Enzyme‐free photothermally amplified fluorescent immunosorbent assay (PAFISA) for sensitive cytokine quantification

Cytokine monitoring has attracted great attention due to its significance in the diagnosis and treatment of many diseases, such as tumors, microbial infections, and immunological diseases. Enzyme‐linked immunosorbent assay (ELISA) is one of the most popular methods in cytokine detection, ascribing to the lavish signal amplification methods in the ELISA platform. In addition to classical enzymes, other signal amplifiers such as fluorescent probes, artificial nano‐enzymes, and photothermal reagents have been applied to reduce the detection limit and produce more sensitive ELISA kits. Due to the accumulative effect of heat, photothermal reagents are promising materials in the signal amplification of ELISA. However, the lack of efficient photothermal generation material at an aggregate scale may delay the further development of this area. In this contribution, based on an efficient organic photothermal aggregate material, an enzyme‐free photothermally amplified fluorescent immunosorbent assay system consisting of an assay microfluidic chip and detecting platform was developed. The photothermal nanoparticles with highly efficient photothermal conversion by harvesting energy via excited‐state intramolecular motions and enlarging molar absorptivity were successfully prepared. The detection concentration at 50 pg/mL of interleukin‐2 was achieved, realizing a signal improvement of detection limits by 20‐fold compared to that of previously reported photothermal ELISA. The microscopic imaging integrated with plane sweeping technology provided high spatial resolution and precision, indicating the potential of achieving high throughput profiling at the microscale. Moreover, as an alternative excitation source, light‐emitting diode not only provided a more affordable and miniaturized detection system but also revealed the great feasibility of intramolecular motion‐induced photothermy nanoparticles for biological analyses.


INTRODUCTION
Inflammation is the body's first line of defense against infection or injury, responding to challenges by activating innate and adaptive responses. [1]The coronavirus disease (COVID-19) pandemic reminded us of the critical role of an effective host immune response and the devastating effect of immune dysregulation. [2]Cytokine storm and cytokine release syndrome are life-threatening systemic inflammatory syndromes involving elevated levels of circulating cytokines and immune-cell hyperactivation that can be triggered by various therapies, pathogens, cancers, autoimmune conditions, and monogenic disorders. [3]It is important for clinicians to recognize a cytokine storm event as it has prognostic and therapeutic implications. [2]Cytokine monitoring is under the spotlight for disease prevention and protection due to the continuous threat of COVID-19 and the biological significance of cytokines. [4,5]The early recognition of cytokine storm through monitoring cytokine levels and prompt interventions may halt the progression of the disease to severe/critical stages and lead to better outcomes.The US Food and Drug Administration has authorized a blood purification product for reducing the amount of cytokines in the bloodstream that controls immune response by filtering the blood and returning the filtered blood to the COVID-19 patient suffering from a cytokine storm. [6,7]Therefore, it is essential to monitor treatment outcomes by detecting blood cytokines.
The classical measurement methods for cytokine detection involves the use of ELISA, [8] flow cytometry intracellular staining, [9] and cytometric bead assay. [10]ELISA is widely used in clinical tests and fundamental research due to its advantages such as standardized procedures, high sensitivity, and relatively simple reading equipment. [8]Most importantly, the ELISA platform has plenty of signal amplification methods, making it a vital technology in modern scientific research.For example, horseradish peroxidase (HRP) is widely applied in classical ELISA kits to catalyze its substrate 3,3ʹ,5,5ʹ-tetramethylbenzidine (TMB) to give a detectable color signal.Besides, the fluorescence immunoassay, showing good compatibility with the currently available analytical platforms, is considered to be one of the most sensitive approaches for screening biomarkers. [11]hang et al. reported an enhanced fluorescence enzymelinked immunosorbent assay (FELISA based on the human alpha-thrombin (HAT) triggering fluorescence "turn-on" signals.Nevertheless, as constrained by the performance of enzymes, their detection sensitivity has not been substantially improved in recent years. [12]Therefore, other signal amplifiers such as artificial enzymes have been developed for novel ELISA applications.Xia et al. reported an enzyme-free signal amplification technique, based on gold vesicles encapsulated with Pd−Ir nanoparticles as peroxidase mimics, for colorimetric assay of disease biomarkers with significantly enhanced sensitivity. [12]Hong et al reported an electrochemical immunosensing fabrication based on bienzyme functionalized Au-PB-Fe3O4 nanoparticles for Carcinoembryonic antigen and a-fetoprotein. [13]And Long et al. developed another enzyme-free system called metal-linked immunosorbent assay (MeLISA).MeLISA was demonstrated to exhibit approximately two magnitudes higher sensitivity and was four times faster for chromogenic reaction than classical ELISA. [14]Photothermy is an emerging signal that can be detected by simple instruments and possesses an accumulative effect suitable for signal amplification.Li et al. reported the photothermal effect of an iron oxide nanoparticles (NPs)mediated TMB-H 2 O 2 colorimetric system and applied it to the development of a new NP-mediated photothermal immunoassay platform for visual quantitative biomolecule detection using a thermometer as the signal reader. [15]In addition to nano-metals, organic photothermal systems are usually investigated due to their defined chemical structures, ease of large-scale preparation, and tunable properties.However, the lack of efficient photothermal generation material at an aggregate scale may delay the further development of this area.
