Identification and Quantification of Anti-Gp.Mur Antibodies in Human Serum Using an Insect-Cell-Based System

Gp.Mur is a clinically relevant antigen of the MNS blood group system that is highly prevalent in several Asian populations. Its corresponding antibody, anti-Gp.Mur, has been implicated in hemolytic transfusion reactions and hemolytic disease of the fetus and newborn. Currently, identifying and confirming anti-Gp.Mur antibody presence in sera via agglutination of a panel of red blood cells (RBCs) is inefficient and difficult to quantify. Using a baculovirus expression system to express Gp.Mur antigen on insect cell surfaces, we have developed a quantitative cell-based system to confirm the presence of anti-Gp.Mur antibody in human serum. We obtained 10 serum samples preidentified as having anti-Gp.Mur antibody and another 4 samples containing noncorresponding antibodies from hospital patients. Insect cells displaying Gp.Mur antigen successfully adsorbed anti-Gp.Mur antibody in the sera and inhibited the RBC agglutination mediated by this antibody. By varying the concentration of Gp.Mur-displaying cells, we could grade levels of RBC agglutination by anti-Gp.Mur antibody. Densitometric analysis further enabled quantitative determinations of hemagglutination inhibition by Gp.Mur-displaying cells. We believe that this cell-based hemagglutination inhibition system greatly improves or supplements existing technology and is a convenient means for accurately identifying and quantifying anti-Gp.Mur antibody.


Introduction
Gp.Mur antigen, previously known as Miltenberger antigen subtype III (Mi.III) or Mi a antigen, is one of several antigens of the MNS blood group system [1,2]. Gp.Mur antigen is a variant of the red blood cell (RBC) glycophorins in the MNS system, and it most likely evolved from crossover events between glycophorin A and glycophorin B [3,4]. In Southeast and South Asian human populations, including Taiwan, incidence of Gp.Mur antigen is higher (ranging from 2% to 7%) relative to white Caucasian (0.0098%) and Japanese (0.0068%) populations, making it one of the most clinically relevant antigens for blood transfusions in those regions [5][6][7][8][9]. Anti-Gp.Mur antibody is one of the most commonly identified RBC alloantibodies among thalassemia and multitransfused patients in Taiwan [10,11]. An aboriginal group in Taiwan, the Ami tribe, presents a markedly high incidence of this antigen (up to 88.4%), which has been credited to their athleticism [9,12]. Furthermore, the Yami and Puyuma tribes exhibit Gp.Mur antigen frequencies of 34.3% and presence in human sera but its further development could also assist in identifying other clinically significant alloantibodies from other blood group systems.

Generation of Baculovirus Expression Construct
The transfer vector [38] was constructed using pBacPAK8 (Clontech, Laboratories Inc., Fremont, CA, USA), with its polyhedrin promoter replaced with the TriEx promoter from pTriEx™-3 vector (Millipore, Burlington, MA, USA). A td-tomato reporter gene driven by an internal ribosome entry site (IRES) from Rhopalosiphum padi virus (RhPV-IRES) [39] was inserted into the vector to act as the fluorescence reporter. A nucleotide sequence encoding full-length Gp.Mur were synthesized according to GenBank accession number AF090739.1 from the National Center of Biotechnology Information (NCBI). Gp.Mur sequence corresponding to amino acids 1-122 was amplified by polymerase chain reaction (PCR) and subcloned into baculovirus transfer vector harboring the GP64 cytoplasmic tail domain (6C) along with a honeybee melittin signal peptide and a hexameric histidine tag (6H) at the N-terminal of Gp.Mur. The expression construct was generated using In-Fusion ® HD cloning kit (Clonetech, Laboratories Inc., Fremont, CA, USA) according to the manufacturer's manual. The Gp.Mur construct was cotransfected with FlashBAC™ baculovirus genome in SF21 cells to generate a recombinant baculovirus expressing Gp.Mur antigen.

