PfHRP2 detection using plasmonic optrodes: performance analysis

The OF-SPR sensors tips were manufactured using FT400 Step-Index multimode fiber with a core diameter of 400 µm (Thorlabs). Pieces of 15 cm OF were cut, and a 4 cm end was stripped using a clamp. The stripped part of the OF was then immersed in acetone to remove the TECS hard polymer cladding, and cleaned manually by a cloth soaked in acetone until total removal of cladding remains. OFs were then cleaved at the right angle using a Vytran compact fiber cleaver (Thorlabs), to obtain 1 cm unclad sensing areas, and cut on the other side to connect the fiber. Final probe lengths of about 5 cm were used to avoid fiber bending during measurements. OF tips were then cleaned into piranha solution (H₂SO₄ and H₂O₂ 3:1) and silanized in methanol using 1% solution of (trimethylsilyls)-3-propanethiol during 15 min, at room temperature (RT). OFs were then dried overnight at RT and mounted on a holder for a single 50 nm gold sputter deposition, which was performed vertically (Fig. 1a, b). The fibers were then directly stored in cleaned Petri dishes on a holder to avoid any dusts contaminations of the fiber tip.

The immobilization of antibodies on the OFs as bioreceptors against PfHRP2 was performed through covalent bonding (Fig. 1c). OFs were immersed during 16 h in closed chambers of mercaptoundecanoic acid 2 mM in absolute ethanol at room temperature (RT). They were then washed in absolute ethanol and placed into 125 µL vials of 50 mM WSC and 50 mM NHS in activation buffer during 10 min at RT. They were then washed with PBS and immersed during 1h30 in anti-PfHRP2 antibodies. The blocking solution was then applied after another rinsing with PBS during 30 min at RT. Finally, the fibers were washed again in PBS and dried for biodetection experiments.

OFs are connected through a bifurcator (Y-bundle FC/PC) to a white light source (halogen) and a spectrometer (Ocean Optics HR4000, with 500–900 nm range and resolution of 0.19 nm). The spectrum released with this OF-SPR technique results in a gaussian curve (Fig. 2) where the dip quality (full width at half maximum: FWHM and intensity: depth) is influenced by the mode propagation angle and the number of reflections occurring inside the OF [33]. Data were analysed through an in-house developed MATLAB program.

Laboratory-adapted strains 3D7 were maintained in continuous culture following the protocol proposed by Trager and Jensen with some modifications. Parasites were cultured in complete medium supplemented at 5% haematocrit (A Rh⁺). Briefly, complete medium was prepared from RPMI 1640 powder (25 mM Hepes, 2 mM L-Glutamine) supplemented with 25 mM NaHCO₃, 0.5% AlbuMax™ II, 2 mM Hypoxanthine and 0.002% Gentamicin. Cultures were incubated at 37 °C under humid atmosphere in a gas mixture containing 5% CO₂. Cultures were diluted with fresh red blood cells every other day [34].

For the purpose of sample preparation, cultures were synchronized 48 h prior to sample collecting to have a majority of ring-stage forms. The sorbitol method was used to achieve synchronicity. Both whole culture and culture supernatant were used for analysis. Briefly, whole cultures were incubated with a 5% sorbitol solution at 37 °C for 15 min. The lot then underwent successive centrifugation cycles and finally put back in continuous culture as described previously. Two days post synchronization, cultures were left to sediment for 90 min in the incubator. After this period, the supernatant was taken up without disturbing the red blood cell layer. The supernatant was centrifuged at 3260 g for 5 min to remove any red blood cell residues. Whole culture samples were taken up and did not undergo any specific treatment before being analysed. Parasitemia and ring-stage percentage were evaluated by optical microscopy (Additional file 1: 1). Blood smears were prepared for each culture and stained (Giemsa 10%, × 1000). Parasitaemia was calculated as follows: [(number of infected red blood cells)/(total number of counted red blood cells)] × 100. Percentage of ring-stage was determined as the [(number of ring-stage infected red blood cells)/(total number of infected red blood cells)] × 100 [35].

The presence of HRP2 proteins in whole culture and supernatant was first qualitatively verified using rapid diagnostic tests before the OF experimentation (Malaria Ag Pf/Pan from Bioline). The HRP2 proteins spiked in PBS were also tested to verify their detection. These RDTs report a sensitivity of 99.7% and a specificity of 99.5%. 5 µL- sample was placed on the round well and diluted with four drops of the included assay diluent. The result was read after 15 min, when the migration of the sample on the colored bands was completed. The color of the control line attests the good migration while the apparition of a second band attests the presence of HRP2 proteins. HRP2 quantification in cultures analysed by the OFs was performed afterwards through an ELISA. Two distinct calibration curves were established using the HRP2-purified protein spiked into the blank culture medium or the culture medium supplemented at 5% haematocrit at different concentrations. The samples were diluted 1:1000 using culture medium or control blood and incubated on the microtitre plate in triplicates following the supplier recommendations.