Traditional planar molecules exhibit strong face-to-face ππ stacking interactions in the aggregate state, resulting in both insufficient radiative decay and non-radiative decay. [16]onsequently, there is an urgent need for highly efficient photothermy generation materials in an aggregate state.Fortunately, for the past two decades, plenty of aggregationinduced emission (AIE) materials have been developed.Among them, some new molecular design strategies have been proposed to manipulate molecular motions in the solution/aggregate state and enhance the molar absorptivity. [17]he 2 tetraphenylethylene-2naphthalene diimide-fused 2-(1,3-dithiol-2-ylidene)acetonitrile (2TPE-2NDTA) nanoparticles with highly boosted photothermy by virtue of internally efficient excited-state intramolecular motions were successfully developed. [17,18]As for the detection methods, the thermal meter was applied for the measurement of temperature changes in previously reported work; however, the sensitivity and accuracy were not adequate to detect trace cytokines. [15]In addition, pulsed lasers, which are usually required for generating photothermy, are associated with the high cost and demand stability requirements.Replacing lasers with commercial light-emitting diode (LED) sources would greatly impact the equipment cost of this new detection platform.
Therefore, in this work, we developed an enzymefree photothermally amplified fluorescent immunosorbent assay (PAFISA) platform based on photothermal aggregates (Scheme 1) integrating with an efficient microfluidic chip.As demonstrated in Scheme 1, through the introduction of microscopic imaging technology and two-dimensional scanning technology, the measurement of the micro-area and mapping in the relative area was realized.The sensitivity was 20-fold higher than that previously reported and the result accuracy was improved by averaging through the multipoint collection.

RESULTS AND DISCUSSION
The 2TPE-2NDTA-02 (TN for short, Figure 1A) was synthesized as previously described [19] (Scheme S1).To fabricate highly efficient photothermal NPs, TN was doped into 1,2distearoyl-sn-glycero-3-phosphoethanolamine-N-[methoxy (poly(ethylene glycol))−2000-biotin] (DSPE-PEG 2000biotin) using the nano-precipitation method, [19] as illustrated in Figure 1B.The size of the TN NPs was 89 ± 0.7 nm as determined by dynamic light scattering and transmission electron microscope analysis (Figure 1C).The zeta potential of NPs was shown to be −16 ± 1.1 mV, consistent with previous results. [19]Furthermore, the size of the NPs did not change for 3 months, [19] indicating the good colloidal stability of the NPs over time.Light absorption is requisite for eligible photothermal materials.According to the intramolecular motion-induced photothermy mechanism, [18] the energy input also influences the photothermal conversion efficiency (PCE) of TN NPs.Although NPs show high PCE under 808-nm pulse laser illumination, [18] the low output power and high equipment cost of the pulse laser still limits the application of the TN NPs.Here, we explored the possibility of using LEDs as an alternative light source.According to the absorption profile of the TN NPs (Figure 1D), we used several high-power (5 W) LEDs with different central wavelengths (710, 730, 760, and 780 nm, respectively) to excite TN NPs. Figure 1E presents the infrared thermal images of TN NPs in aqueous solutions which were excited at 710, 730, 760, and 780 nm using an LED with same output power (5 W), while Figure 1F indicates the corresponding photothermal curves.The relative calculated PCE is shown in Figure 1G.Although the highest local PCE of TN NPs was at 730 nm, the emission profile of the LED bulbs (Figure S2) showed that the tail of the 730-nm LED extended as far as 560 nm, which causes serious overlap with the temperature-dependent signal whose peak is at 577 nm; thus, affecting the measurement accuracy.Conversely, the PCE at 760 nm was still high enough and, as the tail of the 760-nm LED emission ends at 650 nm, it is distant enough from the temperature-dependent emission peak.Furthermore, the temperature increase caused by the 730-nm LED might be too high and affect the biological activity of protein while that induced by the 760-nm LED is at around the body temperature; thus, the 760-nm LED was selected as the excitation source of photothermy after comprehensive consideration.Subsequently the different concentrations TN NPs aqueous solutions were excited at 760-nm using an LED (5 W) for 10 min and then cooled down naturally to room temperature.The temperature changes elevated with time and reached a maximum within 5 min (Figure 1H and Figure S3).The plateau temperature increases of different concentrations of TN NPs were 21.
