The gel test: a new way to detect red cell antigen-antibody reactions

Y. LAPIERRE, D. RIGAL,J.
ADAM,

D. JOSEF,F. MEYER, s. GREBER, AND

c. DROT

is described. It is applicable to most of the tests performed in blood group serology. The procedures are standardized and easy, and they provide clear and stable reactions that improve the interpretation of results. The process uses s ecial microtubes filled with a mixture of gel, buffer, and reagent. Depending on t[e test to be carried out, the test uses a neutral gel containing no reagents (reagents are added to top of gel) or a specific gel containin reagents (e.g., antiglobulin serum or antiA, -8,-D, etc.). A suspension of RBCs (or typing or the direct antiglobulin test) or a mixture of RBCs and serum (for reverse ABO typing or antibody characterization) is centrifuged through the gel under precise conditions. In negative reactions, the RBCs pass through t h e gel and pellet in the bottom of the tube, whereas, in positive reactions, they are trapped in the gel and the reaction may be read for hours afterwards. The test is easy to perform, sensitive, and reproducible. The antiglobulin tests can be performed without washing of the RBCs. There should be a reduction of risk from biohazardous materials. TRANSFUSION 1990;30:109-113.
A new process for the detection of red cell (RBC) antigen antibody reactions

RED CELL (RBC) antigen-antibody reactions are usually detected by agglutination tests, in saline or macromolecular media, with unmodified or enzyme-treated RBCs, and with or without the use of low ionic media, antiglobulin sera (AGS), and/or potentiators and polycat ions. One of the most important factors affecting the accuracy of results is the reading of the reaction, especially when it is weak. To obtain good results, the reaction must be examined by a qualified person within a short space of time, and, even then, it is sometimes difficult to interpret. The gel test was developed initially to standardize agglutination reactions and to fix the agglutinates to allow a simple and reliable reading. This aim was reached by a new process, patented in 1986, in which RBCs are centrifuged through a gel contained in a special microtube. The gel acts as a trap: the free RBCs pellet in the bottom of the tube, while the agglutinates are trapped (fixed) in the gel for hours. The gel test was sensitive and easy to perform: a reagent (e.g., AGS) could be added directly to the gel, and the antiglobulin test could be performed without any washing of RBCs. After 3 years of research and development, a commercial kit was developed. As this is a new test, the main purpose of this article will be to give a general view of the test, rather than an extensive study of its results. Some data will be given, to illustrate some of its applications.
From the Cenfre Regional de Transfusion Sanguine, Lyon, France, and Diamed SA, Murtcn, Switzcrland. Received for publication March 1, 1989; revision received July 27,

Materials and Methods Principle of the gel test

The gel test may be considered as a new way to obtain and to read agglutination. It can be substituted for many traditional tests, unless they require a stage of dispersion of the agglutinates, as do the hexadimethrine bromide (Polybrene, Sigma Chemical Company, St. Louis, MO) meth~ds.~. Three types of gel tests will be described: neutral, specific, and antiglobulin. In these tests, agglutination occurs, during a centrifugation stage, in a gel contained in a microtube. In the neutral gel test, the gel does not contain any specific reagent, but acts only by its property of trapping agglutinates. The RBCs, or the mixture of RBCs and serum, are dispensed in the upper part of the microtube above the gel. An incubation is performed if necessary, and then the microtube is centrifuged under precise conditions. After centrifugation, the negative reactions are clearly different from the positive ones, which vary depending on their strength (Fig. 1). The main applications of neutral gel tests are antibody screening and identification with enzyme-treated or untreated cells, and reverse ABO typing. In the specific gel test, the microtube contains a mixture of gel and specific reagent (e.g., anti-A, -B,-AB, -D, -C, -C, -E, -c, -e, -K). Such gels are useful for antigen determination. The unknown RBCs are suspended in an enzyme solution, which is dispensed on the gel, and centrifuged. At the beginning of the centrifugation, the RBCs will come in contact with the specific reagent and will be agglutinated and trapped in positive reactions; in negative reactions, they will pellet. The specific gels are calibrated for use with enzyme-treated cells, whenever possible, to improve the reaction (e.g., A, B, Rh, K typing). In the antiglobulin gel test, the gels ability to separate RBCs from their suspension medium makes it possible to simplify considerably the process of the test. In an indirect antiglobulin test (IAT), for example (Fig. 2), 50 p L of a 0.8-percent RBC suspension is pipetted onto a gel containing AGS, the serum is added, and the tube is centrifuged after a period of incubation. At the beginning of centrifugation, the RBCs tend to pass through the gel, but the medium in which they are sus-

1989, and accepted August 8, 1989.