Western Blot Analysis
The recombinant Gp.Mur construct was expressed in Hi5 cells by infecting cells with the recombinant baculovirus expressing Gp.Mur antigen using a multiplicity of infection of 1 (MOI = 1). After incubation at 26 • C at different time points, the cells were detached, and the culture medium was removed by low-speed centrifugation. Dulbecco's phosphatebuffered saline (DPBS) was added to remove the remaining culture medium and to wash the cells. The harvested cells were lysed using RIPA lysis and extraction buffer (Thermo Fisher Scientific, Waltham, MA, USA), and the cell lysates were then separated on a 10% sodium dodecyl sulfate-polyacrylamide gel (SDS-PAGE; Omic Bio, Taipei, Taiwan). Proteins on gels were transferred to a polyvinylidene fluoride (PVDF) membrane and Western blotted using mouse anti-His antibody (1:5000, GeneTex GTX628914; GeneTex Inc., CA, USA). As negative control, cells infected with wild-type baculovirus harboring the reporter gene alone (WT-td) was used simultaneously in the assays. As a loading control, glyceraldehyde 3-phosphate dehydrogenase (GAPDH) expression was determined using rabbit anti-GAPDH (1:5000, GeneTex GTX100118; GeneTex Inc., CA, USA).

Human Serum Samples
Human serum samples preidentified as having anti-Gp.Mur antibody by means of the polybrene method [40] were obtained from Linkou Changung Memorial Hospital, Taipei, Taiwan. The sera were blood bank samples that had undergone routine antibody identification as part of the antibody screening process for blood transfusion. Samples deemed positive for anti-Gp.Mur antibody were stored for future research use and included in the current study. Likewise, serum samples positive for antibodies for other blood groups, namely anti-Le b , anti-E, anti-D, and anti-S, were obtained to serve as controls. Upon receipt, the samples were re-tested for anti-Gp.Mur or antibodies of other blood groups using reagent RBCs containing the corresponding antigens (MeDiPro Antibody Screening Cell, Formosa, Biomedical Technology Corp., Taipei, Taiwan). The tests were performed separately on the immediate spin (IS) phase (at room temperature) and the antihuman globulin (AHG) phase (at 37 • C) [17].
All experimental procedures were approved by the Ethics Committees of Linkou Changung Memorial Hospital (IRB No. 201800282B0), Taiwan, and performed in line with the Declaration of Helsinki II principles. Written informed consent was obtained from all participants prior to enrollment in the study.

Preparation of Insect Cell Suspension
For hemagglutination inhibition assay using cell surface display of Gp.Mur antigen, Hi5 cells were infected with the recombinant baculovirus using MOI = 1 in a T150 cell culture flask. As a control, the WT-td baculovirus was used to infect the cells. At 3 dpi, the cells were detached, pelleted down, and the culture medium was removed. After washing twice with DPBS, cell numbers were counted before resuspending them in DPBS to a concentration of 5.0 × 10 6 cells/mL (Supplementary Figure S1). The cell suspension was then stored at 4 • C until use.

Hemagglutination Inhibition Assay
To perform the hemagglutination inhibition assay, varying amounts of insect cell suspension resulting in different cell concentrations (1.0 × 10 5 , 5.0 × 10 4 , and 2.5 × 10 4 ) were pipetted into separate microtubes containing 100 µL of human serum preidentified as harboring anti-Gp.Mur antibody (Supplementary Figure S1). The insect cell-serum suspension was agitated in a mechanical rotator at 1000 rpm for 25 min, after which it was centrifuged at 10,000 rpm for 5 min to separate the adsorbed serum from the insect cells. The adsorbed serum was then transferred to a separate test tube, and 50 µL of indicator RBC expressing Gp.Mur antigen (MeDiPro Antibody Screening Cell, Formosa, Biomedical Technology Corp., Taipei, Taiwan) was added. The RBC-serum suspension was then centrifuged for 30 s at 800 × g. After centrifugation, the cell pellet was gently dislodged and graded for agglutination [17]. For each serum sample, adsorption and hemagglutination inhibition were performed using the phase in which the serum sample showed greatest reactivity in the preliminary test, i.e., samples A-I were adsorbed at room temperature and the RBC agglutination test was done in the IS phase, whereas sample J was adsorbed at 37 • C and the agglutination test was done in the AHG phase. Presence of anti-Gp.Mur antibody was confirmed when serum samples adsorbed on cells displaying Gp.Mur antigen on their surfaces displayed agglutination inhibition or a reduction in the grade of agglutination relative to agglutination by WT-td-infected cells. Photographs of all test tubes used in agglutination assays were taken. To visualize agglutinates under microscopy and to quantify agglutination by means of densitometry, RBC agglutinates were carefully pipetted into 96-well plates and microscopic images were taken. The 96-well plates were placed into a Amersham TM Tyhoon 5 Imager (GE Healthcare, IL, USA) imager to conduct densitometric scans of the agglutinates. The density of the agglutinates as reflected in the scans was quantified using ImageQuant TL 8.2 software (GE, Healthcare, Il, USA) and expressed as signal intensity. The intensity of the agglutinates of the serum samples adsorbed by Gp.Mur-displaying insect cells was compared with that of the WT-td and expressed as % inhibition. To determine the specificity of hemagglutination inhibition, serum samples containing antibodies for other blood groups were used in the assay.

