Protamine sulfate is used routinely to reverse the anticoagulant action of heparin after cardiac and vascular surgery. It is a low molecular weight (4,500 daltons) protein rich in basic amino acids such as arginine. 1 By virtue of its cationic nature, protamine interacts with sulfate groups on, and neutralizes the anticoagulant activities of, heparin. Unfortunately, the use of protamine can cause severe and at times fatal cardiovascular responses. 2–4 Indeed, the administration of heparin and protamine has been suggested as the major cause of morbidity and mortality for patients undergoing cardiopulmonary bypass. 5 In general, protamine toxicity is mediated by means of two pathways: (1) a nonimmunologic pathway; and (2) an immunoglobulin mediated pathway. Anaphylactoid reactions produced by means of the first mechanism, which are manifested by complement activation, thromboxane generation, and histamine release, are more common and less dangerous. They can normally be aborted with slow administration of protamine. Anaphylactic type of responses produced by means of the second pathway, however, are unpredictable, not preventable, and always life-threatening. More than 100 deaths have been attributed to this type of protamine toxicity. 3,4
The mechanism of heparin neutralization by protamine has been thoroughly investigated. It is now clear that protamine competes with antithrombin III (ATIII) for binding to heparin. 6 Due to a stronger affinity to heparin, protamine dissociates ATIII from the heparin-ATIII complex, thereby reversing the anticoagulant function of heparin. It is also known that protamine binds heparin by means of an electrostatic interaction, and that heparin binds ATIII by means of a small pentasaccharide sequence. 7 Based on these findings, it is concluded that effective binding of protamine to the pentasaccharide sequence in heparin, and displacing the complexed ATIII by means of an ionic interaction, may not require the whole protamine molecule, but rather a small fragment encompassing an intact arginine-rich sequence, for favorable electrostatic interaction.
Like any foreign protein, protamine possesses immunogenic potential. Indeed, the life threatening protamine induced toxicity by means of the immunologic mediated pathway is attributed to the immunogenicity (defined as the ability of a substance to induce an antibody) and antigenicity (defined as the ability to be recognized by an antibody) of protamine. It is well recognized that small peptides with a molecular weight in the range of 1,500 daltons or below are usually either weak or completely devoid of immunogenicity. 8 In addition, studies designed to understand the complexity of protein antigens 9 indicated that a chain-shortened peptide fragment, derived either by enzymatic or chemical digestion of the parent protein, is usually accompanied by greatly diminished antigenicity. On the basis of these results, it is concluded that small peptide fragments with molecular weights of < 1,500 daltons derived from a protein by chemical or enzymatic digestion may be devoid of, or at least possess markedly reduced, immunogenicity and antigenicity.
The convergence of these two conclusions leads to the following hypothesis: If a chain-shortened low molecular weight protamine (LMWP) containing the heparin neutralizing domain could be derived from native protamine, it could be a potent and yet nontoxic heparin antagonist. In this study, we present results to validate this hypothesis. LMWP fragments containing an intact arginine sequence and an average molecular weight of approximately 1,100 daltons were prepared by enzymatic digestion of protamine with thermolysin. In vitro studies that used both the anti-Xa chromogenic and aPTT clotting time assays were conducted to examine if such LMWP fragments could actually neutralize the anticoagulant functions of heparin and low molecular weight heparin (LMWH). In vivo studies were also conducted in mice to determine whether these LMWP fragments would indeed possess a significantly reduced antigenicity and immunogenicity.
Materials and Methods
Protamine sulfate (Clupeine from herring), 2,4,6-trinitrobenzene sulfonic acid (TNBS), thermolysin (EC 188.8.131.52), Freund’s adjuvant, and goat anti-mouse IgG alkaline phosphatase were purchased from Sigma Chemical Co. (St. Louis, MO). Porcine intestine heparin (169 IU/mg; average molecular weight of 13,000 daltons), antithrombin III (ATIII), factor Xa, and chromogenic substrate S-2238 were obtained from Pharmacia Hepar Inc. (Franklin, OH). Actin cephaloplastin was obtained from Dade (Miami, FL). The CY216, a low molecular weight heparin (LMWH) fragment with an average molecular mass of 2,500 daltons, was a gift from Sanofi Recherche (Gentilly Cedex, France). The Coomassie Plus reagent for Bradford protein assay was purchased from Bio-Rad Laboratories (Richmond, CA). Fresh frozen human plasma in citrate was obtained from the American Red Cross in Detroit, Michigan.