109

110

LAPIERRE ET AL.

TRANSFUSION
Vol. 30. No. 2-1990

4 5 +/FIG. 1. Different aspccts of agglutination reactions pcrformcd in a gel: negative reaction (0) and mixed ficld agglutination ( + ) arc clearly diffcrcnt from positivc reactions (1 to 5 ) .

M
0

..

ALUMIN IM SHEET
f

3 mm

SPECIFIC GEL -

i""
i'
mm

pended remains above. This results in a good separation between the RBCs and the medium, without an RBC washing phase. This property of the gel test is somewhat similar to methods described p r e v i o ~ s l y . ~RBCs ~' come in contact with AGS in the upper part of the gel, and the separation of positive and negative reactions occurs. As with conventional methods, the sensitivity of the gel antiglobulin test is improved by the use of enzyme-treated cells. For optimal results, special microtubes should be used. The tubes must be wider at the top to allow an incubation of the reagents above the gel. The part containing the gel must be relatively long and narrow, to ensure prolonged contact of the RBCs with the gel during centrifugation and to limit the quantity of gel used and therefore the cost of the test. Furthermore, it seems that if the bottom of the tube is conical, it gives better results. The proper tubes were not available on the market when the test was first described,' and the first trials were performed in a 400-pL microtube that was 45 mm high and had an internal diameter of 4 mm and a conical bottom with a sharp end (Tube 72702, Starstedt, Romnelsdorf, FRG). These tubes were filled with 200 p L of gel before use. A kit was developed in parallel. This kit (Fig. 3) is composed of a 5 x 7-cm plastic card made up of six microtubes, which are filled with 40 pL of gel and sealed with an aluminum sheet. Depending on the test to be carried out, the six tubes may contain the same gel (antiglob-

FIG.3. An cxamplc of the spccial card dcvelopcd for fhe gcl fcst: in this example, the microtubcs contain anti-C, -c, -E, -e, -Kgcls and a negative control.

SERUM RBC

46

U
1) DISPENSING
2) INCUBATION

3) CENTRIFUGATION

48). POSITIVE REACTION

4b) NEGATIVE REACTION

FIG.2. Indirect gcl antiglobulin tcst. Only clear positive and ncgarivc reactions are shown in 4a and 4b, respectivcly; as shown in Fig. 1, variations in rcactivity may be obtaincd.

ulin gel for six tests, neutral gel for six tests) or a combination of reagent gels (e.g., anti-A, -B, -AB, -D, - C + D + E and neutral gel for ABO-Rh typing and anti-C, -c, -E, -e, -Kand neutral gel for Rh-K typing). The cards may be kept at room temperature (the reagent gels are stable for 1 year). The tubes can be re-covered after distribution of a potentially contagious serum, and the rest of the process may be performed without any contact with this serum, even for the antiglobulin test. The gel tests are usually performed with small quantities of reagents. The suspension of RBCs used contains about 0.4 to 0.5 pL of packed RBCs per tube (e.g., 1 drop of 0.8% RBC suspension), which is less than that used with traditional tube tests. This property allows a better serum:cell ratio without any increase of the volume of serum. For some particular applications (cell separation, for example), a more concentrated cell suspension may be useful. Centrifugation of the microtubes must be strictly controlled. False-positive reactions increase if centrifugation is too slow or too short; false-negative reactions increase if centrifugation is too fast or too long. These parameters are slightly dependent on the shape of the tube and the volume of gel per tube. But, time and speed are not the only important parameters of centrifugation: a fundamental point is that the axis of the tube during centrifugation must be strictly in line with the direction of the centrifugal force. Centrifugation at about 10 minutes at 70 x g is usually optimal, especially for the card described previously. A centrifuge that takes into account the specific parameters of this test has been developed to ensure good results. However, other centrifuges may be used if the centrifugation parameters are strictly respected. Several dextran gels may be used to prepare the gel reagents (e.g., Sephadex GlOO Superfine, Sephadex G200 Superfine, or Sephacryl S200; all, Pharmacia Fine Chemicals, Uppsala, Sweden). The buffer in which the gel is prepared may also vary, but it must be compatible with the RBCs (e.g., saline solutions, low ionic media). When the gel is supplied as a powder, it must swell in the buffer; when supplied as a gel, it must be washed and resuspended in the buffer. The conditions of gel preparation (filtration, sterility) must be very strict to ensure good preservation of the reagent gel. Some elements may be added, to preserve the gel (e.g., sodium azide) or to improve the quality of the RBCs, sedimentation (e.g., antibody-free serum). When specific sera are included in the gel, they must be selected carefully to ensure high sensitivity without nonspecificity. This is particularly true for polyvalent AGS,