Stability Assay of Cells Displaying Gp.Mur Antigen
To determine the stability of Gp.Mur antigen expressed in insect cells, we performed hemagglutination inhibition assays as described above on the same serum samples when the cells had been freshly prepared and after the cells had been stored at 4 • C for 30 days.

Calculations and Statistical Analyses
For cell-based ELISA and stability assays, triplicate values were taken and error bars represent the standard deviation of the mean. For hemagglutination inhibition assay, duplicates for each level of treatment were obtained for each serum sample. For quantitative densitometry data, the values represent the average of duplicates and statistical bars represent the range of the values. To compare methods, Pearson's r was calculated in Prism software. To calculate the % inhibition, the visual grading and densitometry reading of the insect cells with the expressed Gp.Mur antigen was subtracted from the reading of the WT-td, and the resulting value was then divided by the reading of the WT-td and multiplied by 100.

Construction of Recombinant Baculovirus That Can Display Gp.Mur Antigen on Its Cell Surface
To display Gp.Mur antigen on insect cell surfaces, we generated a construct that encodes the Gp.Mur extracellular domain and its transmembrane domain fused with the AcMNPV 6C (6H-Gp.Mur-6C; Figure 1A). For control, we generated another recombinant baculovirus construct containing only the red fluorescence reporter gene, (td-tomato) wildtype-td (WT-td; Figure 1B). Expression of the construct was driven by a composite TriEx promoter, which is a combination of mammalian (CMVie), bacterial (T7), and baculovirus (p10) promoters. We fused a honeybee melittin signal peptide to the N-terminal of the Gp.Mur construct to enable extracellular secretion of the recombinant proteins. We assumed the recombinant protein would be anchored on the plasma membrane of the insect cell membrane via the transmembrane domain of Gp.Mur antigen fused with the 6C. A tdtomato fluorescence gene was inserted downstream of the Gp.Mur construct, which was driven by an IRES sequence. The td-tomato fluorescence signal therefore reflected successful baculovirus infection of cells and successful expression of Gp.Mur recombinant protein.
After successful cotransfection of the Gp.Mur construct with the baculovirus genome, the resulting recombinant baculovirus harboring the construct was named v6H-Gp.Mur-6C.

Expression and Surface Display of Gp.Mur Antigen by Recombinant Baculovirus in Insect Cells
To determine if the recombinant baculovirus could express Gp.Mur antigen in Hi5 cells, we infected the cells at different time points. Cells were lysed, and Gp.Mur expression was detected by Western blot analysis using anti-His antibody ( Figure 2A). The data showed that the optimum time point of protein expression was 3 dpi. To better estimate the levels of Gp.Mur surface display in Hi5 cells, we fixed the infected Hi5 cells in the wells of a 96-well cell culture plate and determined antigen-anti-His antibody interactions by means of ELISA. This cell-based ELISA revealed that 6H-Gp.Mur-6C exhibited substantial protein expression relative to WT-td and CO ( Figure 2B).