Prior to biodetection experiments, the functionalization of the OFs and the related immobilization of the pfHRP2 antibodies was monitored in real time. First, the probe stability was verified into PBS during 1 h. The fibers were then immersed overnight into MUA 2 mM in ethanol to form a self-assembled monolayer (SAM) on the gold film. MUA molecules first bound to the surface and then self-organized, leading to a progressive shift of the SPR signal. The fibers were then rinsed and immersed into the solution of antibodies at a concentration of 20 µg/mL during 1h30. Finally, the blocking was also monitored (Fig. 3). This functionalization process shows the efficiency of the receptors binding to the fiber surface. This validation was performed for every lots of biosensors further tested. All these steps were performed in closed vials to avoid evaporation or perturbation of the medium during measurement, such as vibrations of the fiber probe.

Live monitoring biosensors such as SPR-based platforms have to deal with non specific bindings occurring into complex media. Such unwanted interactions lead to false positive detections or block the interaction sites on the surface. To avoid these side effects, the blocking reaction was optimized by testing three well known methods, often used for immunoblotting or ELISA. The efficiency of three blocking agents was tested on top of gold-plated OFs using casein 0.1%, fish gelatin 0.1% and BSA 1% (m/V), adapted from different protocols reported in the literature [36‐38]. The fibers were functionalized using MUA, the amine-coupling kit, and blocked using these procedures. The sensors were then immersed in PBS to check the stability of the probes and verify the strong anchoring of the surface biolayer. Afterwards they were immersed in culture medium, where an efficient blocking should avoid any perturbations of the signal. The fibers were then rinsed and immersed back into PBS. The first measurements show that the fish gelatin is not sufficient enough to block the surface while BSA and casein present better performances. The casein blocking agent also leads to a better stability in both PBS and the culture medium (Fig. 4). That method was, therefore, selected for further biosensing analyses. The latter is also easier to track during the immobilization process than the BSA, probably because its anchoring on the fiber surface provokes a higher surface refractive index shift, due to molecular sizes and weight differences or by its organization on the probe (homogeneity, surface shape). The organization of the blocking layers onto the surface may play a crucial role for the process monitoring and for the protection against non-specific adsorption.

OFs were immersed into PBS spiked with growing concentrations of PfHRP2. First, the fibers were placed into blank PBS and then into PBS + HRP2. The shift was monitored in real time and the final detection was considered 5 min after the immersion of the probe. These shifts are reported in Fig. 5a. The detection starts at around 1 µg/mL in a label-free configuration. However, the needed LOD is often below that level, so that a sandwich bioassay should be foreseen. For instance, the use of coupled nanoparticles or secondary binders are often used to enhance the signal shift due to their higher effects on the surface refractive index. In this study, secondary antibodies were implemented to bind to the anchored HRP2 proteins in a sandwich-like assay. Using these amplifiers, the detection threshold was improved as shown on Fig. 5b where the signal shift after amplification begins at lower concentrations, which were not detected with the label-free configuration. Here, a significant plasmonic shift starts at around 100 ng/mL.

The detection of HRP2 proteins was finally tested in vitro using cultures of P. falciparum. First, the optical fiber probes were immersed in PBS and then in cultures at 5% Ht without parasites (control) to verify their specificity but also to determine its starting SPR wavelength, as it is driven by the effective refractive index of the solution. There was no significant red shift observed in the control medium within 10 min. Its own refractive index is indeed close to the one of our PBS buffer which is equal to 1.3367 (given at 589 nm). The plasmonic probes were then immersed back in PBS to clean their surface and were immersed in Plasmodium culture for 10 min. A progressive red shift in the upper wavelengths of the SPR signal was observed, as presented in Fig. 6a. The fibers were then rinsed and disinfected into ethanol.

The detection of HRP2 inside these tested cultures was also attested by a strip-test showing the presence of the target by the apparition of a red band (Fig. 6b). An ELISA assay was also performed afterwards to quantify the amount of proteins in these cultures (Additional file 1: 2). It appears that the whole cultures tested show a concentration of HRP2 often ranging around 1 µg/mL while its supernatant following centrifugation has a much lower HRP2 concentration. It also seems impossible to precisely link the parasitaemia to the HRP2 expression, as it is often pointed out in the literature. Depending on the culture-time and synchronization, both parasitaemia and concentrations of HRP2 can be highly impacted.

Lab-on-fiber technology has shown major advances over the last decade [39, 40]. The unclad fiber approach presented in this article also remains a practical method to make use of plasmonics inside low-volume samples and to quicken the acquisition time compared to routine assays such as the ELISA [41]. It leads to highly tunable OFs that are interesting for many applications, especially to monitor live molecular events. In this paper, biofunctionalized optrodes were studied and optimized for the detection of HRP2 proteins in the framework of malaria detection. The possibility to use amplifiers to lower the LOD was shown, and the fibers were tested in complex media such as in vitro cultures of P. falciparum. The performances reached in this paper still leave room for improvements, especially to achieve more user-friendly interfaces. However, the many assets of OFs and their ease of implementation would enable substantial alternatives to the traditional equipment currently used at a research level, for instance, their combination with smartphones and portable devices, the elaboration of online bioprofiling databases [30].

The use of other molecular amplifiers to improve the LOD and reach the performances of best-in-class RDTs can also be investigated and optimized, such as the coupling of gold nanoparticles or nano-shell particles, often put in evidence in this field of research [42, 43]. The use of more specific targets such as pLDH or other bioreceptors, such as DNA aptamers or molecularly imprinted polymers (MIPs) could also be investigated to improve biosensing responses [44].