The Effect of Nonionic Surfactant Structure on Hemolysis Masako Ohnishi and Hiromichi Sagitani* Pola Chemical Industries Inc., Yokohama Research Center, Totsuka, 560, Kashiocho, Totsuka-ku, Yokohama, 244, Japan 679 The hemolytic properties of polyoxyethylene-type non- ionic surfactants were investigated in the concentration range of 0.1-5.0 wt/vol%. The effects of hydrophilic- lipophilic balance, molecular weight, chemical structure and solubilizing ability of the nonionic surfactants in hemolysis are discussed. Bulky surfactants with three or four alkyl chains in their molecules, which show low solubilizing ability of lipids, were the least likely to in- duce hemolysis. Morphological observations of erythro- cytes by video-enhanced microscopy showed that linear surfactants induce the formation of spherocytes and cause hemolysis, whereas bulky surfactants induce only morphological changes from discocytes to spheroechino- cytes. KEY WORDS: Bulky structure, hemolysis, morphological change, nonionic surfactant. For intravenous medicines it is important to develop safer solubilizers without hemolytic activity. It is also neces- sary to clarify the relationship between the hemolytic properties of nonionic surfactants and their structures for future synthesis of new types of surfactants. Many sur- factants induce hemolysis near the critical micelle concen- tration (CMC) (1-4}. Hemolysis studies have usuallybeen carried out in the CMC range (10-3-10-5M) (5-7). How- ever, it is more important to understand the hemolytic properties of nonionic surfactants in the larger concentra- tion range used for their practical applications. Nonionic surfactants are used in intravenous medicines as solubi- lizers at a concentration of around 7% (8). We studied the hemolytic activity of polyoxyethylene-type (POE-type) nonionic surfactants in the concentration range of 0.1-5.0 wt/vol% and focussed on the relationship between the structure of surfactants and their hemolytic ability. The hemolytic properties of surfactants depend on fac- tors such as the chemical structure, solubilizing ability and hydrophilic-lipophilic balance (HLB). Isomaa et al. (9) reported that the hemolytic potency of ionic and ampho- teric surfactants decreases with increasing alkyl chain- length. Fukuda (10) showed that POE(n) alkyl ethers com- prising 9-10 oxyethylene units have maximum hemolytic activity, and Kondo et al. (5) reported that hemolytic ac- tivities of POE(n) dodecyl ethers decrease with increas- ing length of the POE chain. Segal et al. (6) showed that POE chainlength does not affect hemolysis in terms of molarity. Miyajima et al. (3) showed that POE(30) choles- teryl ether has the maximum hemolytic activity in the POE(n) cholesteryl ether series, and Zaslavsky et al. (11) reported that hemolysis does not correlate with the HLB of surfactants. It seems from these results that surfac- tants with long alkyl and POE chains show low hemolytic activity, but this has not yet been clarified. We studied the hemolytic activities of nonionic surfactants contain- ing several alkyl chains in the molecule and considered the effects of structure on hemolysis. *To whom correspondence should be addressed. Copyright © 1993 by the American Oil Chemists' Society MATERIALS AND METHODS Surfactants. POE(n) oleyl ether, POE(n) monostearate, POE(20) sorbitan monooleate, POE(n) hydrogenated cas- tor oil, POE(n) sorbitol tetraoleate and POE(20) polyox- ypropylene (POP) (6) 2-decyltetradecyl ether were pur- chased from Nikko Chemicals Co. (Toky~ Japan). POE(20) 2-decyltetradecyl ether was obtained from Nihon Emul- sions Co. (Tokyo, Japan). The formulas of the surfactants are shown in Table 1. Egg yolk lecithin (PL-100LE, 95% purity) was purchased from QP Inc (Tokyo, Japan). Fluo- rescence spectroscopy-grade 5(6)carboxyfluorescein (CF) was obtained from Eastman Kodak (Rochester, NY). These materials were of commercial grade and were used without further purification. Erythrocytes. Blood was drawn from Japanese white rabbits (either sex, 2.0-3.0 kg). Erythrocytes were sepa- rated by centrifugation (2,000 rpm; 2,400 X g) for 5 min and resuspended after washing in an isotonic buffer solu- tion {0.15 M NaC1 and 0.01 M sodium phosphate at pH 7.0). After repeating this procedure three times, the sus- pension was adjusted to 2 vol/vol%. The concentration was measured at 1.14 × 10 s cells/mL with a Coulter counter. Hemolysis. Two mL of the erythrocyte suspension and 2 mL surfactant buffer solution were mixed in a test tube. After gentle mixing, the tubes were stored in a tempera- ture-controlled water bath at 37°C for 15, 30 or 120 min and then centrifuged at 3,500 × g for 15 rain. The extent of hemolysis (as a percentage) was obtained from the ab- sorbance (540 nm} of the supernatant. A completely hemo- lyzed control sample was prepared by dilution of the erythrocyte suspension with a 1.0 wt% aqueous solution of Triton X-100. A stable erythrocyte sample was prepared by the addition of the isotonic buffer solution to the erythrocyte suspension. Preparation of liposomes. Liposomes containing CF were prepared by a similar method to that of Miyajima etal. (4). Egg yolk lecithin (0.075 g) and cholesterol (0.025 g) were dissolved in 5 mL chloroform, and this solu- tion was dried to a thin film with an evaporator and stored in vacuum overnight. The lipid mixture was dispersed in 200 mM CF Tris HC1 buffer solution (20 mM, pH 7.4) by vortexing and sonicating with a probe-type sonicator (Branson Sonic Power Co., Danbury, CT) at 30 W for 15 min under cooling in an ice bath. Uneneapsulated CF was removed from the liposome solution in a Sepharose 4B col- umn at 5°C. The liposome solution (2.8 mL) was mixed with 0.2 mL of 0.1 wt% aqueous surfactant solution at 23 °C and stored for 1 h. The percentage of leaked CF was determined by the change in fluorescence intensity at 520 nm with an exitation wavelength of 470 nm. Complete leakage of CF is equivalent to hemolysis. Phase diagrams. The solubility regions were determined by visual inspection of samples in glass vials kept at 37°C. The samples were prepared by titration of water into a mixture of the other components. Observation of morphology. Erythrocyte suspensions were diluted with 0.15 M NaC1 and 0.01 M phosphate buf- fer at pH 7.0. This solution was then placed on a capillary microslide (Vitro Dynamics Inc., Rockaway, N J) and was JAOCS, Vol. 70, no. 7 (July 1993)
`
`Ex. 2021-0001
`
`

`
`680 M. OHNISHI AND H. SAGITANI TABLE 1 Chemical Formulas of Nonionic Surfactants a POE(n) oleyl ether C18:lHa5-O-(CH2CH20)nH POE(n) monostearate C17H35- C -O-(CH2CH20)nH II O POE(20) sorbitan monooleate H(CH2CH20)a-O-CH-CH-O-(CH2CH20)bH L L CH 2 CHCHCH2-O- C-C17.1H33 \ / L H ' O O O-(CH2CH20)cH POE(20) 2-decyltetradecyl ether C10H21 /CHCH2-O-ICH2CH2OleoH C12H25 POE(20)POP(6) 2-decyltetradecyl ether CloH21 CH3 \ I /CHCH2-O-lCH2CHO16-(CH2CH20)2oH C12H25 POE(n) hydrogenated castor oil O-(CH2CH20)dH I CH2-O-(CH2CH20) a- C -(CH2)loCH(CH2)sCH3 II O ?-(CH2CH20)eH CH-O-(CH2CH20) b- C -(CH2)loCH(CH2)5CH 3 II O O-(CH2CH20)fH I CH2-O-(CH2CH20)c- ~-(CH2)I°CH(CH2)5CH3 O POE(n) sorbitol tetraoleate CH2-O-(CH2CH20)a- ~ -C17'1H33 O " CH-O-(CH2CH20)b- ~-C17:1H33 O I CH-O-(CH2CH20)c~-Ct7:lH33 O CH-O-ICH2CH20)dH I CH-O-ICH2CH20)eH I CH2-O-(CH2CH20)f-~-C17:lH33 O apOE, polyoxyethylene; POP, polyoxypropylene. JAOCS, VoI. 70, no. 7 (July 1993)
`
`Ex. 2021-0002
`
`

`
`NONIONIC SURFACTANTS AND HEMOLYSIS 681 brought into contact with 2 wt% aqueous surfactant solu- tion containing 0.15 M NaC1 and 0.01 M phosphate on the microslide. The hemolytic process was followed with a video-enhanced microscope (VEM} (Nikon, HPD Micro- scope; Hamamatsu System Inc., Argus 100, Hamamat- su, Japan} and was recorded on floppy disk. RESULTS AND DISCUSSION The surfactants tested in this study were all POE-type nonionic surfactants. The hemolytic activities of these sur- factants are shown as the extent of hemolysis expressed as a percentage after mixing erythrocytes and surfactants at 37 °C. Table 2 summarizes the percentage hemolysis of erythrocytes. Hemolysis decreased as the POE chain- length increased in the POE{n) oleyl ether and POE(n) monostearate series, and hemolysis was not observed in the POE in) hydrogenated castor oil or the POEin) sorbitol tetraoleate series. The percentage hemolysis in the POE(20} 2-decyltetradecyl ether solution was lower than that in the POE(20) oleyl ether solution. Branching of the lipophilic part is effective in decreasing the hemolytic ac- tivity of surfactants. The introduction of a POP chain in POE(20) 2-decyltetradecyl ether induces a further reduc- tion in hemolytic activity. POE(20) POP(6) 2-decyltetra- decyl ether has methyl branches in the POP chain. The