Preparation of Low Molecular Weight Protamine Fragments
Protamine does not adsorb light in the ultraviolet region because of the lack of aromatic amino acid residues in its composition. Thus, to monitor liquid chromatographic elution profiles, protamine was first labeled with 2,4,6-trinitrobenzene sulfonic acid (TNBS) to yield absorbance at 340 nm according to the method of Ando et al. 1 A half gram of protamine was dissolved in 1 ml of 0.17 M borate buffer at pH 8.1, and then mixed with 3 ml of TNBS solution (0.2 g/ml). After incubation in the dark for 3 hours at 25°C, the reaction was quenched by the addition of 100 μl of 0.1 N HCl. The TNBS labeled protamine was purified by an automated HPLC system (model LCC-501 FPLC Plus System, Pharmacia Biotech, Inc., Piscataway, NJ) equipped with a Sephadex G-25 column (16 × 500 mm). The elution buffer was 50 mM Tris-HCl, pH 8.1, containing 0.01 M CaCl2. After purification, the TNBS labeled protamine was digested with thermolysin by using a protamine concentration at least 100 times higher than that of thermolysin. The reaction mixture was incubated in the dark for 3.5 hours at 40°C, followed by rapid heating in a boiling bath. An aliquot of the digested protamine was fractionated on a Sephadex G-25 column by using the same FPLC system mentioned above. The elution profile of the low molecular weight protamine (LMWP) fragments was monitored by measuring the absorbance of TNBS at 340 nm. After the elution profile had been established, the LMWP fragments for subsequent studies were prepared in the same manner as described, except for protamine that was not labeled with TNBS. The concentration of protamine and LMWP was determined by the method of Bradford, 10 using the Coomassie Plus dye reagent. Amino acid analysis of the LMWP fragments was performed by the Protein and Carbohydrate Research Center at the University of Michigan, Ann Arbor, Michigan.
Neutralization of Heparin/LMWH by Protamine/LMWP
Heparin (or LMWH) neutralization by protamine and LMWP was measured in plasma by the anti-Xa chromogenic and activated partial thromboplastin time (aPTT) clotting assays according to our previous work. 11 In brief, for the anti-factor Xa assay, a series of plasma samples containing 0.4 mg of heparin (or 0.2 mg of LMWH), 15 IU of ATIII, and various amounts of protamine (0 - 0.8 mg) or LMWP (0 - 0.4 mg) were prepared. One hundred microliters of the plasma sample were then mixed with 800 μl of Tris-HCl buffer (pH 7.4), followed by the addition of 100 μl of Factor Xa (7.1 nkat/ml). After a 3 min incubation at 37°C, 200 μl of S-2238 substrate (0.5 mg/ml) were added. The reaction mixture was incubated for 5 min at 37°C and then quenched by the addition of 200 μl of 20% acetic acid. The absorbance in the solution was measured at 405 nm, and the degree of heparin neutralization is proportional to this measured absorbance.
For the aPTT clotting assay, 15 μl of heparin solution (20 μg/ml) was mixed with 15 μl of solution containing various concentrations of protamine (0–60 μg/ml) or LMWP species (0–100 μg/ml). To the mixture, 100 μl of actin cephaloplastin and 100 μl of plasma were added. After 3 min of incubation, 100 μl of 0.02 M calcium chloride (preheated to 37°C) was added, and the clotting time was immediately measured by using a fibrometer (Fibrosystem, Becton Dickinson Company, Cockeysville, MD).