TRANSFUSION 1990-Vnl. 30. Nu. 2

GEL TEST FOR ANTIGEN-ANTIBODY REACTIONS

111

the majority of which is not suitable for this test because of a high rate of nonspecific reactions. All the above characteristics (special tubes, strict conditions in the preparation of the gel, precise conditions of centrifugation) have led to the development of a kit to ensure optimal conditions and standardization (Diamed SA, Murten, Switzerland).

Comparative study of a gel antiglobulin test (GLUT) o tests ( L U T and MPT) and h ~ conventional
We tested 400 known antibodies in a retrospective study with two conventional tests, the low ionic antiglobulin test (LIAT) and the manual Polybrene test (MPT), and a gel low ionic antiglobulin test (GLIAT). The sera were obtained mostly from patients, but some came from donors or from the Serum, Cells And Rare Fluids exchange group. Three aliquots of each sera were prepared and numbered, and blind testing was performed. When a serum contained more than one antibody, the detection of each antibody was evaluated with suitable reagent RBCs. LIAT. The LIAT was performed in tubes, with slight variations from the method described by Low and Messeter:6 1 drop of serum and 1 drop of a 4-percent RBC suspension in a low-ionic-saline solution (LISS) were incubated for 10 minutes at 37C. The RBCs were washed three times before the addition of 50 p L of AGS (polyclonal anti-IgG and monoclonal anti-C3d; Diaclon Coombs, Diamed). The reactions were read macroscopically. MPT. The MPT was performed exactly as described by Lalezari. This involves incubating RBCs with test serum, adding Polybrene, decanting supernatant, and resuspending the RBCs for the evaluation of agglutination. GLIAT. The GLIAT was performed according to the process originally described. We mixed 1 g of Sephadex GlOO Superfine (Pharmacia), 18 mL of a LISS solution according to Liiw and Messeterh (CRTS de Lyon, Lyon, France), and 2 mL of a polyvalent AGS (Diaclon Coombs, Diamed). We immediately dispensed 200 p L of this gel in a microtube (Tube 72702, Starstedt) and centrifuged it (1200 x g) to eliminate bubbles. One drop of a 0.8-percent RBC suspension in LISS and 25 p L of serum were added. The tube was incubated for 10 minutes at 37C and centrifuged for 10 minutes at 100 x g (CR 4.22, Jouan, Saint-Herblain, France). The reaction was read macroscopically.

frozen and thawed or fresh. RBCs were classified as D when they did not react with anti-D by a slide test but were constantly positive with LIAT. The gel test was performed with a commercial kit (Diamed) prepared as follows: 5 mL of Sephacryl S200 (Pharmacia) was washed twice in saline. After the second washing (centrifugation 5 min at 1250 x g), the supernatant was discarded and replaced by an isotonic imidazol buffer (imidazol 0.014 M, 0.85% NaC1, pH 7.6) to a final volume of 4.5 mL. We added 0.5 mL of antiserum (e.g., anti-A previously tested for specificity, titer, avidity, and reactivity with ,cells). The special microtube was filled with 40 p L A, and 4 of gel and sealed. A 4-percent suspension of the RBCs to be typed was prepared by mixing 10 p L of packed cells and 200 p L of a bromelin buffer (extracted bromelin 10 g/L, imidazol buffer 0.015 M,0.85% NaCI, pH 7.2, stabilizers; DiaBrom, Diamed). After a 10-minute incubation, 10 p L of the suspension was pipetted onto the specific gel and a negative control gel. The card was then centrifuged in a special centrifuge for 10 minutes at 70 x g (Diacent 2 with ID carrier, Labomed SA, Murten, Switzerland).

