Ouabain and Digoxin Between Two Related Digitalis Glycosides

Monoclonal Antibodies That Distinguish Between Two Related Digitalis Glycosides, Ouabain and Digoxin

This information is current as Behnaz Parhami-Seren, Charles Bell, Michael N. Margolies of September 28, 2021. and Garner T. Haupert, Jr. J Immunol 1999; 163:4360-4366; ; http://www.jimmunol.org/content/163/8/4360 Downloaded from

References This article cites 37 articles, 15 of which you can access for free at: http://www.jimmunol.org/content/163/8/4360.full#ref-list-1

Why The JI? Submit online. http://www.jimmunol.org/

• Rapid Reviews! 30 days* from submission to initial decision

• No Triage! Every submission reviewed by practicing scientists

• Fast Publication! 4 weeks from acceptance to publication

*average by guest on September 28, 2021

Subscription Information about subscribing to The Journal of Immunology is online at: http://jimmunol.org/subscription Permissions Submit copyright permission requests at: http://www.aai.org/About/Publications/JI/copyright.html Email Alerts Receive free email-alerts when new articles cite this article. Sign up at: http://jimmunol.org/alerts

The Journal of Immunology is published twice each month by The American Association of Immunologists, Inc., 1451 Rockville Pike, Suite 650, Rockville, MD 20852 Copyright © 1999 by The American Association of Immunologists All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. Monoclonal Antibodies That Distinguish Between Two Related Digitalis Glycosides, Ouabain and Digoxin1

Behnaz Parhami-Seren,* Charles Bell,2† Michael N. Margolies,* and Garner T. Haupert, Jr.3†

The exogenous digitalis glycosides, ouabain and digoxin, have been widely used in humans to treat congestive heart failure and cardiac arrhythmias. Several reports have also pointed to the existence of endogenous ouabain- and digoxin-like compounds, but their precise roles in mammalian physiology and various disorders of the circulation are not clear. In an attempt to produce specific Abs for the purification and identification of endogenous ouabain-like compounds, somatic cell fusion was used to produce mAbs specific for ouabain. Our attempts to produce ouabain-specific mAbs were unsuccessful when ouabain was coupled to exogenous proteins such as bovine ␥-globulins, BSA, and human serum albumin. However, when ouabain was coupled to an Ab of A/J mice origin and the same strain of mouse was used for immunization with ouabain-Ab conjugate, three Abs (1-10, 5A12,

and 7-1) specific for ouabain were obtained. In assays of fluorescence quenching and saturation equilibrium with tritiated ouabain, Downloaded from Ab 1-10 exhibited 200 nM affinity for ouabain. These three mAbs are distinguished from existing Abs to ouabain and digoxin by their specificity for ouabain and lack of cross-reactivity with digoxin. Specificity studies showed that the loss of cross-reactivity was correlated with the presence of a hydroxyl group at either position 12␤ (digoxin) or 16␤ (gitoxin) of the steroid ring. These Abs can be used to develop assays for detection and characterization of ouabain-like molecules in vivo. The Journal of Immunology, 1999, 163: 4360–4366. http://www.jimmunol.org/ igitalis glycosides have been used widely in the clinical Whether this endogenous Oua is truly endogenous, vs accumu- treatment of congestive heart failure and certain cardiac lated from the environment, or exists in the circulation as a regu- D arrhythmias for hundreds of years. The therapeutic index latory hormone is currently debated. In an effort to answer these of these drugs is narrow, and clinical toxicity is frequent (1). The crucial questions, several groups have developed polyclonal anti- advent of somatic cell fusion techniques (2) permitted the produc- Oua Abs for use in RIA and ELISA-type biological assays. Using tion of mAbs with high sensitivity and specificity to digoxin (Dig)4 different Abs, findings have continued to be contradictory. Some (3) for use in the study of Ab structure-function relationships, de- laboratories have obtained immunoassay data supporting not only velopment of clinical assays, and emergency treatment for life- the existence of OLC in mammalian plasma and tissues (9, 10), but threatening digitalis overdoses (4, 5). that OLC can be secreted by adrenal cells in culture in response to by guest on September 28, 2021 In recent years, a body of evidence has accumulated to suggest receptor stimulation (11) and feeding of steroid hormone precur- the existence in mammals, including man, of compounds with bi- sors (12). To our knowledge, only one group has provided phys- ological properties similar to or identical with those of the plant- icochemical structural analysis of immunoreactive isolates. Liquid derived cardiotonic steroids. Structural analysis of tissue and chromatography mass spectrometry using selected ion recording plasma extracts has indicated that one of these compounds is iden- and positive electrospray mass spectrometry indicated a compound tical in structure to the plant-derived cardenolide, ouabain (Oua) in tissue isolates and adrenal cell culture supernatants, respec- (6, 7) (Fig. 1). This Oua-like compound (OLC) has been impli- tively, with a molecular mass identical with that of authentic plant- cated in the control of renal sodium excretion, blood pressure reg- derived Oua (12). ulation, cardiac muscle performance, and the pathogenesis of hy- Other investigators, using their own polyclonal Abs to Oua, pertension through endogenous regulation (or dysregulation) of have questioned the existence of authentic Oua in plasma and ad- ϩ ϩ Na ,K -ATPase (sodium pump) in pertinent target tissues (8). renal cell culture supernatants and have provided data that structural identity of OLC from these sources with plant Oua is un- likely. The primary basis for these conclusions rests on the *Department of Surgery, Massachusetts General Hospital and Harvard Medical demonstration that HPLC retention times for immunoreactive School, and †Renal Unit, Medical Services, Massachusetts General Hospital, Harvard OLC and authentic plant Oua spiked into the chromatographic Medical School, Charlestown, MA 02129 samples are different (13–15) Received for publication May 14, 1999. Accepted for publication July 28, 1999. One obvious explanation for these discrepant results is that the The costs of publication of this article were defrayed in part by the payment of page assays employed do not recognize the same compounds; that is, charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. that the polyclonal Abs are not specific for plant Oua or a putative Oua isomer of mammalian origin. One approach to enhance the 1 This work was supported by National Institutes of Health Grant RO1HL52282 (to G.T.H) and in part by National Institutes of Health Grants R29AI33175 (to B.P.S.) specificity of and/or provide a standardization for immunoassay and RO1 CA24432 and HL47415 (to M.N.M.). detection of OLC would be the development of anti-Oua mAbs 2 Current address: Franklin Pierce Law Center, Two White Street, Concord, NH with high specificity for Oua. To our knowledge, there is only one 03301. report in the literature of an mAb to Oua, but this Ab showed a 3 Address correspondence and reprint requests to Dr. Garner T. Haupert, Jr., Renal high degree of cross-reactivity with Dig, the cardiac glycoside in Unit, CNY-8, Massachusetts General Hospital, Building 149, 13th Street, Charles- town, MA 02129. E-mail address: [email protected] prevalent clinical use (16) (Fig. 1). 4 Abbreviations in this paper: Oua, ouabain; Dig, digoxin; BGG, bovine ␥-globulin; We set out to raise such Abs by techniques previously used in OLC, ouabain-like compound(s); HSA, human serum albumin. our laboratory for the production of anti-Dig mAbs (3). Initial