Expression and Surface Display of Gp.Mur Antigen by Recombinant Baculovirus in Insect Cells
To determine if the recombinant baculovirus could express Gp.Mur antigen in Hi5 cells, we infected the cells at different time points. Cells were lysed, and Gp.Mur expression was detected by Western blot analysis using anti-His antibody ( Figure 2A). The data showed that the optimum time point of protein expression was 3 dpi. To better estimate the levels of Gp.Mur surface display in Hi5 cells, we fixed the infected Hi5 cells in the wells of a 96-well cell culture plate and determined antigen-anti-His antibody interactions by means of ELISA. This cell-based ELISA revealed that 6H-Gp.Mur-6C exhibited substantial protein expression relative to WT-td and CO ( Figure 2B).

Expression and Surface Display of Gp.Mur Antigen by Recombinant Baculovirus in Insect Cells
To determine if the recombinant baculovirus could express Gp.Mur antigen in Hi5 cells, we infected the cells at different time points. Cells were lysed, and Gp.Mur expression was detected by Western blot analysis using anti-His antibody (Figure 2A). The data showed that the optimum time point of protein expression was 3 dpi. To better estimate the levels of Gp.Mur surface display in Hi5 cells, we fixed the infected Hi5 cells in the wells of a 96-well cell culture plate and determined antigen-anti-His antibody interactions by means of ELISA. This cell-based ELISA revealed that 6H-Gp.Mur-6C exhibited substantial protein expression relative to WT-td and CO ( Figure 2B).   Samples obtained from the blood bank of Linkou Changung Memorial Hospital, preidentified as containing anti-Gp.Mur antibody, were validated as having that antibody at the IS and AHG phases using reagent RBCs expressing Gp.Mur (MeDiPro Antibody Screening Cell). Most of the serum samples presented a strong hemagglutination reaction at the IS phase (grades ranging from 3+ to 4+), except for sample J (grade 2+) ( Table 1). Sample J exhibited a higher grade of 3+ at the AHG phase, perhaps because it had been exposed to Gp.Mur antigen more than once [41].

Adsorption Using Gp.Mur Antigen Expressed in Insect Cells
Because RBCs host multiple antigens, a positive hemagglutination result means that the serum sample contains antibody or antibodies to at least one of the many antigens on the surface of the RBCs. Suspected antibody identity must be confirmed by further phenotyping a patient for the absence of corresponding RBC antigens or by reacting the serum with another batch of reagent RBCs that host the corresponding antigens. However, if a suspected antibody is adsorbed using a specifically expressed antigen and then subjected to an RBC hemagglutination assay, that would directly confirm antibody identity. Here, we expressed Gp.Mur antigen on the surface of insect cells, acting as the adsorbing material, to confirm the presence of anti-Gp.Mur antibody in serum samples. Varying concentrations of insect cells were added to serum samples and then indicator RBCs containing the target antigen were added (Supplementary Figure S1). Resulting RBC hemagglutination was visually graded [17].

Grading of Hemagglutination
To determine if adsorption of anti-Gp.Mur antibody by Gp.Mur antigen expressed in insect cells inhibited hemagglutination, we first employed the conventional grading system that ranks hemagglutination on a scale of 0 (no agglutination) to 4+ (a single solid agglutinate) [17]. This approach revealed varying degrees of hemagglutination among samples, though all 10 samples presented inhibition relative to WT-td cells. Hemagglutination inhibition assays were performed at the IS phase for samples A-I, whereas it was performed at the AHG phase for sample J. We observed a dose-response inhibitory effect (Figure 3), with the highest cell concentration (1.0 × 10 5 ) resulting in almost complete inhibition of hemagglutination (Figure 3; samples A, C, and H). However, this system of grading is highly subjective, especially when evaluating lower grades of agglutination.

Quantitative Grading of Hemagglutination
To better quantify the degree of hemagglutination inhibition by the adsorbing cells and given the dose-response effect we observed, we adopted a densitometric approach to quantitatively analyze hemagglutination in detail. We carefully pipetted the RBC agglutinates into 96-well plates and took images under microscopy (Supplementary Figure  S2). The plates were then placed in an Amersham TM Typhoon 5 Imager for densitometric scanning of the agglutinates. The scans were quantified using ImageQuant TL 8.2 software (Figure 4). Quantitative measurements of the intensity of the densitometric scans supported the existence of a dose-response inhibitory effect, with the highest insect cell concentrations resulting in the strongest inhibition ( Figure 4 and Table 1).

