO-Fc; -0.6 to 0.4  V for 

simultaneous testing at 100  mV/s. 

Ab1: Primary antibody; Ab2: Secondary antibody; BSA: Bovine serum albumin; CE: Co