FIGURE 1. Schematic representation of Oua and Dig structures, showing the numbering system of the cardenolide steroid rings. http://www.jimmunol.org/ attempts were unsuccessful; all the mAbs recognized the Oua-pro- subsequent fusion experiments, a similar immunization protocol was used, tein conjugate, but not the hapten Oua itself. By using a novel Ag but a different strain of mice (A/J, The Jackson Laboratory) and different presentation technique, we were able to overcome this problem of immunizing Ags (Oua-BGG, Oua-HSA, or Oua coupled to 26-10 Ab) were used. Mice that were immunized with Oua-26-10 Ab conjugate received specificity for the Oua-protein complex. We report here the iden- six additional booster injections of 10 ␮g of Ag in soluble form every 15 tification of mAbs with high specificity for Oua that do not rec- days (hyperimmunized). Before fusion, mouse sera were tested for Ab ognize the clinically used cardiac glycoside, Dig. We propose that titers. Fifty percent binding to Oua-protein conjugates was achieved at further development of these Abs could make possible standard- 30,000- to 45,000-fold serum dilutions. Fusions were conducted using Sp2/0-Ag14(Sp2/0) cell lines (23). After ization in bioassays and allow clarification of ambiguities in the fusion, cells were distributed into 96-well microtiter plates. literature regarding the presence, source, pathogenetic role, and by guest on September 28, 2021 mammalian biosynthetic possibilities of OLC. Immunoassays for selection of Oua-specific mAbs Clones producing Oua-specific mAbs were selected by testing the ability of Materials and Methods culture supernatants from wells showing cell growth to bind to immobi- Cell lines lized Oua-protein conjugates in ELISA assays. Fifty microliters of a solution of Oua-protein conjugates (5 ␮g/ml in PBSA (0.15 M NaCl, 0.1 M The generation, selection, and characterization of cell lines producing the sodium phosphate, and 0.02% sodium azide, pH 7.2)) were immobilized in 26-10 (IgG2a, ␬) and 36-71 (IgG1, ␬) mAbs were previously reported (3, the wells of microtiter plates. The binding of mAbs in the culture super- 17). Ab 26-10, which was obtained from the spleen cells of A/J mice natants was detected using HRP-goat anti-mouse Ab (Sigma) (24). The end immunized with Dig-coupled BSA (Dig-BSA), exhibits an affinity of 9.1 ϫ point of the reaction was determined after addition of 25 ␮l of 2 M phos- Ϫ9 ϭ Ϯ ϫ Ϫ8 10 M for Dig and cross-reacts with Oua (Ka 6.0 0.4 10 M) phoric acid in an ELISA reader at 450 nm. Clones were selected for further (18). Ab 36-71 was also derived from spleen cells of A/J mice and is study if the OD was Ն1.0 for Oua-protein conjugates and Յ0.2 for ϭ 450 specific for the hapten p-azophenylarsonate, with a binding constant Ka uncoupled protein. Clones from the wells that tested positive in direct Ϫ 1–4 ϫ 10 7 M (19, 20). binding assays were transferred to 48-well microtiter plates. Inhibition ELISA was used to determine whether the binding of Abs in Synthesis and characterization of hapten-protein conjugates the culture supernatants to immobilized Oua-protein conjugates was inhibited by free Oua. Thus, the binding of 25 ␮l of culture supernatants to Oua, Dig, other cardiac glycosides (Table II), and steroid hormones (cor- immobilized Oua-coupled protein was tested in the presence of either 25 ␮l tisone, corticosterone, and progesterone) were purchased from Sigma (St. of a solution of 100 ␮M Oua or 25 ␮l of 1% BSA, both in PBS. Clones that Louis, MO). Oua was covalently coupled through its terminal rhamnose exhibited Ն40% inhibition were subcloned and studied further. moiety to a number of proteins as previously described (21). Ags included The isotypes of mAbs were determined using an isotyping ELISA kit Oua-BGG (bovine ␥-globulin; United States Biochemical, Cleveland, OH), (Zymed, San Francisco, CA). Oua-HSA (Miles Laboratories, Elkhart, IN), and Oua-BSA. Oua was also coupled to the affinity-purified mAb 26-10. Oua-BGG contained an aver- Affinity purification of mAbs age of 2.5 Oua residues/molecule of BGG; Oua-BSA, Oua-HSA, and Oua- 26-10 Ab conjugates contained 0.5, 1.0, and 1.5 Oua residues/molecule of Oua-specific Abs were purified from1lofculture supernatant by affinity protein, respectively, as determined by their absorption spectrum in con- chromatography on Oua-BGG-Sepharose. Abs were concentrated using centrated H2SO4 (22). Centriprep (30,000 m.w. cut-off; Amicon, Beverly, MA) and were sub- jected to gel filtration on Ultrogel ACA34 columns (LKB, Bromma, Swe- Immunization and fusion den) to separate the monomer mAbs from aggregated ones. All immunizations were given i.p. For production of Oua-specific mAbs, Affinity determinations two strains of mice and different Oua-protein conjugates were used. In the first attempt, BALB/c mice (The Jackson Laboratory, Bar Harbor, ME) Competition ELISA was used first to determine the relative affinity of each were immunized i.p. with 100 ␮g of Oua-BSA emulsified in IFA. They mAb for Oua and Dig. The 96-well PVC plates were coated with 50 ␮lof were again immunized 3 wk later with 50 ␮g of Oua-BSA in CFA. Ten 5 ␮g/ml Oua-BGG in PBSA. First, we determined the Ab concentration days later mice received 10 ␮g of soluble Oua-BSA. Two weeks later (3 that was not in excess of immobilized Ag. Using the direct binding assay days before fusion) mice were boosted with 10 ␮g of soluble Oua-BSA. In described above, the concentration of Ab at which 50% binding was 4362 OUABAIN-SPECIFIC mAb