Results Comparative study of GLUT, LUT, and MPT

Table 1 shows the results with these three tests. All Rh antibodies detected by LIAT were detected by GLIAT and MPT. Some antibodies were defected only by GLIAT and a few more only by MPT. All the Kell system antibodies detected by MPT were detected by LIAT and GLIAT, both of which proved to be more sensitive than MPT. Several antibodies were detected only by GLIAT and one only by LIAT. In the Duffy, Kidd, and Ss groups, the three techniques gave similar results, but it has to be noted that GLIAT detected all the antibodies, whereas LIAT and MPT missed some. In the group of naturally occurring antibodies, some antibodies were detected only by MPT and some only by GLIAT. These two tests proved to he more sensitive than LIAT, which is not necessarily an advantage, for antibodies that are generally nonsignificant clinically. In the group of rare antibodies, GLIAT appeared to be at least as sensitive as MPT and LIAT.

Evaluation of the sensitivity of GLUT and GET using a calibrated anti-D

Table 2 shows the results obtained by the different tests. The gel tests showed a high sensitivity and a good reproducibility.

Evaluation of the sensitivity of GLIAT and a gel enzyme test

A calibrated anti-D (CNTS, Orsay, France), containing 11.7
pg per mL (58.5 IU) of anti-D was used to evaluate the relative

Efficiency of n gel typing system for weak phenotypes

The gel test gave stronger reactions in all 66 samples of weak phenotypes tested (Table 3).

sensitivity of a GLIAT and a gel enzyme test (GET). Three conventional tests were performed in parallel: LIAT, MPT, and a conventional two-stage papain test. T h e calibrated serum was diluted in saline containing 10-percent inert AB. Each dilution was numbered at random by a technician and tested blind by another technician. The manipulation was repeated 10 times, with different people, at different times. The GET was performed with a 0.8-percent suspension of papain-treated RBCs and a neutral gel (antiglobulin serum was replaced by a 10% dilution of A13 serum in saline). Volumes, temperature, incubation, and centrifugation parameters were the same as in GLIAT.

Discussion
T h e G L I A T is probably the most valuable application of the gel process. It is easier to perform and to read than the traditional antiglobulin test, and it has more reliable results. It may b e used for direct o r indirect tests and may b e applied to grouping, antibody characterization, and compatibility testing. Like most methods, GLIAT may miss some antibodies, but it appears to be one of the most potent tests and perhaps the safest for routine use. In our experience, it sometimes detected antibodies

Efficiency of a gel typing system for weak phenotypes

We tested 66 RBCs with weak antigens (e.g., A,, A,, B,,

D) with the corresponding specific gels. The RBCs were either

112

LAPIERRE ET AL.

TKANSFUSION
Vol. 30, No. 2-1940

Table 1. Comparison of GLlAT*, conventional LlAT+, and MPT? retrospective study of 400 known antibodies Antibody Rh (222) [D(99), C(231, ~(241, E(62), e(7), Ce(l), C'"(6)l Kell (78) [K(67), k(6), K P W I Duffy, Kidd, Ss (30) [FYe(l4), Fyb(2), Jk"(71, Jkb(2), S(3), ~(211 Naturally occurring (53) [Lea(16), Leb(8), M(18), N(1h P,(7), HI(3)l Various (18) [Ve1(3), U(2), Ch"(2), Yt"(l), Lub(l),L W ) , Lan(l), JMH(3), Ge3(1), Jk3(1), Kpb(2)] Total (400) 'Gel low ionic antiglobulin test. tLow ionic antiglobulin test. $Manual Polybrene test. Number (%) of sera yielding specific positive reactions. GLIAT 196 (88Yo) 77 (99%) 30 (100%) 42 (79%) 14 (78%) LIAT 125 (56%) 72 (92%) 26 (87%) 6 (11%) 14 (78%) MPT 201 (91%) 32 (41%) 26 (87%) 33 (62%) 9 (50%)

359 (90%)

243 (61%)

301 (75%)

Table 2. Anti-D

Evaluation of the sensitivity of various agglutination tests using dilutions of a calibrated anti-D LIAT' 4.6 1.9 GLIATt 2.3 1.3 ET* 4 2.9 GET4 MPT 1.4 0.7 1.7 1.0

Lowest detectable concentration (nglmL) SEM

'Low ionic antiglobulin test. ?Gel low ionic antiglobulin test. $Enzyme test. Gel enzyme test. IlManual Polybrene test.

