Kai 1 and Kai 2: Characterization of these dog erythrocyte antigens by monoclonal antibodies

Dog Erythrocyte Antigens (DEA) have thus far been found by sensitizing dogs with canine allogeneic blood and are clinically important regarding blood transfusion incompatibilities, but remain poorly defined. The goals of this study were to discover and characterize two DEAs, named as Kai 1 and Kai 2. The monoclonal antibodies were produced by mouse hybridoma techniques and examined by ELISA isotyping, immunoblotting, and affinity chromatography. Canine blood samples were typed and the development of alloantibodies was examined in transfused dogs. The monoclonal Kai 1 and Kai 2 antibodies were isotyped as IgM kappa and IgG3 lamda, respectively, and identified two different erythrocyte membrane proteins of 200 kDa and 80 kDa in molecular weights, respectively. Either Kai 1 or Kai 2 can be expressed but not both in an individual dog. There were no naturally occurring anti-Kai 1 or Kai 2 alloantibodies. In addition, Kai 1- and/or Kai 2- dogs developed Kai 1 and Kai 2 alloantibodies, respectively, when transfused with mismatched blood. This is the first discovery of canine blood types by screening monoclonal antibodies. Kai 1 and Kai 2 are novel blood types which can induce anti-Kai 1 or anti-Kai 2 alloantibodies when Kai 1- and/or Kai 2- dogs are transfused with Kai 1+ or Kai 2+ blood. These canine blood types may explain some of the blood incompatibilities and transfusion reactions observed in dogs in clinical practice.


Introduction
Dogs have been used in early xenotransfusions to humans as well as animal models to characterize transfusion reactions [1]. In veterinary clinical practice, when anemic or bleeding dogs are transfused, blood type incompatibilities have been documented based upon hemolytic transfusion reactions and incompatible agglutination crossmatch test results [2,3]. Based upon experimentally sensitizing dogs with canine blood transfusions, eight Dog Erythrocyte Antigens (DEA) were classified with polyclonal alloantibodies by an international committee in 1974, but currently only reagents for DEA 1, DEA 3,DEA 4, and DEA 7 are commercially available. Dogs are either positive or negative for a particular blood type except for the DEA [3,4]. Recent flow cytometry and strip kit typing studies with monoclonal antibodies reveal that dogs are either DEA 1-or weakly to strongly DEA 1+ [5], but the biochemical and molecular basis remains elusive for DEA 1 and other DEAs. Additionally, non-officially classified red blood cell (RBC) antigens have been identified in dogs, including Dal (Dalmatian) [6,7] and Shigeta dog blood types [8]. Furthermore, dogs with acute hemolytic transfusion reactions following a second transfusion [2,9] as well as major crossmatch incompatibilities after a first transfusion have been reported [3], indicating the development of alloantibodies to unknown RBC antigens different from conventional DEAs.
By using mouse hybridoma technique, we obtained two monoclonal antibodies against two erythrocytic membrane antigens which did not coexist in any dogs but could both be absent. The two monoclonal antibodies (mAb) and corresponding antigens, named Kai 1 and Kai 2 (meaning 'dog' in Korean), were biochemically characterized by enzyme-linked immunosorbent assay (ELISA), immunoblot and affinity chromatography. Their antigenicity was examined in dogs after receiving inadvertently mismatched transfusions.

Animals and samples
Female mice (BALB/c, 6 weeks old, Daehan Animal, Seoul, South Korea) were used. All mice had free access to food and water and were maintained on a 12 hr light-dark cycle. Ethylenediaminetetraacetic acid (EDTA) blood samples from non-transfused and previously transfused dogs were obtained from small animal hospitals and Korea Animal Blood Bank (KABB, Sokchosi, South Korea). We obtained written consent from the dog owners to allow for experimental use of the laboratory results of blood samples collected during medical treatment (S1 file). All animal procedures were approved by the Institutional Animal Care and Use Committee at the Daegu Haany University.
Screening and production of monoclonal antibodies against canine blood: Kai-1 and Kai-2 Hybridomas producing monoclonal antibodies were made according to the previous method [10,11] with slight modifications. In brief, mice (n = 6) were sensitized with canine RBC from two Korean Mastiff dogs: 0.4 μM aminopterin, and 16 μM thymidine (Sigma), the hybridoma cells causing agglutination specific to the blood used for sensitization were selected in direct hemagglutination assays. The positive hybridoma cells were then inoculated intraperitoneally into pristine-primed female BALB/c mice (n = 6) to yield the ascites containing the monoclonal antibody.