achieved was ascertained. Inhibition of binding of Abs to Oua-BGG was determined by adding 25 ␮l of Ab (concentrations as determined above) and 25 ␮l of free Oua (0.001–200 ␮M, 2-fold dilutions). The percent in- Ϫ hibition is the ratio (OD450 in the presence of 1% BSA OD450 in the ϫ presence of Oua)/(OD450 in the presence of 1% BSA) 100. The relative afﬁnity (IC50) is the Oua concentration that inhibits 50% of the binding of Ab to Oua-BGG.

The equilibrium binding constant (Ka) of Oua-specific mAb 1-10 was also determined by fluorescence quenching using a Hitachi F-4500 fluorescence spectrophotometer (Hitachi, San Jose, CA). The excitation and emission wavelengths were 270 and 340 nm, respectively. Eight incremental additions of 20 ␮lof10Ϫ6 M Oua in 2 ml of Ab solution in PBSA (12–20 ␮g) followed by four incremental additions of 20 ␮lof10Ϫ5 M Oua were made. The initial fluorescence reading was diminished by 70–75%. Control titrations were conducted by adding Oua to 2 ml of an mAb solution with unrelated specificity (36-71 mAb). Fluorescence quenching was repeated with 1-10 and 36-71 mAbs using 10Ϫ6 and 10Ϫ5 M Dig in PBSA.

The Ka was calculated using a curve-fitting program (19). FIGURE 2. Inhibition of binding of mAbs to Oua-BGG by free Oua. The affinity for the 1-10 mAb was confirmed using an equilibrium sat- Abs were titrated in direct binding assays to determine the concentration uration method with [3H]Oua or [3H]Dig (DuPont-NEN, Boston, MA) as equivalent to 35–50% binding to Oua-BGG. Binding of Abs to Oua-BGG ␮ described previously (18, 25). Briefly, 22 g of mAbs (1-10 or 36-71) were was determined in the presence or the absence of 0.00035–200 ␮M free added to different concentrations of either tritiated Oua or Dig (0.08–20 2 5 Oua. Ab binding was detected using HRP-goat anti-mouse Ab. The control nM, 4.5 ϫ 10 to 4.5 ϫ 10 cpm, 2-fold dilutions). Following incubation Downloaded from mAb 26-10 was raised against Dig-BSA and cross-reacts with Oua. The at room temperature for 1 h, samples were filtered through glass fiber to separate bound from free hapten, and the filters were washed with 10 ml percent inhibition was calculated as described in Materials and Methods. of cold PBSA. 3H-labeled ligand in the filters was measured by liquid scintillation counting. Affinity data were analyzed using the LIGAND program (26). The specificity of these mAbs was tested in inhibition assays using Dig. As shown in Fig. 3 and Table I, three mAbs showed