Table 3. Blood grouping with the gel test: 66 samples with weak phenotypes Phenotype A3 A, 8. D" Number tested 9 5 2 50 Agglutination (slideltube) LIAT' NTS NT NT GTt

+/-

+ +

Neg

++

++ + + ++

'Low ionic antiglobulin test. tGel test. *Not tested.

that were potentially significant clinically before the conventional tests did. For instance, a man with strong antic, -Fy", and -s was given emergency transfusion of 4 C - , s - , Fy(a - ) RBC units that were compatible with the usual antiglobulin test. Three of the 4 units were incompatible with GLIAT, and an additional anti-Jkb was identified with this test. There was no posttransfusion rise in hemoglobin concentration, anti-Jkb was easily detectable by the usual techniques 3 days later, and anemia was corrected by C - , s - , Fy(a-), Jk(b-) RBCs. GLIAT is sometimes more sensitive that conventional low ionic antiglobulin tests in detecting some antibodies

not normally considered to be of clinical significance (anti-My -N, -Leb, -Lea...), and this may be considered as a disadvantage. If the incubation is conducted at 37"C, such antibodies are detected by the gel test with a relatively low rate. The gel may also be used for direct antiglobulin tests (DAT). In this case, it should be interesting to have access to a panel of gel reagents with different specificities (polyvalent serum, anti-IgG, anti-C3d). When a specific gel is used for blood grouping, it must be standardized carefully. Several factors may influence the sensitivity of the gel test, but the choice of the reagent and its concentration in the gel are two crucial points. If the sensitivity obtained with the specific gel is too high, weak phenotypes may not be differentiated easily from normal phenotypes: the reactions with normal phenotypes must be strongly positive (grade 5 ) , but it is useful to obtain weak agglutination (grades 1 to 4) with weak phenotypes. This is achieved by a careful selection and dosage of the specific reagent in the gel. However, the sensitivity must be sufficient to distinguish weak reactions from negative ones (grade 0). Generally, problems that occur when grouping with a gel test are the same as those that occur with a conventional test (e.g., false-positive results when dealing with RBCs having a positive DAT, weak agglutination with weak phenotypes, mixed-field agglutination). In every case, weak reactions remain easily distinguishable from mixedfield agglutination: in this case, some cells are retained at the upper part of the gel, whereas others have pelleted at the bottom ( 5 Fig. 1). With the usual quantities of antisera, the limit of detection of a minor population of cells is about 10 percent. Although not as sensitive as special methods for detecting a minor population of cells, the gel test is more sensitive than most of the usual methods (e.g., tubes, microplates). With some modifi-

TRANSFUSION 1990-VoI. 30. No. 2

GEL TEST FOR ANTIGEN-ANTIBODY REACTlONS

113

cations of the process (use of more concentrated RBC suspension), the gel test allows the detection of a smaller percentage of cells (1%)and the separation of the two populations. After separation and washings, other tests may be performed on nonagglutinated cells. The small quantity of blood required for the gel test may be a n advantage when testing babies: less than 10 p L is by far enough to type a sample for A, B, D, C, E, c, e, and K antigens. Neutral gels can be used for reverse ABO typing, screening and identification of antibodies with untreated or enzyme-treated RBCs, and cold agglutinin titrations. Reverse ABO typing may be performed with untreated or enzyme-treated cells. The first method is approximately as sensitive as the usual saline tests; the second method is more sensitive but increases the number of unwanted reactions. The neutral gels with enzyme-treated cells (GET) can also be used to detect irregular antibodies. Incubation may be carried out at any temperature, depending on the need, but a low temperature will give more undesirable reactions. The sensitivity and specificity of such tests are comparable to those of conventional enzyme tests. The sensitivity of the gel test may be due to several factors. A small quantity of RBCs is sufficient for its performance, which allows an improvement of the serum:cell ratio.8 If the gel is prepared in a low ionic medium and the incubation conducted in the same sort of medium, then the whole process is carried out in low ionic condition^.^ The absence of washing could lead to a higher sensitivity, by limiting the elution of the antibodies from RBCs. The reliability of the gel test is due to the ease of reading, the possibility of re-reading, and the fact that the whole process is relatively independent of the skill of the manipulator. As the method makes it possible to separate cells from serum without the risk of aerosol formation, it may be useful for handling serum samples that contain dangerous pathogens. The cost of the gel test is dependent on the laboratorys size and organization and the type of application used. Although the test is not yet available everywhere, some data based on the European situation may be given. The equipment required for a manual use (working station, dispensers, pipette, special centrifuge) is at present available for $3600. In this configuration, the centrifuge allows 864 reactions per hour to be performed (e.g., 72 blood typing tests on 1 2 parameters, for a total of 864 antiglobulin tests). The average retail price of a tube containing antiglobulin gel is $0.27. It replaces the tubes, the AGS for one test, and the saline for washing; reduces the amount of test RBCs and LISS needed; and allows the replacement of a cell washer by the previously described centrifuge, which is cheaper. For an Rh and K profile (C, c, E, e, K, control), the average price is about $2.25 each ($0.38 per reaction). The prices of other