1 μg/mL), and 2.0 • C (0.01 μg/mL).The TN NPs represented a good function relation between concentration and photothermal conversion within the range from 0.01 to 100 μg/mL.Furthermore, the stability of the photothermal performance under 760-nm excitation over 21 days has also been checked (Figure S4).Thus, all the data indicating TN NPs' great potential for designing and speculating temperature for specific biological applications.
The ratiometric fluorescence methods have been frequently applied for temperature sensing since they can eliminate variations such as indicator concentration, measurement geometry, light source intensity, and light field to realize higher accuracy. [20,21]As shown in Figure 2A and the upper panel of Figure 2B, the PAFISA chip was constructed with a glass substrate, temperature sensing film, and a polydimethylsiloxane (PDMS) slab.In brief, the temperature sensing film consisted of Ru(phen) 3 , perylene, and poly(styrene-co-acrylonitrile) (PSAN).The commercially available Ru(phen) 3 , whose emission peak was at 577 nm, served as a temperaturesensitive probe; [22] while perylene was used as a reference dye since temperature changes seldom showed effects on its three luminescence intensity peaks (442 nm, 466 nm, and 500 nm). [20]As for the film matrix, compared with the polyacrylonitrile (PAN), which was a conventional paired matrix for mixing Ru(phen) 3 , [22][23][24][25][26] PSAN exhibited similar characteristics as PAN in low gas permeability conditions for avoiding oxygen quenching of luminescence while maintaining the ability of polystyrene (PS) to be activated and bind with PDMS and capture antibodies. [27]The addition of (3aminopropyl)triethoxysilane (APTES) for surface treatment after oxygen plasma activation promoted the sensing layer to bond with PDMS and immobilize capture antibodies due to the aminosilane. [28]The absorption and emission spectra of temperature sensing film doped with Ru(phen) 3 and perylene are demonstrated in Figure 2C.The absorption range of this sensing film was between 350 and 500 nm, indicating a wide range of excitation light choices.These dual luminophores were used as a temperature indicator, which were obtained using the ratio of the peak of Ru(phen) 3 (I  = 577.23 ) to the peak of perylene (I  = 466.45).To examine the deposition of capture antibodies on the temperature sensing film within the PDMS chamber, a commercial interleukin-2 (IL-2) murine ELISA kit was adopted.The chip presented a blue color after injecting TMB solution (lower panel of Figure 2B), which indicated that capture antibodies were successfully bound onto the film surface and that the chip with sensing dye in polymer film would not hinder the binding of capture antibodies.