Specificity of anti-Oua mAbs minimal (5A12) or absent (7-1 and 1-10) cross-reactivity with Dig, http://www.jimmunol.org/ as their binding to Oua-BGG could not be inhibited with concen- Competition ELISA was used to determine the cross-reactivity of the ␮ mAbs with different digitalis glycosides (Oua, Dig, gitoxin, and digitoxin trations of free Dig as high as 100 M. One mAb (2H8) cross- ϭ ␮ ␮ and their derivatives listed in Tables I and II) and with endogenous steroid reacted with Dig (IC50 3 M). Approximately 0.006 M Dig hormones (cortisone, corticosterone, and progesterone). In these assays was required for 50% inhibition of 26-10 binding. The relative binding of mAbs to Oua-BGG was determined in the presence or the ab- ␮ affinity of 26-10 for Dig in inhibition assays is in agreement with sence of various concentrations (0.00035–200 M) of free digitalis gly- 3 ϫ Ϫ9 cosides and steroid hormones as described above. the previously reported Ka values using [ H]Dig (9.1 10 M) (25). The Ka of 26-10 for Oua was previously reported to be 40- fold less than that for Dig using [3H]Oua (18); in the inhibition Results assays reported here this difference is 62-fold (Table I). by guest on September 28, 2021 In an attempt to produce Oua-specific mAbs, Oua was coupled to Because the 2H8 mAb cross-reacts with Dig, we excluded this different protein carriers. From the fusion of spleen cells of A/J and Ab from further study. mAbs 5A12, 7-1, and 1-10 were affinity BALB/c mice that were immunized with Oua-BSA, Oua-HSA, or purified, concentrated, and applied to a gel filtration column Oua-BGG, Ͼ1000 clones exhibited significant specific binding to (ACA34). Fig. 4 demonstrates the aggregation pattern of these Abs Oua-protein conjugates, but the binding of very few clones could in neutral buffer (PBSA). All mAbs formed aggregates, but mAb be inhibited by free Oua. These clones had low relative affinity for 1-10 had the lowest amount of aggregates and the highest amount ϭ Ϫ4 Oua (IC50 10 M). It appeared that Oua was being recognized of monomer Ab. The cell line producing mAb 1-10 secreted high by the Ab-producing cells in vivo mainly in the context of the levels of Ab, ϳ15 mg of purified Ab from1lofculture superna- epitopes on the protein carrier. To overcome the problems associ- tant. In contrast, hybridoma clones 5A12 and 7-1 were low pro- ated with protein carrier immunogenicity, we coupled Oua to the ducers. The level of production of mAbs and their aggregation anti-Dig 26-10 mAb that was derived from A/J mice (3) and then patterns are important for the practicality of large scale production, hyperimmunized A/J mice with the Oua-26-10 Ab conjugate. purification, and stability. The inhibition assays were repeated us- From the fused splenocytes of two immunized mice in two differ- ing all three affinity-purified Abs. Similar relative affinity values ϭ ␮ ent fusion experiments a total of 600 clones were screened for their (IC50 7–25 M) were obtained for all affinity-purified Abs (data binding to Oua-BGG and BGG. Sixty clones were found to pro- not shown). duce Abs that bound to Oua-BGG but not to BGG (data not Fluorescence quenching analysis of anti-Oua Abs indicated that shown). the fluorescence emission only of 1-10 mAb (but not that of 5A12,

Inhibition assays were performed to identify clones that produce 7-1, or 26-10) can be quenched upon addition of free Oua. The Ka mAbs, the binding of which to Oua-BGG could be inhibited with of 1-10 mAb for Oua was 3 Ϯ 1 ϫ 10Ϫ7 M using fluorescence free Oua at micromolar concentrations. Four clones (5A12 and quenching (Fig. 5). The fluorescence emission of 1-10 mAb was 2H8 from the first fusion, and 7-1 and 1-10 from the second fusion) comparable to that of control Ab upon addition of 10Ϫ6 and 10Ϫ5 were selected for further studies. The mAb 2H8 was IgG2a ␬, and M free Dig, confirming that this mAb does not cross-react with the other mAbs were IgG1 ␬. Fig. 2 shows the inhibition pattern of free Dig in solution.

each mAb with Oua. The high affinity Dig-specific mAb 26-10, Using a saturation equilibrium assay with tritiated Oua, the Ka ϭ ϫ Ϫ8 ϫ Ϫ7 which cross-reacts with Oua (Ka for Oua 6 10 M) (18), was of 1-10 mAb was measured (2.4 10 M) and was similar to that used as control. The binding of all four mAbs to Oua-BGG could obtained by fluorescence quenching (data not shown). The satura- be inhibited with free Oua in a concentration-dependent manner. tion equilibrium assay was repeated using [3H]Dig. Ab 1-10 did Approximately 7–25 ␮M Oua was required to achieve 50% inhi- not capture sufficient tritiated ligand for measurement. bition for Oua-specific mAbs. For 26-10 mAb, 0.37 ␮M Oua was The binding specificity of Oua-specific mAbs to closely related required for 50% inhibition (Fig. 2 and Table I). analogues of Oua and Dig was determined by competition ELISA. The Journal of Immunology 4363

Table I. Fine speciﬁcity of Oua-speciﬁc mAbsa

␮ IC50 ( M)

Analogueb 26-10c 1-10c 7-1c 5A12c

Oua 0.37 25 20 10 Ouabagenin 0.40 30 30 16 Strophanthidin 0.20 10 10 4 Acetylstrophanthidin 0.20 6 6 2 Acovenoside 0.20 8 8 5 Convallatoxin 0.10 3 3 2 Helveticoside 0.20 150 150 150 Digitoxin 0.02 3 3 3 Digitoxigenin 0.04 4 4 4 Dig 0.006 NId NI NI Digoxigenin-3,12-diacetate 6 NI NI NI Gitoxin 1.5 NI NI 100 Gitoxigenin-3,16-diacetate NI NI NI NI 16-Acetylgitoxin 0.5 NI NI NI Oleandrin Ͼ100 NI NI NI Oleandrigenin Ͼ100 NI NI NI Cortisone 100 NI NI NI Downloaded from Corticosterone NI NI NI NI Progesterone 2 NI NI NI

a Speciﬁcity of mAbs was determined in ELISA competition assays. Binding of mAbs to Oua-BGG was determined in the presence or absence of increasing concentrations (0.00035–200 ␮M, 2-fold dilutions) of inhibitors using HRP-goat anti-mouse Abs. The ␮M concentration required for 50% inhibition of the binding of Abs to Oua-BGG was calculated as described in Materials and Methods. b

Analogues were prepared at 2–5 mM concentrations in 70% ethanol and diluted into PBS. http://www.jimmunol.org/ c mAbs were diluted in 1% BSA/PBS and titrated using direct binding ELISA to determine the Ab concentration equivalent to 35–50% binding to immobilized Oua-BGG. d NI, no inhibition was observed at highest inhibitor concentration (100 ␮M).