Comparison of the Grading Methods
These assays demonstrated that RBC hemagglutination was inhibited when anti-Gp.Mur antibody in the sera was adsorbed by insect cells expressing Gp.Mur antigen. At a cell concentration of 1.0 × 10 5 , all samples presented an agglutination grade of 0 to 1+, with an average % inhibition of 82%. At 5.0 × 10 4 cells, sera presented agglutination grades of 0 to 3+ and an average % inhibition of 55%. At 2.5 × 10 4 cells, agglutination grading ranged from 1+ to 4+, with an average % inhibition of 28%. When we assayed %inhibition more quantitatively by means of densitometric analysis, we recorded 89%, 64%, and 27% inhibition for insect cell concentrations of 1.0 × 10 5 , 5.0 × 10 4 , and 2.5 × 10 4 , respectively (Table 1). To compare these two grading approaches, we used a Pearson's coefficient correlation (Pearson's r) to assess their relationship. We observed moderate positive correlations (r = 0.63 for 1.0 × 10 5 , r = 0.67 for 5.0 × 10 4 , and r = 0.58 for 2.5 × 10 4 ) between the two grading approaches, indicating that they were comparable. We compared the means of % inhibition generated according to the two grading approaches by means of paired Student t-test and the results were not significant (p > 0.05).

Quantitative Grading of Hemagglutination
To better quantify the degree of hemagglutination inhibition by the adsorbin and given the dose-response effect we observed, we adopted a densitometric appr ( Figure 4). Quantitative measurements of the intensity of the densitometric scans supported the existence of a dose-response inhibitory effect, with the highest insect cell concentrations resulting in the strongest inhibition ( Figure 4 and Table 1).

Determining the Specificity of Gp.Mur Antigen Expressed in Insect Cells
Because there are numerous RBC antigens against which antibodies can develop, the binding specificity of recombinant RBC antigens to their corresponding antibodies is very important. In the above-described experiments, we have shown that anti-Gp.Mur antibody can recognize Gp.Mur antigen expressed in insect cells. We also wanted to determine if the Gp.Mur antigen expressed in insect cells binds nonspecifically to other RBC antibodies because such cross-reactivity could lead to inaccurate or confusing results with significant medical implications. To do that, we used serum samples containing antibodies from different blood groups, namely anti-Le b , anti-E, anti-D, and anti-S. The S antigen is highly homologous to the Gp.Mur antigen, so cross-reactivity with it may compromise test results if both anti-S and anti-Gp.Mur antibodies are present in a serum sample. We subjected these sera to the same procedures for antibody adsorption, grading, and quantification as conducted on sera preidentified as having anti-Gp.Mur antibody. Hemagglutination inhibition assays showed no observable or quantitative differences in levels of inhibition between WT-td-infected cells and v6H-Gp.Mur-6C-infected cells ( Figure 5; Table 2), so our Gp.Mur antigen was highly specific.

Stability of the Gp.Mur Antigen Expressed in Insect Cell
We wanted to assess if the Gp.Mur antigen expressed in insect cells is stable under conditions of long-term storage. For this experiment, we used freshly prepared suspensions of insect cells expressing Gp.Mur antigen, some of which was analyzed immediately and the remainder stored at 4 • C for 30 days and then subjected to the same analyses. There were no differences in the degree of hemagglutination inhibition between freshly prepared and stored samples (p > 0.05) (Supplementary Figure S3).