Two different approaches were applied to calibrate the temperature sensing film, in which diffused or focused excitation light (405 nm, in purple color) was adopted sep-arately.To make measurements on a micrometer-sized area, a homemade confocal microscopy system was used to focus the 405-nm laser to a diameter of approximately 3 μm, which improved the detection precision and spatial resolution (Figure 2D).The chips were evenly heated by a hot plate from room temperature to 80 • C at a stepwise of 2 • C. The 405-nm laser was used to excite the mixed dye, and the intensity ratio of the emission peak at 577.23 nm to the one at 466.45 nm was collected in real time as the temperature-dependent signal at each temperature (Figure 2E).The intensity at 577.23 nm, which came from Ru(phen) 3 decreased exponentially as temperature increased while the intensity at 466.45 nm showed small changes, which provided the ratio of intensities R = I  = 577.23 ∕I  = 466.45 ) at different temperatures for calibration.Five different positions of the sensing film detection provided the signals at their corresponding temperature after equilibrium, which generated a calibration plot that yielded intensity ratios with small standard deviations.The decay of intensity ratio with increasing temperature could be described by a good exponential fitting with high sensitivity that achieved a 0.03 • C on the average slope (Figure 2F).The fluorescence ratiometric system composed of Ru(phen) 3 and perylene exhibited excellent temperaturedependent performance.This high sensitivity could be a good amplifier of the signal that causes the temperature change and improve the assay's detection limit.
To investigate the photothermal effect of TN NPs in the PAFISA chip, TN NPs suspensions with different concentrations (100 pg/mL-100 μg/mL) were injected into the PAFISA chip and the temperature increase induced by the 760-nm LED irradiation was measured.During the detection process, a 20× objective was driven by an X-Y scanning piezo stage, and spatial mapping (Scheme S2) with an area around 100 × 100 μm on the chip was performed with a step precision of 10 μm (lower panel of Figure 3A).This detection method provided a static average signal that could reduce the disturbance resulting from the localized uneven spatial distribution of capture antibodies, which was an inherent drawback of traditional ELISA assays.Consequently, the PAFISA chips filled with NPs at different concentrations were illuminated at 760 nm using an LED for 250 s, and the signals were recorded in real-time (shown in Figure 3B).After 140 s, each chip reached a temperature plateau with a constant value, illustrating that the PAFISA chips could be readout within only 5 min.Figure 3C showed a corresponding emission profile of different concentrations after 250 s of illumination.To visualize the photothermal signal vividly, mapping results were also obtained, as shown in Figure 3D.According to the color bar, the raising concentration of NPs accompanied by temperature increases induced by the 760 nm LED excitation, elucidated that the photothermal effect was successfully performed and that the heat was cumulative in confined chambers of PAFISA chips with various concentrations of TN NPs.
The statistical results are shown in Figure 3E, for NPs with concentrations between 0.1 ng/mL and 1 μg/mL, a linear range of corresponding temperature difference from 0.42 ± 0.14 • C to 8.36 ± 0.26 • C was observed, which could be used for antigen analysis.
PAFISA chip and relevant detection optical path were successfully developed (Scheme 1 and upper right panel of Figure 4A).IL-2 was chosen as a cytokine model to demonstrate the detection ability of PAFISA chips.According to the design principle, IL-2 would bind to the capture antibodies that were coated on the surface of the temperature sensing film.Next, antigens would be detected by IL-2 biotinylated detection antibodies and then bind to the streptavidin and biotinylated photothermal aggregates (TN NPs).Consequently, the concentration of antigens could be calculated based on the temperature increase associated with the accumulative photothermy generated from TN NPs.The temperature could be monitored and visualized by the temperature sensing film and a home-made confocal system.The temperature increase corresponded to the increase in NP concentration, which in turn indicated the antigen concentration.The IL-2 concentration to temperature plot showed a linear increase in temperature from 50 pg/mL to 50 μg/mL with an average standard deviation of 0.22 • C (Figure 4B).This result elucidated that the temperature difference detected on the sensing layer was able to be utilized to analyze the IL-2 concentrations as low as 50 pg/mL, which was 20 times more sensitive than the photothermal ELISA report previously. [15]oteworthily, the maximum temperature in all the groups has been tuned to a temperature approximate to the body temperature, thus guaranteeing the sample's bioactivity.Moreover, the specificity of the PAFISA developed in this work was investigated by using different proteins as samples (shown as the lower panel of Figure 4A), and the data are shown in the Figure 4C,D.Proteins with a concentration of 50 μg/mL including bovine serum albumin (BSA), cytochrome, fibrinogen (Fg), hemoglobin (Hgb), and phosphate buffered saline (marked as blank) were applied in this cross-testing assay.The data showed that a significant temperature increase signal was only found in the IL-2 group, while there was no temperature change in other groups.This result demonstrated the high selectivity and accuracy of PAFISA in specific protein detection.When performing classical ELISA, the TMB substrate is favorable, as the system is sensitive to color change and the change can be easily observed.Although TMB solution could be visually observed and detected by ultraviolet-visible (UV-vis) absorbance, the light-sensitive feature [29] limits its detection window and storage life. [30]urthermore, the enzymes used for catalysis and the addition of stop reagents make the systems less controllable.Compared to the enzymatic TMB generation-based ELISA assay, the PAFISA chips could be stored for at least 3 days (Figure S5) since the photothermal NPs were stable and would not be affected by the ambient optical environment.This could offer enough time for a standard assay procedure and allow for performing the assay in an environment without light shelter, which overcame an unavoidable drawback associated with traditional enzyme-based assays.

