A simple and intuitive methodology for estimating red blood cell surface antigen expression variation using optical cell-detachment technique


 The analysis of surface antigens on cells, especially red blood cells (RBCs), has attracted increasing attention due to the recognition of antigenic variation that can facilitate early diagnoses. This paper presents an alternative methodology to estimate the variation of surface antigen expressions using an optical cell-detachment technique to validate the binding of individual RBCs stuck on corresponding antibody-coated surfaces. The detachment tests were implemented by an optical tweezers with gradually decreasing laser powers associated with serial antibody dilutions. Then, the antigen expression variation was estimated based on the known antibody dilution folds. The B- and B3-types of RBCs were selected for the demonstration subjects. With the semi-quantitative analysis, the proposed methodology was successfully verified for evaluating the variation of the RBC surface antigen expressions. The analysis result shows good consistency with the literature’s findings.


Introduction
Numerous antigens composed of peptides, proteins, and other molecules are present on the cell surface. The antigens are the critical molecular markers of cell functions and lineages [1]. The antigenic change in quantity is crucial, especially for hematologic development, immune response, and tumor progression [2][3][4]. For instance, the human leukocyte antigens (or CD antigens) expressed on immune cells participate in or perform functions in the immune response [5]. RBCs can be classi ed by ABO blood group system based on inherited differences and expression in cell surface antigens [6]. Also, many reports have shown that blood group antigen expression is correlated with tumor progression [7][8][9].
Flow cytometry and enzyme-linked immunosorbent assays (ELISA) have been developed for the analysis of cell surface antigens. For instance, the ow cytometric analysis results performed by Cho et al. reveals that the transfect expressed 35.5% of the total B antigen produced on the B101 allele transfect [10]. Chen et al. showed that the percentage of B antigen expression on B3 cells is 40.92% of B1 [11]. However, the conventional methods have limitations due to low sensitivity and the need for large clinical samples [12,13]. For example, ow cytometry requires more than 5,000 to 10,000 copies of the target antigen on one cell to ensure quantitative analysis [12] , whereas 50,000 cells per well are needed for ELISA to obtain high optical density values and low background noise during detection [14].
More recently, the focus is shifted to the individual cell analysis to minimize sample size and enhance sensitivity. Subsequently, the optical tweezers and atomic force microscope (AFM) have been widely employed to the single-molecule biophysics research [15][16][17][18]. Optical tweezers, in contrast to AFM, are useful for single cell manipulation with precise force measurement [15]. The exerted force of optical tweezers is in the range of pico-Newton to nano-Newton scales and is resulted from the laser interaction with dielectric micro-object, with optical force ranges in the pico-Newton to nano-Newton scales. In 1989, direct trapping of RBCs using optical tweezers was reported in the pioneering works of Ashkin et al. [19].
Subsequently, optical tweezers are now being used in the investigation of an increasing number of biochemical and biophysical processes [20][21][22][23][24][25][26][27][28][29][30]. Many easy-to-use commercial products of optical tweezers have been promoted. Although optical tweezers has been widely used as a manipulation tool in interdisciplinary elds, it is still rare to employ this technology as a biosensor by the "detachment" operation.
In this study, we extended our preliminary sequent works [31][32][33] based on the speci c antigen-antibody interactions principle and optical manipulation technique to develop an alternative methodology for estimating the variation of antigen expressions on RBCs. Herein, the optical tweezers was employed as a biosensor to detach an individual RBC off from an antibody-coated surface with serial dilution folds and laser powers for detecting the RBC binding.

Preparation of RBCs and antibody-coated slides
Standardized RBCs were provided by Formosa Biomedical Technology Corp. (Taipei, Taiwan) and diluted 800 fold with PBS before the addition of 0.1 g/ml BSA to prevent the RBCs sticking together and block non-speci c interactions with antibodies or the slide. One drop (~0.02 ml) of the RBCs solution was incubated on the antibody-coated slides for 20 minutes at room temperature just prior the measurement.
The antibody-coated slides were prepared by protein adhesion on surfaces coated with poly-L-lysine [34,35], an e cient method to prepare antibody microarrays [36]. Cover slides (Paul Marienfeld GmbH & Co. KG/Germany) were cleaned with acidic alcohol (1% HCl in 70% ethanol), rinsed thoroughly in ultrapure H2O, incubated at room temperature in a 1:10 poly-L-lysine solution (Sigma-Aldrich #P8920) for 5 minutes, and then dried in a 60°C oven for 1 hour. Solutions of anti-A and anti-B monoclonal antibodies (1 mg/ml) were provided by Thermo Fisher Scienti c Inc. (Waltham, USA) and diluted with PBS solutions (Sigma-Aldrich #P4417) with serial dilution folds. Poly-L-lysine coated slides were incubated in antibody solutions at room temperature for 1 hour, then for 5 minutes in 0.05 g/ml BSA solution to block nonspeci city and stored at 4°C.