The immunogenicity of protamine and LMWP was examined in mice. The production of polyclonal antibodies was performed according to the method of Cooper and Paterson. 12 Fifteen thin-skinned albino mice, 10 for protamine and 5 for LMWP (due to a limited supply of this material, only 5 mice were used), were included in this pilot study. Each mouse was immunized with 50 mg of either protamine or LMWP in complete Freund’s adjuvant (CFA). Four weeks later, animals were bled and the first booster was given with 5 mg of protamine (or LMWP) in incomplete Freund’s adjuvant (IFA). Second and third boosters were given at 2 week intervals. Afterward, blood was collected, allowed to clot, and centrifuged to collect serum. All sera were tested for IgG antibodies by using the modified ELISA method of Singh and Tingle. 13 Protamine or LMWP was used to coat the wells of the microplate to capture related antibodies. The detection antibody was goat-antimouse-IgG alkaline phosphatase reacted with p-nitrophenylphosphate to produce the absorbance readings at 405 nm. The antigenicity of LMWP was examined by following the same ELISA method described above, but by using a LMWP coated microplate for antibody detection in the protamine immunized mice.
The TNBS labeled protamine was digested with thermolysin, a protease that does not cleave the arginyl bonds. The primary reason for selecting thermolysin was to avoid excess digestion of protamine and retain the structural integrity of its arginine sequence. In addition, protamine is also known to contain a few cleavable sites (e.g., Val, Leu, etc.) for thermolysin. 1 Figure 1 depicts the elution profiles from a Sephadex G-25 column of native protamine (first peak) and thermolysin treated protamine (second peak). The shift of the protamine peak toward the right indicated that the products of thermolysin treated protamine were low molecular weight fragments. These fragments were referred to as low molecular weight protamine (LMWP), showing a molecular weight distribution range of 400 to 1,400 daltons, with the peak around 1,200 daltons.
The peak fractions from #15 through #18 in Figure 1 were pooled for subsequent studies. This pooled fraction was not homogeneous and contained at least 3 to 4 LMWP fragments as resolved by reverse-phase liquid chromatography (data not shown). Amino acid analysis of this pooled fraction revealed a composition of 5–6 arginine, 1 proline, 1 serine, and 1 alanine residues, along with trace amounts of threonine, valine, and glycine residues. To evaluate the functions of these LMWP fragments in neutralizing the anticoagulant properties of heparin and LMWH, both the aPTT clotting assay and the anti-Xa chromogenic assays were adapted. As shown in Figure 2, both protamine and LMWP effectively neutralized heparin induced aPTT activity. However, the LMWP was not as effective as protamine, because twice as much LMWP (∼3.7 mg) as protamine (∼1.7 mg) was required to completely neutralize the aPTT activity of heparin (1 mg). This finding was somewhat anticipated, considering the nonspecific nature of the aPTT assay, which is a general method based on the measurement of activation of the intrinsic system of the coagulation cascade. 14 A wide variety of coagulation factors, including thrombin, are involved in the neutralization of aPTT activity. Therefore, it is of no surprise that LMWP with its significantly shortened chain length would require a much higher dose to completely displace thrombin from its binding to heparin, as reflected by the aPTT data.
Interestingly, Figure 3 shows that LMWP was more effective than protamine (on a weight basis) in neutralizing the anti-Xa activity of heparin, as measured by the chromogenic assay by using S-2222 as the specific substrate. Complete neutralization of the anti-Xa activity of the same dose of heparin (40 μg) was achieved with 15 nmole (67 μg) of protamine and 25 nmole (30 μg) of LMWP. These estimated neutralization end-points seem to be reliable and consistent, because the doses required for 50% heparin neutralization (shown by arrows and determined as the half-points of ΔAbs405nm in Figure 3) occurred approximately at one-half the doses (i.e., 7 nmole and 12 nmole for protamine and LMWP, respectively) required for 100% neutralization.
Findings in Figure 3 suggest that LMWP may bind heparin more specifically at heparin’s pentasaccharide sequence than does protamine. To confirm this speculation, experiments involving the neutralization of the anti-Xa activity of low molecular weight heparin (LMWH) by both protamine and LMWP were conducted. As shown in Figure 4, LMWP was indeed more effective than protamine in neutralizing the anti-Xa activity of CY216 (20 μg). At the same dose of 30 μg, LMWP completely neutralized the anti-Xa activity of CY216, whereas protamine only yielded approximately 60% neutralization. It should be pointed out that CY216 is one of the LMWH compounds currently used clinically.