CONCLUSION
In this paper, we introduced a highly efficient AIE photothermy material integrating with microfluidics and feasible optics for investigating cytokine detection assay.This enzyme-free PAFISA system consisting of an assay microfluidic chip and detecting platform was developed, and we showed that it is practical, accurate, and potentially high-throughput.The photothermy aggregates 2TPE-2NDTA nanoparticles displayed high efficiency for trace cytokine detection by using a 760-nm LED to illuminate.In this system, the fluorescence intensity ratio of Ru(phen) 3 to perylene acts as a temperature signal amplifier, which improves the detection sensitivity compared to that of an absorption-based assay.Furthermore, the adoption of microscopic imaging combined with plane sweeping technology in the system improved its spatial resolution, accuracy, and precision, allowing for detecting concentrations as low as 50 pg/mL, which is 20 times more sensitive than previously reported methods.By replacing the enzyme-based signals with fluorescence signals of temperatures generated using sensing films with NPs, we made our assay intuitive and user-friendly, which allows manipulating it in an ambient optical environment.Furthermore, the expensive pulsed laser was replaced by the low-cost LED bulb as the excitation source, which makes our platform more economical, portable and allows for miniaturization.Overall, this PAFISA system exploits photothermic NPs to detect cytokines injected onto a microchip with thermal sensing films and provides a new opportunity for using photothermal characteristics in high throughput screening and detection at microscale.

Description of the measurement system
As shown in Figure 1C,D, the measuring system can be divided into two parts, namely the excitation section and the detection section.In the excitation section, a 5 W LED bulb whose central wavelength is 760 nm was chosen as the excitation light source to heat the TN NPs.Different temperature changes depending on different TN NPs concentrations could be detected by the thermosensitive dye coated on the microfluidic chip.This temperature change can be accurately detected by the detection section.A 405 nm continuous laser beam (Radius 405, Coherent, Santa Clara, CA, USA) was used to pump the thermosensitive dye, and the temperature-dependent spectrum was expanded using a spectrometer (Shamrock 193i, Andor, Belfast, Northern Ireland) and recorded using the CCD (Newton EM CCD, Andor).To make measurements on a micrometer-size area, a homemade confocal microscopy system mounted with a 20× objective (LMPlanFL N 20×/0.40,Olympus, Tokyo, Japan) was used for the detection section to focus the laser to a diameter of approximately 3 μm to improve the detecting precision.Furthermore, the objective was guided using an X-Y scanning piezo stage (N12, Coremorrow, Harbin, China), which makes spatial discrimination possible.The linkage and control program of the spectrometer/CCD and the piezo stage was described by LabVIEW, the detailed control flow diagram is shown in Scheme S2.A bandpass filter (FF01-715/SP-25, Transmission Band: 450-700 nm, Tavg > 93%, Semrock, Rochester, NY, USA,) was set before the spectrometer, which allowed the separation of the signal from the 405-nm laser and the 760-nm LED background.