Optical tweezers for RBC manipulation
The schematic of the optical tweezers system based on an inverted microscope platform (Olympus IX51) is shown in Figure 1. A continuous-wave Nd-YAG laser (Onset Electro-Optics, model # ISF064-1000P) at λ=1064 nm focalized by a high NA (1.3) microscope objective/oil (UPLFLN100XO2, Olympus) provides the trapping beam with minimum and maximum available powers of 4 mW and 250 mW, respectively. Herein, the laser power was measured at the microscope objective. The RBCs solutions were con ned to an isolated sample chamber comprising two cover slides (170 μm thickness), and a double-faced tape (120 μm thickness) to eliminate ow disturbance. The chamber can be moved with an XYZ-axis nanopositioner (Physics Instrument, NanoCube®P611), while the laser is focused at a xed position in the chamber. The optical dragging speed was kept low (5 μm/sec) for the static test during detachment, so the solution viscous effect can be ignored.
The basic principle of optical tweezers to produce optical forces for manipulating micron-sized dielectric objects have been described previously [37][38][39]. A laser beam is focused by a high numerical aperture (NA) of a microscope objective to a spot (less than 1 μm) in a transparent micro-object, generating an optical trapping force. The force (F) can be expressed as , where P is the laser power, n is the relative refractive index, is a dimensionless parameter related with object dimension, NA, wavelength, polarization, beam pro le, and spot size. Essentially, 1 mW of laser power approximately generates 1 pN force for a 1 μm diameter sphere [40]. In this study, the experiment subject, RBC, is 7.5~8.5 μm in diameter, with a width of 1.7~2.2 μm in the ring and 0.5~1 μm at the center [41]. During cell-detachment testing, the optical trapping spot exerts at a constant volume at the ring edge of an attached RBC. And, the relative index (n), the NA, the wavelength, the polarization, the beam pro le, and the spot size are consistent in the experiment with the same optical tweezers laser and medium. Hence, the parameter is constant so that the optical force (F) is linearly proportional to the laser power (P).

Demonstration of RBC detachment
First, the negative control experiment, i.e. the case of non-speci c antibody-antigen interaction was performed. Figure 2.(a) schematically illustrates a non-attached RBC trapped by optical tweezers, showing the RBC is vertically aligned by the optical torque ( Figure 2.(b)). The suspended RBC can be freely dragged by the optical tweezers in solution (Figure 2.(c)) at the threshold power, i.e. 4 mW. The sequent lms of the manipulation are shown in Fig. 2d. In contrary, when the RBC antigens speci cally interact with the antibody-coated surface, the RBC attaches to the slide (Figure 3.(a)). For example, Figure   3.(b) represents an A-type RBC stuck on the anti-A-coated surface even using the maximum available power (250 mW) of optical detachment. The basic qualitative tests of speci c antibody-antigen interactions were veri ed using blood types A, B, and O (Table 1.). The RBC samples were dropped on the functionalized slides coated with the associated anti-A and anti-B antibodies. The criterion of celldetachment demonstration is strict. Each test repeated at least 5 times for 5 cells, that is, 25 continuous trials to validate the antigen-antibody binding (marked as ). In other words, once a RBC was detached in any trial, the test was deemed invalid (marked as ).