Purification of antibodies
Ascite was filtered through a 0.45 μm syringe filter and 5-fold diluted with PBS, before passing the fluid through protein L or protein G columns (GenScript, Piscataway, NJ, USA). The column resins were washed 3 times with PBS, and immunoglobulins were eluted with 0.1 M glycine (pH 3.0) buffer and neutralized by addition of 1M Tris-HCl (pH 8.0). The purified monoclonal antibodies were further diluted with PBS and analyzed by bicinchoninic acid (Sigma) and sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) assays by comparing them to mouse IgG (Abcam, Cambridge, UK).

Isotype determination of Kai antibodies
ELISA plates coated with goat anti-mouse heavy-chain antibodies (IgG1, IgG2a, IgG2b, IgG3, IgA and IgM) or goat anti-mouse light-chain antibodies (kappa or lambda) were used (Thermo Scientific, Rockford, IL, USA). The purified antibodies against the two canine blood samples were added to each well and the isotypes were then spectrophotometrically determined by measuring the absorbance at 450 nm above 0.2 optical densities.

Western blotting with canine RBC membrane proteins
Canine RBC membrane proteins were precipitated using a glycoprotein isolation kit (Thermo Scientific) and the protein concentrations were determined (Bio-Rad DC protein assay kit II, Bio-Rad, CA, USA). The lysates containing 50-100 μg of protein were mixed with 4x NuPA-GE-LDS sample buffer, and the proteins were separated through NuPAGF™ Bis-Tris gels with 1x NuPAGE MES SDS running buffer (Thermo Scientific, Rockford, IL, USA). The gels were electro-transferred onto a Hybond ECL transfer membrane with transfer buffer (25 mM Tris, 250 mM glycine, 20% methanol) at 300 mA for 90 min. To block nonspecific binding sites, the membranes were treated with 5% nonfat dry milk in Tris-buffered saline with Tween 20 (TBST) for 90 min at room temperature (about 22˚C). The membranes were incubated overnight with Kai 1 (1:250 titer) or Kai 2 (1:100) antibody in TBST containing 3% nonfat dry milk at 4˚C. The membranes were washed with TBST three times for 30 min and incubated with goat anti-mouse IgM or IgG HRP conjugated secondary antibody (1:3000) in TBST that contained 3% nonfat dry milk for 90 min at room temperature. After incubation with secondary antibody, the membranes were washed with TBST three times for 30 min at room temperature. The proteins were detected by an enhanced chemiluminescence detection system.

Affinity chromatography of Kai antibodies
Affinity purification was accomplished by using Affi-Gel affinity chromatography according to manufacturer's instructions (Bio-Rad, CA, USA). Briefly, 1 mg of anti-Kai 1 or anti-Kai 2 antibodies were added into columns filled with Affi-Gel slurry. After overnight incubation at 4˚C, the column resins were blocked with 100 mM ethanolamine and washed three times with PBS. Affinity proteins for IgM or IgG were purified using the same columns as those used for antibody purification.

Comparison of canine blood types between Kai types and commercially available DEA 1 test kit
To determine any relationship between Kai types and DEA 1, DEA 1+ or DEA 1-samples (total n = 50), typed with Rapid Vet-H DEA 1 card test (DMS Laboratories, Flemington, NJ, USA), were compared to typing results with Kai antibodies.