Table I shows the relative affinity (IC50; i.e., micromolar concen- ble I). Surprisingly, all three mAbs bound to digitoxin at micro- tration of free inhibitor required for 50% inhibition) of mAbs for molar concentrations (2–4 ␮M). The cross-reactivity of 1-10 mAb each inhibitor, compared with that of 26-10 mAb. All three mAbs with digitoxin was confirmed using fluorescence quenching. As exhibited similar, but not identical, fine specificities for Oua ana- shown in Fig. 5, digitoxin, but not Dig or gitoxin, inhibited the

logues. The absence of the rhamnose sugar of Oua at position 3 of fluorescence emission of 1-10 mAb in a pattern similar to that of by guest on September 28, 2021 Ϯ ϫ Ϫ7 the steroid ring (Fig. 1) did not substantially affect binding, as Oua. A Ka of 4.9 0.8 10 M was obtained for digitoxin. ouabagenin binding was indistinguishable from that of Oua (Ta- bles I and II). However, the relative affinities of Abs for helveti- Discussion coside (strophanthidin digitoxoside) were reduced 6- to 15-fold Oua is a cardiac glycoside (m.w., 584.7) found in certain plant compared with their affinities for Oua, indicating that the nature of species, such as the seeds of Strophanthus gratus (27). Oua and the attached sugar affects binding for Oua analogues lacking the OLC have also been found in humans and animals (7, 28–31). 1␤- and 11␣-OH substitutions (Tables I and II). Neither Dig nor Although its function in plants is not known, in mammals Oua and gitoxin inhibited the binding of mAbs to Oua-BGG, although Ab ϭ ␮ OLC may play a role in the regulation of sodium balance, arterial 5A12 exhibited cross-reactivity with gitoxin (IC50 100 M; Table I). None of the three mAbs reacted with the endogenous steroid hormones cortisone, corticosterone, and progesterone (Ta-

FIGURE 4. Elution patterns of Oua-speciﬁc mAbs from an ACA34 gel ﬁltration column. One liter of Ab-containing culture supernatants was passed through an Oua-BGG-Sepharose column. Abs were eluted with 0.2 FIGURE 3. Inhibition of binding of mAbs to Oua-BGG by free digoxin. M ammonia into tubes containing 1.5 M Tris, pH 4.5, and concentrated Binding of Abs to Oua-BGG was determined in the presence or the absence using Centriprep. Concentrated Abs were loaded onto ACA34 columns, of 0.0001–100 ␮M free Dig as described in Fig. 2. The results for Ab 7-1 which were equilibrated with PBSA. One-milliliter fractions were are identical with those for 1-10 mAb (symbols obscured). collected. 4364 OUABAIN-SPECIFIC mAb

the Oua specificity relative to that of Dig of 26-10 Ab would require extensive alterations of the binding site. Thus, we used the traditional somatic cell fusion method to produce Oua-specific mAbs. Fusion of the spleen cells of mice immunized with Oua coupled to BSA, HSA, or BGG with plasmacytomas yielded a very large number of clones secreting mAbs specific for the Oua-protein carrier. In every fusion three kinds of specificities could be detected. The first group (32%) secreted Abs that bound to Oua-protein conjugates; they did not cross-react with either Dig-protein conjugates or protein carriers alone (data not shown). The specificity of the second group of mAbs, which constituted 54% of the clones, was directed against Oua-protein conjugates, which cross-reacted with Dig-proteins but not with protein carrier. The third group of mAbs FIGURE 5. Plots of quenching of fluorescence of 1-10 anti-Oua mAb Ϫ Ϫ (14%) bound only the protein carrier. We were surprised by the and control (36-71) mAb vs free hapten. Hapten (10 6 and 10 5 M) was added to 12–20 ␮g of Ab (in 2 ml of PBSA) as described in Materials and fact that the binding of mAbs to Oua-protein conjugates could not Methods. Quench data were transformed using a computer-assisted curve- be inhibited by micromolar concentrations of free Oua. Because

fitting program (19) to determine the intrinsic affinity (Ka) of 1-10 mAb for we were searching for high affinity Abs, all the inhibition screen- Oua and digitoxin. ings were performed in the presence of 100 ␮M free Oua. This Downloaded from indicated that either Oua is not immunogenic in vivo or the immunogenicity of the protein carriers is greater than that of Oua, pressure, and vascular smooth muscle tone under normal circum- thus shifting the specificity of the Abs toward the protein. stances and have a pathophysiologic role in common clinical dis- To avoid the problems associated with protein carrier immuno- orders, such as essential hypertension, pregnancy-induced hyper- genicity, Oua was coupled to 26-10 Ab. Because the 26-10 Ab was tension, cardiac failure, and chronic renal failure (8, 32). Two derived from A/J mice, the same mouse strain was used for im- http://www.jimmunol.org/ major issues that need to be clarified are 1) the source(s) (including munization with the Oua-26-10 conjugate. Among 60 clones that in vivo vs ex vivo) of Oua or OLC molecules found in humans and secreted mAbs exhibiting specific binding to Oua-BGG, only the animals, and 2) the actual function of Oua or OLC in vivo as a binding of four Abs was inhibited with free Oua. These results can regulator of cardiovascular physiology. The availability of specific be explained in two different ways. First, because OLC exists in molecular probes and reliable methods of detecting and measuring vivo, the immune system may be tolerant to Oua, and thus Oua can endogenous or exogenous Oua is the prerequisite to successfully be recognized only in the context of exogenous proteins. This ex- investigating these issues. plains why the specificity of Abs secreted by clones isolated from Three anti-Oua mAbs were produced by somatic cell fusion.