Chip fabrication
The temperature sensor layer was made using a coating solution: 5 mg tris(1,10-phenanthroline) ruthenium (II) (trimethylsilyl)propane sulfonate ((Ru(phen) 3 )TSPS), 1.25-mg perylene and 500-mg poly(styrene-co-acrylonitrile) (PSAN) were mixed in chloroform (3.4 mL; all from Sigma Aldrich, St. Louis, MO, USA) and stirred overnight at room temperature.The Ru(phen) 3 -perylene-PSAN mixture was spread through blade coating (BDG218, Biuged, Guangzhou, China) on a glass surface.A thin film was generated after evaporation of the solvent.The resin substrate was patterned through threedimensional printing to obtain the master mold.The microchamber was cast onto a PDMS slab of 20-mm length, 10-mm width, and 3.5-mm height that had a recess of 8-mm length, 4-mm width, and 1-mm height.The PDMS elastomer and the curing agent were mixed at a ratio of 10:1 and poured onto the silicon wafer mold.PDMS was degassed on the mold in a vacuum desiccator chamber and then heated at 80 • C in an oven for a 1-h curing period.Then, PDMS was peeled off from the mold and cut out with two 1-mm holes for the tubing.
The PSAN was treated with a 1 min oxygen plasma discharge, followed by 1% APTES solution coating on the surface for 20 min, which promoted the sensing layer to bond to PDMS and immobilize capture antibodies.The PSAN film was dried with nitrogen or air after discarding the 1% APTES solution.The PDMS microchamber surface was rinsed with isopropanol (IPA) and distilled water to remove the dust.To bond the PDMS microchamber and the PSAN surface, both PDMS and the PSAN were treated with an oxygen plasma discharge, which was generated at 30 W using a plasma cleaner (PDC-002, Harrick Plasma, New York, NY, USA).

S C H E M E 1
Schematic diagram of the photothermally amplified fluorescent immunosorbent assay (PAFISA) chip.

F
I G U R E 1 (A) Chemical structure of 2TPE-2NDTA-02.(B) Schematic illustrations of TN nanoparticles (NPs) fabrication by encapsulating DSPE-PEG2000 with 2TPE-2NDTA-02.(C) Dynamic light scattering (DLS) histogram and transmission electron microscope (TEM) image of TN NPs.Scale bar: 50 nm.(D) The absorption profile of the nanoparticle.(E) Infrared thermal images of TN NPs in aqueous solution (100 μg based on TN) under various irradiation of 5 W light-emitting diode (LED) with different central wavelengths for 5 min.(F) Temperature changes in the TN NPs solutions as a function of time.The solutions were irradiated using a 5 W LED and different central wavelengths for 10 min followed by natural cooling.(G) The calculated photothermal conversion efficiency (PCE) of TN NPs under different wavelength excitation.(H) Photothermal conversion behavior of TN NPs in aqueous solution at different concentrations (0.01, 0.1, 1, 10, and 100 μg/mL) under 760 nm LED (5 W) irradiation.

F
I G U R E 2 (A) Process flow of the photothermally amplified fluorescent immunosorbent assay (PAFISA) microfluidic chip integrated with a temperature sensor.(B) Upper panel: A ready-to-use PAFISA chip, with a coin of 25 mm in diameter for size reference, lower panel: the result of the deposition of capture antibodies on the temperature sensing film within the polydimethylsiloxane (PDMS) chamber using a commercial interleukin-2 (IL-2) murine enzyme-linked immunosorbent assay (ELISA) kit.(C) Absorption and emission profile of the mixed dye consisting of Ru(phen) 3 and perylene at room temperature.(D) Schematic diagram of the temperature calibration of the temperature sensing film.(E) Representative normalized emission profile of temperature sensing film under different temperatures.(F) Static result of the intensity ratio as a function of temperature.

F
I G U R E 3 (A) Schematic representation of TN nanoparticles (NPs) measurement in a chip chamber using thermal detection.(B) Photothermal curves for different TN NPs concentrations (100 pg/mL-100 ug/mL) irradiated at 760 nm using a light-emitting diode (LED) (5 W) for 250 s. (C) Corresponding emission profile of temperature sensing film at 250 s. (D) Temperature mapping at different concentrations of the NPs.(E) Plot of temperature changes with different concentrations.F I G U R E 4 (A) Sensitivity and specificity assessment of photothermally amplified fluorescent immunosorbent assay (PAFISA).(B) Static result of the temperature changes at different concentrations.The inset represents the temperature mappings corresponding to different concentrations.(C) Photographs and (D) temperature difference based on PAFISA results from interleukin-2 (IL-2) and other five different interfering substances.