Immunogenicity of both protamine and LMWP was examined in mice by monitoring the production of polyclonal antibodies. Although the antibody titer was not very high (only 1:50 dilution), 4 of 10 mice showed a positive reaction for antibodies to protamine (see the shaded bars in Figure 5). On the other hand, none of the five mice immunized with LMWP exhibited any detectable evidence of antibody production when examined by using the LMWP coated wells. Although the data seem to suggest an absence of immunogenicity for LMWP, they are inconclusive due to the lack of a statistically sufficient number of test animals in this study.
The antigenicity of LMWP, which is defined as its ability to be recognized by anti-protamine antibodies, was also examined by using the same ELISA method, but by using a LMWP coated microplate for antibody detection in protamine immunized mice. As shown by the solid bars in Figure 5, anti-protamine antibodies exhibited an extremely low level of cross-reactivity, if any at all, toward LMWP, and the absorbance readings were less than one-tenth for LMWP when compared with those for protamine. These findings confirm our second hypothesis, i.e., LMWP derived from protamine with a molecular mass of approximately 1,200 daltons is markedly devoid of immunologic properties.
The use of protamine sulfate after cardiac and vascular surgery is a step routinely taken to alleviate postoperative bleeding. Despite its frequent use, protamine remains implicated in uncommon but dramatic reactions. Although the commonly occurring anaphylactoid responses can be largely aborted with conventional pressor agents and slow protamine administration, the immunoglobulin mediated anaphylactic responses are unpredictable, life-threatening, and not manageable by any clinical intervention, thereby posing a major safety threat. The LMWP fragments, which retain full heparin neutralizing ability and yet are devoid of the immunologic properties of protamine, therefore, offer significant clinical benefit.
The finding that on a same weight basis LMWP is more effective than protamine in neutralizing the anti-Xa activity of heparin is of great clinical importance. It is well established that the prelude of the heparin induced anti-Xa event lies in its binding with ATIII by means of a specific pentasaccharide sequence. 7 To this end, a variety of LMWH compounds containing an ATIII-binding pentasaccharide domain have recently been successfully derived from heparin in an attempt to replace heparin with an improved anticoagulant without the risk of bleeding. 7,15 Although these LMWH compounds may be safer as anticoagulants than heparin as demonstrated in both laboratory and clinical settings, their uses are nevertheless not yet completely exempted from bleeding risks. 7,15 Added to this scenario is a clinical dilemma: unlike heparin, the anticoagulant functions of LMWH cannot be effectively neutralized by protamine 7,15 or any currently existing neutralizing agents, such as platelet factor 4. 16 Thus, if the LMWP fragments could completely reverse the anticoagulant activity (i.e., anti-Xa activity) of LMWH compounds, as our preliminary data on neutralization of CY216 indicates (see Figure 4), then it would potentially encourage the clinical acceptance of LMWH as a safer, all-purpose anticoagulant drug.
The presence of markedly reduced antigenicity (i.e., cross reactivity) toward anti-protamine antibodies by LMWP is also of particular clinical significance. Aside from its use in heparin neutralization, protamine has another major medical application. It is routinely combined with insulin to formulate NPH (neutral protamine Hagedorn) insulin or PZI (protamine zinc insulin). These formulations are shown to retard the adsorption of insulin and allow insulin-dependent diabetic patients to achieve euglycemia with less frequent insulin injections. However, this previous exposure to protamine renders diabetic patients highly susceptible to a severe protamine response. Nell and Thomas 17 reported that 30% of NPH treated diabetes had IgG anti-protamine antibodies by ELISA assay. In a recently published report, Lavinson and Ohm 18 estimated that in insulin-dependent diabetes, there was a 10-fold increased risk of a hemodynamically significant protamine reaction at the time of cardiac surgery compared with nondiabetic controls. The lack of cross-reactivity between LMWP and anti-protamine antibody would enable this large population of diabetic patients, who would be given LMWP as a substitute for protamine during heparin reversal, to be protected from risks of immunoglobulin mediated, severe protamine responses.
A thorough review of the already established mechanisms of protamine induced adverse responses by means of the nonimmunologic pathway suggests that the majority of such responses are attributed to the “crosslinking” ability of protamine due to its polycationic and polymeric nature. For instance, complement activation, which is one of the main events in this nonimmunologic pathway, is primarily due to crosslinking of heparin by protamine to form antigen-antibody like large network structures. 19 The LMWP species, which possess significantly shorter chain lengths than protamine, may have no such crosslinking ability and, as a consequence, may produce less toxicity. Further investigation is currently in progress in our laboratory.