mice immunized with Oua coupled to BSA, HSA, or BGG was by guest on September 28, 2021 Each Ab was analyzed for its affinity and fine specificity for Oua directed against Oua-protein carriers and not Oua alone. An alter- and related cardiac glycosides. Using solid phase competition as- native explanation is that in Oua the steroid ring is attached says, an IC range of 7–25 ␮M for Oua was obtained for these 50 through a single sugar (rhamnose; Fig. 1), which may allow the mAbs (Fig. 2 and Table I). The affinity (Ka) of one mAb (1-10) was measured by two other methods (fluorescence quenching and sat- attached proteins to sterically hinder the cardenolide moiety of uration equilibrium) and was found to range from 0.24–0.3 ϫ Oua. Anti-Dig mAbs can be elicited more easily, because in Dig 10Ϫ8 M (240–300 nM; Fig. 5). These affinities are sufficiently high the steroid ring is attached via the tridigitoxose (Fig. 1); thus, the to allow the Ab to be used in different methods of Oua detection. sugars may act as a spacer between the steroid ring of Dig and the Although two mAbs with high affinities (2.0 ϫ 10Ϫ7 and 1.2 ϫ protein. Also, there are significant structural differences between 10Ϫ9 M) for Oua were previously reported (16), both cross-reacted the steroid ring substitutions of Oua and those of Dig (Fig. 1 and ␤ ␤ ␣ with Dig, the form of cardiac glycoside widely prescribed for the Table II). Oua has four OH groups at steroid positions 1 ,5 ,11 , treatment of heart failure and certain arrhythmias. Such cross-re- and 19, while Dig does not share any of these OH groups; Dig has ␤ activity would probably be problematic, particularly in human an OH group at steroid position 12 . Such differences could be studies. The mAbs here reported are distinguished from the earlier sufficient for a molecule to be recognized as self or non-self by the ones in their specificity for Oua and their lack of cross-reactivity cells of the immune system. with Dig (Fig. 3 and Table I). Cross-reactivity of anti-Oua mAbs with digitoxin was an unex- In some cases, a combination of methods of Ab engineering pected finding. Comparison of digitoxin with Oua, Dig, and gi- using phage display and molecular modeling have been used to toxin reveals that Oua and digitoxin both lack OH groups at po- change the specificity of Abs that bind to closely related analogues sition 12␤ or 16␤ of the steroid ring, while Dig and gitoxin contain of an Ag. For example, the affinity and specificity of an anti- OH groups at 12␤ and 16␤, respectively (Table II). Oua and digi- cortisol mAb that cross-reacted with prednisolone and dexameth- toxin differ with respect to their sugars at position 3␤ (rhamnose vs asone was improved by 8- and 5-fold, respectively, using random digitoxose, respectively), and digitoxin also lacks the OH group at mutagenesis of its V region genes (33). An anti-estradiol Ab frag- positions 5␤,11␣, and 19 (Table II). ment that cross-reacted with testosterone was mutagenized to im- The reason for heteroclicity of anti-Oua mAbs is not known. We prove its specificity (34). Similar methods have been used for anti- cannot rule out the possibility that the chemical identity of Oua is testosterone mAb cross-reactive with dehydroepiandrosterone altered in the Oua-protein conjugate, but assays of fluorescence sulfate (35) and anti-hydroxyprogestrone mAb cross-reactive with quenching and saturation equilibrium demonstrated that anti-Oua cortisol (36) to improve affinity and specificity. Our own efforts mAbs can bind free Oua in native form in solution, thus indicating (data not shown) in engineering Oua-specific Ab fragments using that the ouabain structure has to some extent been preserved. Al- the existing Dig-binding mAbs 26-10 and 40-50 (37) have not though only speculation, the observation that immunization of been successful to date. Maintaining high affinity and increasing mice with Oua-26-10 complex resulted in Abs with higher affinity The Journal of Immunology 4365

Table II. Structural characteristics of Oua and Dig analoguesa

Substitutions at Steroid Positions

Analogue 1␤ 3␤ 5␤ 11␣ 12␤ 16␤ 19

Oua OOH L-rhamnose OOH OOH OOH Ouabagenin OOH OOH OOH OOH OOH Strophanthidin OOH OOH AO O O A Acetylstrophanthidin OCOCH3 OH O O Acovenoside A OH 6-deoxy-3-O-methyl-L-talose CH3 Convallatoxin L-rhamnose OOH AO Helveticoside Digitoxose OOH AO Digitoxin Tridigitoxose Digitoxigenin OOH Dig Tridigitoxose OOH O O O Digoxigenin-3,12-diacetate OH OCOCH3 OCOCH3 Gitoxin Tridigitoxose OOH O O Gitoxigenin-3,16-diacetate OCOCH3 OCOCH3 O 16-Acetylgitoxin Tridigitoxose OCOCH3 O Oleandrin Oleandrose OCOCH3 O O Oleandrigenin OH OCOCH3 Downloaded from a Cardenolide numbering scheme is shown in Fig. 1. for digitoxin, which was not the immunizing Ag, could be ex- It is surprising that a panel of mAbs obtained from three inde- plained if Oua were modiﬁed to a digitoxin-like compound in vivo pendent fusions all exhibited the same unique speciﬁcity. This may after the complex was processed for presentation to T and B cells. indicate that the in vivo immune response to Oua is restricted,