Alloantibody detection in transfused dogs
Blood samples from dogs (n = 4) were tested in a tube agglutination assay applying the anti-Kai 1 and anti-Kai 2 antibodies as well as for the presence of anti-Kai 1 and/or anti-Kai 2 antibodies. In brief, blood samples were obtained from regular blood donors as well as recipient dogs transfused with Kai 1+ or Kai 2+ blood. The samples were washed in saline and resuspended to a 2~3% suspension in physiological saline solution. In a 3 ml plastic tube, equal volumes of plasma and RBC suspension were mixed and incubated at 37˚C for 15 min. After centrifuging for 1 min at 1000g, the tubes were gently mixed to dislodge the pelleted RBCs. The degree of agglutination strength was macroscopically graded from negative (0) to positive (4+).

Production of Kai 1+ and Kai 2+ monoclonal antibodies
Hybridoma cells from a total of 4800 wells were screened for specificity to the canine blood used to sensitize mice. The cell lines were discarded if they agglutinated non-specifically or cross-reacted to all RBCs. We isolated two hybridoma cell lines which produced different, but not cross-reacting, antibodies against the RBCs from the two dogs typed as DEA 1.

Isotypes of purified anti-Kai antibodies
Using the Proteins L and G columns and SDS-PAGE, the anti-Kai 1 immunoglobulins (Ig) were captured by the Protein L column, showing distinct bands at~80 kDa and~25 kDa, reflecting the heavy and light chain Ig, different from mouse IgG, thereby suggesting the presence of an IgM (Fig 1A). On the other hand, the anti-Kai 2 Ig were captured by the Protein G column and revealed~50 kDa and~20 kDa bands identical to the heavy and light chain of murine IgG, respectively ( Fig 1B). Moreover, by ELISA pre-coated with goat anti-mouse heavy-chain antibodies (IgG1, IgG2a, IgG2b, IgG3, IgA and IgM) or goat anti-mouse light-chain antibodies (kappa or lambda), the anti-Kai 1 antibody was isotyped as IgM with kappa light chains, whereas anti-Kai 2 antibody reacted with antibodies against IgG3 heavy and lamda light chain. These results indicate that anti-Kai 1 and anti-Kai 2 antibodies belong to subclasses of IgM kappa and IgG lamda, respectively (Fig 2).

Distribution of Kai types in DEA 1+ and DEA 1-dogs
Utilizing these two mAbs in a conventional tube agglutination typing assay produced either strongly positive (3+/4+) or negative (0)

Identification of the Kai 1 and Kai 2 antigens by Kai mAbs
By immunoblotting under reducing conditions, the anti-Kai 1 mAb detected specifically a major band >170 kDa and two minor bands of~55 and~70 kDa in samples from Kai 1+ but not Kai 1-blood (Fig 3A). With the anti-Kai 2 mAbs, only one major band between 70 and 130 kDa was seen with samples from Kai 2+ dogs (Fig 3B). For more precise identification of protein bands specific to Kai 1 or Kai 2 antibodies, affinity chromatography was performed and revealed a major band of~200 kDa and split bands~50 kDa with the anti-Kai 1 mAb. On the other hand, anti-Kai 2 mAbs bound a prominent band at~80 kDa (Fig 4).

Detection of alloantibodies against Kai 1 and Kai 2
Based upon major crossmatch tests with Kai 1 and Kai 2 mismatched blood, pre-transfusion compatibility testing did not show any agglutination reactions. However, when crossmatching plasma from previously transfused dogs~21 days after transfusion, +1 to +3 agglutination reactions were observed when the recipient was Kai 1-or Kai 2-, and the donor was Kai 1+ or   Kai 2+ (Table 2). Those recipients were DEA 1+ but not typed for other blood groups. Many blood clots were microscopically seen in a Kai 2+/DEA 1- (Fig 5A) and a Kai 1+/DEA 1+ blood-transfused dog (Fig 5B).