This research was supported in part by NIH grants HL38353 and HL55461.
1. Ando T, Yamasaki M, Suzuki K: Protamines: Isolation, characterization, structure, and function, in Kleinzeller A (ed), Molecular Biology, Biochemistry, and Biophysics, vol. 12. New York, Springer-Verlag, 1973, pp. 1–109.
2. Metz S, Horrow JC: Protamine and newer heparin antagonists, in Stoelting RK (ed), Pharmacology and Physiology in Anesthetic Practice, vol. 1. Philadelphia, Lippincott, 1994, pp. 1–15.
3. Horrow JC: Protamine: A review of its toxicity. Anesth Analg 64: 348–361, 1985.
4. Lindblad B: Protamine sulfate: A review of its effects: Hypersensitivity and toxicity. Eur J Vasc Surg 3: 195–201, 1989.
5. Weiler JM, Gellhaus MA, Carter JG, et al: A prospective study of the risk of an immediate adverse reaction to protamine sulfate during cardiopulmonary bypass surgery. J Allergy Clin Immunol 85: 713–719, 1990.
6. Okajima Y, Kanayama S, Maeda Y, et al: Studies on the neutralizing mechanism of antithrombin activity of heparin by protamine. Thromb Res 24: 21–29, 1981.
7. Rosenberg R: The heparin-antithrombin system: a natural anticoagulant mechanism, in Colman RW, Marder VJ, Hirsh J (eds), Hemostasis and Thrombosis: Basic Principles and Clinical Practice, 2nd edition. Philadelphia, Lippincott, 1987, pp. 1373–1392.
8. Sela M: Antigenicity: Some molecular aspects. Science 166: 1365–1374, 1969.
9. Lerner RA: Antibodies of predetermined specificity in biology and medicine, in Dixon FJ (ed), Advances in Immunology, vol. 36. New York, Academic Press, 1984, pp. 1–44.
10. Bradford MM: A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding. Anal Biochem 72: 248–254, 1976.
11. Yang VC, Bernstein H, Cooney CL, Kadam J, Langer R: Removal of the anticoagulant activities of the low molecular weight heparin fractions and fragments with flavobacterial heparinase. Thromb Res 44: 599–610, 1986.
12. Cooper HN, Paterson Y: Production of antibodies, in: Coligan JE, Kruisbbek AM, Margulies DH (eds), Current Protocols in Immunology
. New York, Green Publishing Associates and Wiley-Interscience, 1991, sect. 2.4.1–2.4.7.
13. Singh VK, Tingle AJ: Detection of circulating immune complexes by C1q-microplate ELISA system. J Immunol Methods 50: 109–114, 1982.
14. Walenga JM, Fareed J, Hoppensteadt D, Emanuele RM: In vitro evaluation of heparin fractions: Old vs. new methods. Crit Rev Clin Lab Sci 22: 361–389, 1986.
15. Rang HP, Dale MM, Ritter JM, Gardner P: Hemostasis and thrombosis, in Rang HP, Dale MM, Ritter JM, Gardner P (eds), Pharmacology. New York, Churchill Livingstone, 1995, pp. 331–350.
16. Cook JJ, Niewiarowski S, Yan Z: Platelet factor 4 efficiently reverses heparin anticoagulation in the rat without adverse effects of heparin-protamine complexes. Circulation 85: 1102–1109, 1992.
17. Nell LJ, Thomas JW: Frequency and specificity of protamine antibodies in diabetic and control subjects. Diabetes 37: 172–176, 1988.
18. Levinson MM, Ohm W: The pathophysiology and treatment of severe protamine reactions in cardiovascular surgery, in The Heart Surgery Forum on Internet: email@example.com.,
June 3, 1995.
19. Rent R, Ertel N, Eisenstein R, Gewurz H: Complement activation by interaction of polyanions and polycations. I. Heparin-protamine induced consumption of complement. J Immunol 114: 120–124, 1975.