Antibody dilution method
The "antibody dilution method" was specially proposed in this study. The binding strength between an RBC and an antibody-coated surface, i.e. RBC-antibody a nity, is proportional to the antibody-antigen associations, which depends on the quantities of the antigens/RBC surface and the antibodies/functionalized surface. Once the binding condition is validated by the cell-detachment test, the antigen expression variation can be estimated with the known antibody dilution folds.
Serial slides coated with decreasing antibody concentrations were prepared by dilution and stored at 4°C before testing. The dilution folds increase by an exponent of 2, but will be adjusted according to the test requests. For example, the dilution folds increase by 512 after 2048, such as 2560, 3072, 3584…etc for getting more precise testing. The RBC-antibody a nity was quanti ed by the highest dilution at which the RBCs could not be detached by the optical tweezers from the functionalized surface.

Optically detachment tests with dilution method
Rare blood group (e.g. subtype) discrepancies occur when antigens -A and -B have a weak expression, i.e. there are much fewer antigens at the RBC surface [42]. In this study, A-and B-type RBCs were selected as the subjects to implement the optical cell-detachment method. B3-type RBCs, the most common B subtype in the Asian population [11] were selected as the subject to compare the antigen expression variation with B-type RBCs.
The detection strategy includes two steps. First, the antibody concentration is high enough to interact with all the antigens at the contacted part of an RBC surface. The quantity of the antibody-antigen bindings is consistent. The stuck strength of an RBC on the functionalized surface is large enough to oppose the optical detaching pull even using the highest power (250 mW). In the sequent tests, as the antibody concentration becomes lower by dilution, the quantity of the antibody-antigen bindings is gradually decreasing until the RBC can be detached using the power of 250 mW. At this time, the antibody dilution fold reaches a critical value. The results listed in Table 2. indicate that A-and B-type RBCs shows the same result, varifying the reliability of the method. The binding of B-type RBCs can be validated until the fold of antibody dilution is 4608 while the highest dilution for B3-type is 1024, showing an apparent distinction that is consistent with the previous illustration. The difference in antibody dilution folds implies the variation of the surface antigen expression. In the second step, the antibody concentration continually decreases, and the optical power for detaching an RBC becomes lower than 250 mW. The laser powers gradually decreased by 5 mW for further dilutions until 10 mW, and then gradually decreased by 1 mW. For instance, the B-type RBCs were detached when the dilution fold is 5120 at 250 mw power in the rst step tests. Subsequently, the applied powers were gradually decreased until the binding could be validated, i.e. the RBCs could not be detached. Finally, the binding strength was too small to oppose the optically detaching pull even using the lowest available power (4 mW). At this time, the antibody dilution fold reaches the maximum value. The results listed in Table 3. indicate that the binding of B-type RBCs can be validated until 7168-fold dilution while the highest dilution for B3-type RBCs is 2560. As a result, the more precise detection can be concluded.   Table 3 shows that the antibody dilution fold (D) is inversely proportional to the laser power (P). The R-squares of B-type and B3-type are 0.99 and 0.98, respectively, validating the method. The phenomenon implies the higher dilution (lower antibody concentration), the lower power (lower optical force) needed to detach the RBC binding. With this relationship, the RBC surface antigen expression variation can be semi-quantitatively analyzed. The maximum antibody dilution in the B-type test is 2.7 times (7168/2560) of the B3-type, meaning that the antigen expression on the B3-type RBC is 37.5% of which on the B-type RBC. The result is consistent with the previously mentioned literature [11,12] using similar types of RBCs. Thus, the proposed approach was veri ed. After decades of development, convenient commercial optical tweezers machines and automatic manipulation systems with micro uidic chips [43][44][45][46] are becoming available and popular. The automatic operation would promise more optical detachment trials for more detailed dilution folds to obtain a higher resolution. Therefore, a more precise quantitative estimation could be obtained.

Conclusion
A novel methodology using the optical cell-detachment technique has been developed to estimate the variation of RBC surface antigen expression in this study. It suggests an alternative approach for the whole-cell-based antibody-antigen interaction analysis. The experiment results are consistent with literature that used conventional methods. One drop of blood from the ngertip was more than enough to achieve a sensitive detection. If necessary, a higher resolution of estimation could be obtained by more detailed dilution folds. The proposed concept is easy-to-use and intuitive without complicated biological treatment processes and optical force calculations. With the development of automatic routine procedures, we believe this approach could be widely capable with potential applications in the future, such as molecular biology, biophysics, and micro-/nano-technology engineering.