However, to our knowledge there is no experimental evidence for explaining the low frequency of Oua-specific clones in fusion http://www.jimmunol.org/ targeted modification of self-Ags by the immune system. experiments. mAbs elicited against Dig (3, 25, 37, 38) exhibit varying spec- Using a novel Ag presentation technique, we have been able to ificity patterns for related cardiac glycosides. Such mAbs can bind obtain mAbs to the cardiac glycoside, Oua, whereas immunization Dig and digitoxin equally well or distinguish these two analogues with more traditional hapten-protein complexes produced mAbs to by up to a 1000-fold difference. The three anti-Oua mAbs reported the complex, but not to Oua itself. The particular advantage of here are unique in binding with high affinity to digitoxin, but not these mAbs is that they do not cross react with either endogenous to Dig. In addition, they do not cross-react with gitoxin as do adrenal steroids or Dig, the primary cardiac glycoside used in clin- mAbs elicited against Dig (3, 38). This indicates that in anti-Oua ical practice. Although they are heteroclitic and do react with some Abs, binding site complementarity around the 12␤ OH is probably of the cardenolides tested, none of these latter has been reported as by guest on September 28, 2021 very tight. an isolate of mammalian origin. The mAbs herein described can The chemical nature and the structure of endogenous digitalis- thus provide more specific molecular probes to assess the putative like factors have remained elusive. Some investigators have iden- role of endogenous Oua in mammalian physiology and in the tified an OLC in human plasma (7), while others (8) have isolated pathophysiology of the prevalent human cardiovascular diseases, a compound from human urine that was indistinguishable from hypertension and congestive heart failure. It is important that anti- Dig based on physico-chemical analysis and immunoreactivity Oua Abs be validated by demonstrating their utility as reagents to with anti-Dig IgG. These Abs also neutralized the potency of the detect OLC. Dig-like compound. In addition, a Dig-like immunoreactive factor was isolated from mammalian adrenal cortex that exhibited similar Acknowledgments chromatographic and spectral properties as Dig (39). Thus, it is possible that both endogenous OLC and Dig-like compounds exist We thank Lihua Zhang and Rou-Fun Kwong for technical assistance. in vivo in mammals; if this is so, only specific probes would distinguish between them. The purpose of having a panel of anti-Oua References mAbs that do not cross-react with Dig is to assure that the struc- 1. Smith, T. W. 1975. Digitalis toxicity: epidemiology and clinical use of serum tural nature of the purified OLC is that of Oua and not Dig. In concentration measurements. Am. J. Med. 58:470. addition, we focused on the production of anti-Oua mAbs to aid in 2. Kohler, G., and C. Milstein. 1975. Continuous cultures of fused cells secreting determining the molecular identity of the OLC we previously iso- antibody of predefined specificity. Nature 256:495. 3. Mudgett-Hunter, M., N. M. Margolies, A. Ju, and E. Haber. 1982. High-affinity lated from hypothalamus (29). monoclonal antibodies to the cardiac glycoside, digoxin. J. Immunol. 129:1165. From the clinical point of view, OLC has been implicated in the 4. Smith, T. W., V. P. Butler, E. Haber, H. Fozzard, F. I. Marcus, W. F. Bremner, I. C. Shulman, and A. Phillips. 1982. Treatment of life-threatening digitalis in- pathophysiology of human essential hypertension and congestive toxication with digoxin-specific Fab antibody fragments. N. Engl. J. Med. 307: heart failure (8, 32). Patients with these disorders, who will be 1357. subjects of clinical studies to verify a role for OLC, are often 5. Smith, T. W., E. Haber, L. Yeatman, and V. P. Butler. 1976. Reversal of ad- vanced digoxin intoxication with Fab fragments of digoxin specific antibodies. treated with Dig. Thus, the availability of our mAbs may allow N. Engl. J. Med. 294:797. study of these patients to verify a role for OLC even if they are 6. Schneider, R., V. Wray, M. Nimtz, W. D. Lehmann, U. Kirch, R. Antolovic, and treated with Dig. W. Schoner. 1998. Bovine adrenals contain, in addition to ouabain, a second inhibitor of the sodium pump. J. Biol. Chem. 273:784. Recently, we produced an additional anti-Oua mAb (8E4) with 7. Hamlyn, J. M., M. P. Blaustein, S. Bova, D. W. DuCharme, D. W. Harris, Ϫ8 an affinity of 1.8 ϫ 10 M for Oua (unpublished observations). F. Mandel, W. R. Mathews, and J. H. Ludens. 1991. Identification and charac- The pattern of specificity of 8E4 is similar to that of the mAbs terization of a ouabain-like compound from human plasma. Proc. Natl. Acad. Sci. USA 88:6259. reported here; i.e., 8E4 binds to Oua and digitoxin, but not to Dig 8. Goto, A., K. Yamada, N. Yagi, M. Yoshioka, and T. Sugimoto. 1992. Physiology and gitoxin, in contrast to the 26-10 Ab. and pharmacology of endogenous digitalis-like factors. Pharmcol. Rev. 44:377. 4366 OUABAIN-SPECIFIC mAb