reactions (with neutral or specific gels) vary around these values, depending on the demand and the price of the reagent included in the gel. The labor costs are more difficult to evaluate and vary depending on the activity of the laboratory. In Europe, where the gel test is beginning to be used routinely in several big laboratories for antibody screening and compatibility testing, it is generally admitted that it significantly reduces labor and training costs. For other applications (e.g., blood typing), the gel test may have the same effect in small laboratories, but not in large ones, especially if they use automation. It is hoped that, in the future, the test will be automated. The dispensing of reagents will not be more difficult to automate than with other tests, and i t may be easier. The reading of the reaction will benefit from a special reader. A kinetic reading could probably further improve the quality of the results. The gel test is already, however, a valuable tool for immunohematologists, especially for antiglobulin tests and for other agglutination tests in small laboratories, or for less-experienced technicians.
Acknowledgments
The authors thank Carleen Billaud, Francoisc Juron-Dupraz, Brigitte Pochon, and Paul Debeaux for their valuable assistance.

References
1. Lapierre Y. ProctdC de misc en Cvidencc dagglutinats Ctythrocy-

2.

3.
4.
5.

6.
7.

8.
9.

taires. French patent applicafion 8502010. Bull Off PropriClC Indust 1986;33:2577321. Lalezari P. A new method for detection of rcd blood cell antibodies. Transfusion 1968;8:372-80. Lalezari P, Jiang AF. The manual Polybrene test: a simple and rapid procedure for detection of red cell antibodies. Transfusion 1980;20:206-11. Graham HA, Hawk JB, Chachowski R, Savitz SR. A new approach to prepare cells for the Coombs tcst (abstract). Transfusion 1982; 22:408. Kankura T, Kurashina S, Nakao M. A gcl filtration technique for separation of crythrocytcs from human blood. J Lab Clin Mcd 1914; 83:840-4. Lbw B, Messeter L. Antiglobulin test in low-ionic strength salt solution for rapid antibody screening and cross-matching. Vox Sang 1974;26:53-61. Issitt PD. Applied blood group serology. 3rd ed. Miami: Montgomery Scientific Publications, 1985:49-50. Hughes-Jones NC, Polley MJ, Telford R, Gardner B, Kleinschmidt G. Optimal conditions for detecting blood group antibodies by the antiglobulin test. Vox Sang 1964;9:285-95. Ahn JH, Rosenfield RE, Kochwa S. Low ionic antiglobulin tests. Transfusion 1987;27: 125-33.

Y. Lapierre, MD, MSc, Chef de Service, Centre Regional de Transfusion Sanguine (CRTS) de Lyon. Antenne de Ihbpital Lyon-Sud 69310, Pierre-Btnite, France. [Reprint requests] D. Rigal, MD, PhD, Chcf dc Scrvice, CRTS de Lyon. I. Adam, Gencral Manager, Diamed SA, Murten, Switzcrland. D. Josef, Director of Research and Development, Diamed SA F. Meyer, MD, MSc. Chef dc Servicc, CRTS de Lyon. S. Greber, Chief of the Laboratory, Diamed SA C. Drot, Technicicnne Supcricurc, CRTS de Lyon.