Discussion
While based upon experimental and clinically mismatched transfusions more than a dozen blood groups have been described in dogs, the nature of these canine alloantibodies and RBC antigens have remained mostly elusive. We produced two distinct monoclonal antibodies which are strongly agglutinating and bind two apparently new dog erythrocyte antigens named Kai 1 (200 and 50 kDa) and Kai 2 (80 kDa). The Kai 1 and Kai 2 antibodies were isotyped as IgM kappa and IgG3 lamda, respectively, and identified two different dog erythrocyte membrane proteins of 200 kDa and 80 kDa, respectively. We show that Kai 1-and/or Kai 2dogs developed anti-Kai 1 and anti-Kai 2 alloantibodies, respectively, when transfused with mismatched blood demonstrating the clinical importance of these Kai blood groups. In veterinary transfusion medicine, hybridoma techniques have thus far only been applied to produce mAbs against canine DEA 1.1 (later named DEA 1) and DEA 3 antigens, feline A and B antigens and the equine Ca antigen [11][12][13][14]. In our attempt to produce mAbs against DEA 1.1 and DEA 1.2 antibodies reported here, we generated and characterized two mAbs against canine erythrocyte antigens, named Kai 1 and Kai 2. The anti-Kai 1 mAb isotyped by ELISA as IgM kappa and was a much stronger agglutinin (1:2056) compared to the anti-Kai 2 IgG3 lamda (1:32) against the Kai+ erythrocytes. Indeed, IgG mAbs can also effectively agglutinate erythrocytes [15], are more stable than IgM in accelerated storage tests, and frequently exhibit higher affinity than IgM [16]. The two anti-DEA 1 mAbs used in typing kits are either IgM (DMS laboratories Inc., Tempe, AZ) or IgG (Alvedia, Lyon, France) [17]. The anti-DEA 3 mAb was not characterized and is currently not available for typing. And all other alloantibodies for typing dogs are polyclonal antibodies generated accidently or experimentally by sensitizing dogs by RBC transfusions [3].
The generated anti-Kai 1 and anti-Kai 2 alloantibodies cause strong agglutination reactions in the tube, as shown here, and gel column tests used in our recent typing survey [18]. While both techniques readily differentiate Kai+ and Kai-RBCs; the gel column technique is simpler to read and photographically capture for archiving, but requires a special centrifuge. Thus, these Kai mAbs are suitable as typing reagents in clinical settings. In the surveys reported here from Korea (with mostly Mastiffs as blood donors typed) and previously from North America  [18], Kai 1+/Kai 2-(42% in Korea; 94% in North America) dogs were commonly found, while Kai 1-/Kai 2+ (37% in Korea; 1% in USA) and Kai 1-/Kai 2-(20% in Korea; 5% in North America) were less frequently observed. The variation in the two surveys regarding the prevalence of each Kai type may be related to differences in geographical areas and breeds typed as well as the relatively small size of the surveys. Noteworthy, neither survey detected any dogs which were Kai 1+/Kai 2+ suggesting the two antigens may not be coexisting in any dogs. When comparing the Kai typing results and DEA 1 with available mAbs from the small survey of 50 dogs reported here and >500 dogs from North America reported recently [18], it is evident that Kai 1+ or Kai 2+ were more frequently DEA 1+, but could also be rarely DEA 1-. Similarly, Kai 1-and/or Kai 2-dogs could be DEA 1+. Thus, the anti-Kai and anti-DEA 1 mAbs do not recognize the same blood types in dogs. This was surprising as the Kai mAbs were generated against supposedly against DEA 1.1+ and DEA 1.2+ blood. The blood initially used for sensitizing mice was typed with reagents from polyclonal antisera (KABB) and Shigeta DEA 1.1+ (1.1B+) and as DEA 1.2+ (1.2B+). Unfortunately, neither the original dogs used for sensitization nor the Shigeta typing reagents were available to further compare and investigate these apparent discrepancies. Furthermore, in our previously reported survey, a relationship between DEA 3, DEA 4 and DEA 7 as well as Dal could not be identified [18]. Also, it has become evident that the DEA 1 blood group system is composed of varied DEA 1 types and not DEA 1.1 and 1.2 (see below) Despite considerable research efforts over the past decades, the nature of the canine blood group antigens remains largely unknown. Some DEAs have been assigned specific molecular weights, but those studies have not progressed