9. Harris, D. W., M. A. Clark, J. F. Fisher, J. M. Hamlyn, K. P. Kolbasa, 25. Schildbach, F. J., I. R. Near, E. R. Bruccoleri, E. Haber, D. P. Jeffrey, S.-C. Ng, J. H. Ludens, and D. W. Ducharme. 1991. Development of an immunoassay for J. Novotny, S. Sheriff, and N. M. Margolies. 1993. Heavy chain position 50 is a endogenous digitalislike factor. Hypertension 17:936. determinant of affinity and specificity for the anti-digoxin antibody 26–10. 10. Ferrandi, M., P. Manunta, S. Balzan, J. M. Hamlyn, G. Bianchi, and P. Ferrari. J. Biol. Chem. 268:21739. 1997. Ouabain-like factor quantification in mammalian tissues and plasma. Hy- 26. Munson, P. J. 1983. A computerized analysis of ligand binding data. Methods pertension 30:886. Enzymol. 92:543. 11. Laredo, J., J. R. Shah, Z.-R. Lu, B. P. Hamilton, and J. M. Hamlyn. 1997. An- 27. Jacobs, W. A., and N. M. Bigelow. 1932. Ouabain or g-strophanthin. J. Biol. giotensin II stimulates secretion of endogenous ouabain from bovine adrenocor- Chem. 96:647. tical cells via angiotensin type 2 receptors. Hypertension 29:401. 28. Ludens, J. H., M. A. Clark, K. P. Kolbasa, and J. M. Hamlyn. 1993. Digitalis-like 12. Perrin, A., B. Brasmes, E. M. Chambaz, and G. Defaye. 1997. Bovine adreno- factor and ouabain-like compound in plasma of volume-expanded dogs. J. Car- cortical cells in culture synthesize an ouabain-like compound. Mol. Cell. Endo- diovas. Pharmacol. 22:S38. crinol. 126:7. 29. Tymiak, A. A., A. J. Norman, M. Bolgar, C. G. DiDonato, H. Lee, L. W. Parker, 13. Lewis, K. L., G. T. Yandle, G. J. Lewis, A. M. Richards, B. G. Pidgeon, L.-C. Lo, N. Berova, K. Nakanishi, E. Haber, et al. 1993. Physicochemical char- J. R. Kaaja, and M. G. Nicholls. 1994. Ouabain is not detectable in human acterization of a ouabain isomer isolated from bovine hypothalamus. Proc. Natl. plasma. Hypertension 24:549. Acad. Sci. USA 90:8189. 14. Gomez-Sanchez, E. P., M. F. Foecking, D. Sellers, M. S. Blankenship, and 30. Haupert, G. T., Jr., and J. S. Sancho. 1979. Sodium transport inhibitor from C. E. Gomez-Sanchez. 1994. Is the circulating ouabain-like compound ouabain? bovine hypothalamus. Proc. Natl. Acad. Sci. USA 76:4658. Am. J. Hyperten. 7:647. 31. Haupert, G. T., Jr., C. Carilli, and L. C. Cantley. 1984. Hypothalamic sodium- 15. Doris, P. A., A. Hayward-Lester, D. Bourne, and D. M. Stocco. 1996. Ouabain transport inhibitor is a high affinity reversible inhibitor of Naϩ,Kϩ-ATPase. production by cultured adrenal cells. Endocrinology 137:533. Am. J. Physiol. 247:F919. 16. Terano, Y., A. Tomii, and F. Masugi. 1991. Production and characterization of 32. Blaustein, M. P. 1996. Endogenous ouabain: Role in the pathogenesis of hyper- antibodies to ouabain. Japan. J. Med. Sci. Biol. 44:123. tension. Kidney Int. 49:1748. 17. Marshak-Rothstein, A., M. Siekevitz, M. N. Margolies, M. Mudgett-Hunter, and M. L. Gefter. 1980. Hybridoma proteins expressing the predominant idiotype of 33. Chames, P., S. Coulon, and D. Baty. 1998. Improving the affinity and fine spec- the anti-p-azophenyarsonate response of A/J mice. Proc. Natl. Acad. Sci. USA ificity of an anti-cortisol antibody by parsimonious mutagenesis and phage dis- 77:11120. play. J. Immunol. 161:5421. Downloaded from 18. Schildbach, F. J., J. D. Panka, R. D. Parks, C. G. Jager, J. Novotny, 34. Lamminmaki, U., B. O. Villoutreix, P. Jauria, P. Saviranta, M. Vihinen, A. L. Herzenberg, M. Mudgett-Hunter, E. R. Bruccoleri, E. Haber, and L. Nilsson, O. Teleman, and T. Lovgren. 1997. Structural analysis of an anti- N. M. Margolies. 1991. Altered hapten recognition by two anti-digoxin hybrid- estradiol antibody. Mol. Immunol. 34:1215. oma variants due to variable region point mutations. J. Biol. Chem. 266:4640. 35. Hemminki, A., S. Niemi, A.-M. Hoffren, L. Hakalahti, H. Soderlund, and 19. Sharon, J. 1990. Structural correlates of high antibody affinity: Three engineered K. Takkinen. 1998. Specificity improvement of a recombinant anti-testosterone amino acid substitutions can increase the affinity of an anti-p-azophenylarsonate Fab fragment by CDRIII mutagenesis and phage display selection. Prot. Engin. antibody 200-fold. Proc. Natl. Acad. Sci. USA 87:4814. 11:311. 36. Iba, Y., N. Hayashi, J.-I. Sawada, K. Titani, and Y. Kurosawa. 1998. Changes in

20. Parhami-Seren, B., P. H. Kussie, R. K. Strong, and M. N. Margolies. 1993. http://www.jimmunol.org/ Conservation of binding site geometry among p-azophenylarsonate-specific Abs. the specificity of antibodies against steroid antigens by introduction of mutations J. Immunol. 150:1829. into complementarity-determining regions of the VH domain. Prot. Engin. 11: 21. Smith, T. W., V. P. Butler, Jr., and E. Haber. 1970. Characterization of antibodies 361. of high affinity and specificity for the digitalis glycoside digoxin. Biochemistry 37. Jeffrey, P. D., J. F. Schildbach, C.-Y. Y. Chang, P. H. Kussie, M. N. Margolies, 9:331. and S. Sheriff. 1995. Structure and specificity of the anti-digoxin antibody 40–50. 22. Brown, B. T., and S. E. Wright. 1960. Absorption spectra of cardiac glycosides J. Mol. Biol. 248:344. and aglycones in sulfuric acid. J. Am. Pharm. Assoc. 49:777. 38. Mudgett-Hunter, M., W. Anderson, E. Haber, and N. M. Margolies. 1985. Bind- 23. Shulman, M., C. D. Wilde, and G. Kohler. 1978. A better cell line for making ing and structural diversity among high affinity monoclonal anti-digoxin antibod- hybridomas secreting specific antibodies. Nature 276:269. ies. Mol. Immunol. 22:477. 24. Parhami-Seren, B., and M. N. Margolies. 1996. Contribution of heavy chain 39. Shaikh, I. M., B. W. C. Lau, B. A. Siegfried, and R. Valdes. 1991. Isolation of junctional amino acid diversity to antibody affinity among p-azophenylarsonate digoxin-like immunoreactive factors from mammalian adrenal cortex. J. Biol.

speciﬁc antibodies. J. Immunol. 157:2066. Chem. 266:13672. by guest on September 28, 2021