bulky structure of the binding site between the hydro- philic group (POE chain} and the lipophilic group (alkyl chain} may be more effective in reducing hemolytic activ- ity than the branched alkyl chain. POE{n) hydrogenated castor oil and POE{n} sorbitol tetraoleate have three and four binding sites with alkyl chains in their molecules, respectively, so there is no hemolysis with these surfac- tant systems because of their bulky structure. Time dependence of hemolysis was observed in the POE(20) sorbitan monooleate, POE(20) 2-decyltetradecyl ether, POE(20) POP(6) 2-decyltetradecyl ether, POE(50) oleyl ether and POE(55) monostearate systems. POE(20) sorbitan monooleate, POE(20) 2-decyltetradecyl ether and POE(20) POP{6) 2-decyltetradecyl ether do not have linear structures. Although POE{50) oleyl ether and POE(55) monostearate have linear structures, their hemolytic ability may be low because of their high hydrophilicity and, therefore, these surfactants belong to the category of slow-reaction hemolytic surfactants described by Weltzien {12}. In hemolysis, concentration dependence was not ob- served in any surfactant systems except in the POE(55) monostearate, POE(20) 2-decyltetradecyl ether and POE(20) POP{6) 2-decyltetradecyl ether systems. All these surfactants belong to the surfactant group showing time- dependent hemolysis. Surfactants with moderate hemo- lytic activity show time- and concentration-dependent hemolysis; however, surfactants of low hemolytic activ- ity do not show such time and concentration dependence in this concentration range. The influence of molecular weight of surfactants on hemolysis is shown in Figure 1, which compares percent- age hemolysis and the molecular weight of surfactants. The percentage hemolysis has a tendency to decrease with an increase in molecular weight of nonionic surfactants. Only the results of the 5 wt/vol% surfactant aqueous solu- tions are shown in Figure 1 because the 0.1 wt/vol% and 1.0 wt/vol% solutions showed the same tendency. The solubilizing power of surfactants with long oxyethylene chains decreases because of their low ability to aggregate {13,14}. The hemolytic ability of surfactants with high molecular weight was low in the POE-type nonionic TABLE 2 Percentage of Hemolysis by Some Types of Nonionic Surfactants a Surfactant concentration (wt/vol%): 0.1 1.0 5.0 Incubation time (min}: 15 30 120 15 30 120 15 30 120 Surfactant MW HLB-n b Hemolysis in %c,d POE(10} oleyl ether 708 14 100 100 100 100 100 100 100 100 100 POE(20) oleyl ether 1148 17 100 100 100 100 100 100 100 100 100 POE(50) oleyl ether 2468 18 52 74 78 49 57 65 32 32 53 POE(25) monostearate 1429 15 -- 80 82 -- 83 87 -- 85 85 POE(40) monostearate 2089 17 -- 72 76 -- 78 81 -- 66 77 POE(55) monostearate 2749 18 -- 0 18 -- 1 44 -- 26 72 POE(20) sorbitan monooleate 1326 15 0 -1 27 0 -1 65 0 48 65 POE(20) 2-decyltetradecyl ether 1233 2 65 100 100 100 100 100 100 100 POE(20) POP(6) 2-decyltetradecyl ether 1582 11 -1 -1 0 -2 -2 7 -2 -2 35 POE(20) hydrogenated castor oil 1874 10 -1 -1 -2 -1 -1 0 0 2 2 POE(40) hydrogenated castor oil 2754 12 -2 -2 -2 -2 -2 -2 -2 -2 -2 POE(60) hydrogenated castor oil 3634 14 --1 --1 -1 --2 -2 -1 -2 -2 -3 POE(40) sorbitol tetraoleate 3072 12 0 0 -1 --1 "1 --2 0 0 -1 POE(60) sorbitol tetraoleate 3952 14 --2 --2 -2 -1 -1 -1 -4 -4 --3 ~Abbreviations as in Table !. MW, molecular weight; HLB, hydrophilic-lipophilic balance. bHLB numbers were calculated by the emulsification method (Ref. 15). CErythrocyte concentration: 1 vol/vol%. d0% and 100% hemolysis were obtained by the addition of isotonic buffer solution and 1.0 wt% Triton X-100 aqueous solution, respectively. JAOCS, Vol. 70, no. 7 (July 1993)
`
`Ex. 2021-0003
`
`

`