or confirmed by others. In affinity immunochromatographic assays reported here, the anti-Kai 1 mAb identified a protein of~200 kDa and~50 kDa (two split bands), while the anti-Kai 2 mAb detected a protein of~80 kDa. These proteins have not been further purified and characterized by sequencing. Interestingly, the anti-Kai 1 and anti-Kai 2 binding proteins of 200/50 kDa and 80 kDa concur with earlier protein studies for DEA 1.1 [11] and DEA 1.2 [19], respectively, despite the above discordant Kai and DEA 1 blood typing results. However, it would be highly unusual to have two different proteins for DEA 1.1 and DEA 1.2 be part of the same blood group system. Recent studies using the available anti-DEA 1 mAbs failed to confirm these earlier protein studies on DEA 1 [17]. In the meantime, further flow cytometric studies with currently available anti-DEA 1 mAbs have shown that DEA 1.1 and DEA 1.2 are the same antigen and did not correlate to the variations in the expression of DEA 1 with DEA 1-and weakly, moderately and strongly DEA 1 + dogs [17]. Furthermore, a preliminary genome wide association study found a single nucleotide polymorphism in a region on chromosome CFA27 for DEA 1 [20]. Further biochemical and molecular genetic studies will be needed to characterize and clarify the relationship of the proteins found by immunoblotting and reactions with various monoclonal and polyclonal antibodies. Maybe the Kai 1 and Kai 2 may correspond to the originally described erythrocyte proteins DEA 1.1 and DEA 1.2 and the currently used anti-DEA 1 mAbs recognize another RBC antigen.
While originally debated dogs generally do not have naturally occurring alloantibodies to known blood groups [4] and pregnancy does not induce alloantibodies due to their endoepithelial placenta which is impermeable to proteins [21]. Similarly, we show here as well in the previous study that Kai 1-and/or Kai 2-dogs do not have anti-Kai 1 or anti-Kai 2 alloantibodies in their plasma, respectively. However, when evaluating the plasma from a small number of previously transfused DEA 1+ dogs, we discovered anti-Kai 1 and/or anti-Kai 2 alloantibodies in plasma from those typed as Kai 1-and/or Kai 2-. This documents the antigenic nature of both of these erythrocyte membrane proteins and thereby making Kai 1 and Kai 2 qualify as new independent blood group systems.
However, the clinical importance of developing anti-Kai 1 and/or anti-Kai 2 alloantibodies still needs to be determined by showing accidental or experimental acute hemolytic transfusion reactions in previously transfused Kai 1-and/or Kai 2-dogs receiving a second transfusion more than four days apart. The development of post-transfusion alloantibodies has been shown clinically in a few surveys [3,22], but only in rare occasions the causative blood type mismatch such as DEA 1.1, DEA 4 and Dal, was found. As the polyclonal anti-Dal [23,24] and monoclonal anti-Kai reagents become more available for typing and recipients can be typematched beyond the currently recommended DEA 1, sensitization and transfusion reactions can be further avoided. In the meantime, the current recommendation of transfusing DEA 1or DEA 1 matched blood at time of first transfusion initially and also crossmatch compatible blood for any subsequent transfusion (>4 days) may be extended to including typing for Kai dogs with incompatible crossmatch to find more readily compatible donors which had been also typed from Kai [18]. It should be noted that despite 36 blood group system and 352 antigens in humans [25], routine typing remains fairly restricted to ABO and Rh (and possibly Kell) unless in special circumstances or when the patient is found to be crossmatch incompatible [26].
In conclusion, two blood groups named Kai 1 and Kai 2 were identified and characterized with mAbs which differ based upon typing surveys from DEA 1 and other known and tested blood types. Affinity immunoblots identified distinctly different erythrocyte membranes which coincide with earlier described DEA 1.1 and DEA 1.2 antigen sizes. And while the antigenicity of these antigens was shown in crossmatch tests, the clinical importance of these induced alloantibodies needs to be still determined. Blood typing beyond DEA 1 typing including Kai and Dal typing may be helpful to explain blood type incompatibilities in previously transfused dogs and donors.