Discussion
In this study, we report a practical use of insect cell surface antigenic display for hemagglutination inhibition assay. We have demonstrated that insect cells displaying Gp.Mur antigen on their surfaces can substantially inhibit anti-Gp.Mur antibody in serum samples from patients preidentified as having that antibody. Moreover, as a significant advantage over the conventionally adopted tube method, we could quantitatively assess levels of hemagglutination inhibition by means of densitometry using our insect-cell-based system. Thus, our method can replace or supplement the conventional confirmatory method that involves using several RBC reagents.
We expressed Gp.Mur antigen on both Hi5 insect cells via recombinant baculovirus. We generated v6H-Gp.Mur-6C, which displayed a high level of surface expression of the antigen, demonstrating that the baculovirus 6C provides anchorage of the protein on insect cells. Baculovirus expression vector systems (BEVS) are superior to other expression systems such as yeast and bacteria as they exhibit better capacity for retaining post-translational modifications, protein oligomerization, and membrane protein expression [33]. Because our expressed recombinant protein was well recognized by the anti-Gp.Mur antibody in the human serum samples, as demonstrated by our hemagglutination inhibition assays, the recombinant Gp.Mur antigen is strongly expressed in insect cell membranes and with an appropriate antigenic configuration.
Although post-translational modifications (PTM) in insect cells differ slightly from mammalian cells, such as the inability of insect cells to produce end products containing terminal galactose and sialic acid [30], thousands of functional recombinant proteins, including human and animal-based vaccines, have been developed using BEVs [42]. Furthermore, based on our data, our recombinant Gp.Mur was able to maintain its antibody-binding capability as shown in our hemagglutination inhibition assay results.
The conventional method of RBC antibody identification makes use of a panel of phenotyped human RBCs. Because RBCs contain multiple antigens, a positive agglutination reaction means that antibodies exist in the sera against one or several antigens on the RBCs. As an indirect means to identify the antibody in a serum sample, a panel of human RBCs is typically used so that antigens from the RBCs that do not show an agglutination reaction with the serum can be ruled out. However, this method is ineffective in cases of autoantibodies, antibodies against high-frequency antigens, and antibodies that display a dosage effect. Given the limitations of this conventional identification system, a binary "yes or no" system has been proposed [18], which can only be achieved if the relevant RBC antigens are expressed individually. Previously, researchers have expressed and purified several of the relevant antigens using mammalian expression systems [20,21]. However, protein purification is tedious and requires up to 5 µg of the purified protein to inhibit the corresponding antibodies [19,23]. In the current study, we used a baculovirus system for high-level protein expression. Further development of our insect-cell-based system would circumvent the cost and effort associated with protein purifications. To the best of our knowledge, no insect cell-based hemagglutination inhibition assay for identifying human RBC antibodies has been reported previously.
Insect cell surface display of Gp.Mur antigen enabled specific recognition of the anti-Gp.Mur antibody in our patient serum samples. We also conducted hemagglutination inhibition assays on our recombinant baculovirus for other antibodies (anti-Le b , anti-E, anti-D, and anti-S), but the recombinant Gp.Mur antigen expressed in insect cells could not recognize those antibodies. Because S antigen shares conserved sequences with Gp.Mur antigen, the fact that our Gp.Mur antigen expressed in insect cells could not inhibit the hemagglutination activity of anti-S antibody strongly suggests that our recombinant Gp.Mur antigen is highly specific for anti-Gp.Mur antibody. Such specificity is highly important for identifying RBC antibodies as it informs which blood or blood products can be transfused into patients. Thus, the enhanced specificity of hemagglutination inhibition by recombinant RBC antigens could greatly aid in identifying the presence not only of the corresponding antibodies but also of antibodies of other specificities. Significantly, our recombinant Gp.Mur antigen expressed in insect cells is very stable, fully retaining its inhibitory characteristics even after storage at 4 • C for 30 days.

Conclusions
Anti-Gp.Mur antibody is one of the most commonly identified RBC antibodies in humans in Taiwan [6,10,43], and it may be highly prevalent in Southeast Asia, South Asia, and some parts of East Asia [7], although detailed prevalence and incidence studies on this antibody are lacking. This antibody has been implicated in hemolytic transfusion reactions and hemolytic disease of the fetus and newborn [14][15][16]. The technology that we have developed to express Gp.Mur antigen on insect cell surfaces by means of BEVS can be used to confirm and quantify anti-Gp.Mur antibody in human serum samples. The current approach for confirming anti-Gp.Mur antibody presence involves agglutinating a serum sample suspected of hosting anti-Gp.Mur antibody with another RBC reagent harboring Gp.Mur antigen. Hemagglutination inhibition using purified recombinant RBC antigen has been demonstrated to aid in the identification of the corresponding antibody as well as other nonspecific antibodies [19,21,23]. Our insect cell surface display technology for Gp.Mur antigen represents a convenient supplementary or replacement approach for the critical confirmation of anti-Gp.Mur antibody in human sera.