`682 M. OHNISHI AND H. SAGITANI to aS 0 II O 100 50 ~O O ~o y = -0.0327X + 109 o ~r = -0.7609 0 0 • 0 0 0 0 I I I I tO00 2000 3000 4000 Roleculav weight FIG. 1. Hemolysis and molecular weight of nonionic surfactants. Surfactant concentration: 5 wt/vol%/system, incubation time, 30 vain. surfactant system because hydrophilicity increases with an increase in the POE chain. Therefore. the relationship between molecular weight and hemolysis by surfactants of the same HLB number should be considered to under- stand the effects of molecular weight on hemolysis. The HLB number of POE(10) oleyl ether, POE(60) hydroge- nated castor oil and POE(60) sorbitol tetraoleate is 14 (15) (Table 2). The hemolytic abilities of POE(60) hydrogenated castor oil and POE(60) sorbitol tetraoleate, having higher molecular weights, are lower than that of POE(10) oleyl ether. Molecular weight is also important in reducing hemolysis, as well as in reducing HLB. The molecular weight of POE(50) oleyl ether, POE(55) monostearate and POE(40) hydrogenated castor oil are within the 2,500-3,000 range (Table 2), and the hemolytic abilities of the former two are higher than that of the lat- ter. The former surfactants have a single alkyl chain in the molecule and the latter has three alkyl chains. It is effective to introduce a bulky structure at the binding site to decrease hemolysis. Hemolysis starts with solubilization of erythrocyte lipids and proteins by surfactants {6,16,17). This is why many investigations concerned with hemolysis and sur- factants were carried out with surfactant solutions near the concentration range of CMC {1-4). However, POE(n) hydrogenated castor oil and POE(n) sorbitol tetraoleate do not induce hemolysis at 5 wt/vol%. The CMC of POE(60) hydrogenated castor oil was shown to be below 0.01 wt/vol% by the solubilization ofp-dimethylaminoazo- benzene; therefore, the mechanism of hemolysis of surfac- tants cannot be explained only by solubilization. Miya- jima et aL. (3) pointed out that the hemolytic activity of POE(n) cholesteryl ethers was not the same as the order of their CMC. Lecithin liposomes encapsulating CF were prepared as an erythrocyte model, and the leakage of CF from lipo- somes caused by the addition of surfactants was mea- sured. The percentage hemolysis and the leakage of CF are shown in Table 3. There is a good relationship between these two results. Percentage hemolysis and leakage of CF decrease with an increase in polyoxyethylene chainlength in the linear alkyl nonionic surfactant series. POE(n) hydrogenated castor oil shows the lowest percentage he- molysis and leakage. From these results, hemolysis is related to the interaction of lipids in erythrocyte mem- branes and nonionic surfactants. If a nonionic surfactant is a good solubilizer of lipids {lecithin), a mixed solution of lecithin and surfactant gives an isotropic micellar solu- tion. The solubilization of lecithin by nonionic surfactants was investigated from ternary phase diagrams of leci- thin/nonionic surfactant/aqueous solution systems (Fig. 2). Aqueous ethanol solution was used as a solvent instead of pure water to avoid complicating the phase diagrams. The mixture of lecithin and nonionic surfactants was dissolved in ethanol, and water was then added to the solu- tion. The isotropic solution region is shown as the shaded area in Figure 2. POE(10) oleyl ether aqueous solution shows a wide region of isotropic surfactant solution [Fig. 2(a)]. More hydrophilic POE(50) oleyl ether does not show such a wide range of isotropic solution [Fig. 2(b)]. These two phase diagrams indicate that POE(50) oleyl ether cannot solubilize as much lecithin in the micelles as POE(10) oleyl ether can. POE(50) oleyl ether may be too hydrophilic to produce mixed micelles with lecithin, and POE(60) hydrogenated castor oil also cannot solu- bilize much lecithin in the micelles [Fig. 2(c)]. POE(60) hydrogenated castor oil is a good solubilizer for lipophilic medicines (e.g., vitamin K), but is not good for lecithin. Morphological changes of erythrocytes by nonionic sur- factants are shown in Figure 3. The erythrocyte structure rapidly changed into a spherocyte within 30 s of contact when POE(10) oleyl ether was added to the erythrocyte solution. The size gradually increased, and the ceils burst within 1 rain. Thus, hemolysis caused by POE(10) oleyl ether appears to be due to osmotic lysis {18}. After about the same contact time, erythrocytes remained as disco- cytes in the POE(60) hydrogenated castor oil solution. The TABLE 3 Correlation Between Hemolysis and Carboxyfluorescein (CF) Release from Liposomes Caused by the Addition of Nonionic Surfactants Hemolysis CF release Surfactant (%)a (%)b POE(10} oleyl ether 100 98 POE(20) oleyl ether 100 96 POE(50) oleyl ether 22 17 POE(25) monostearate 85 54 POE(40) monostearate 66 26 POE(55) monostearate 26 13 POE(20) POP(6} 2-decyltetradecyl ether - 2 24 POE(20) hydrogenated castor oil 2 8 POE{40) hydrogenated castor oil - 2 9 POE(60) hydrogenated castor oil - 2 7 POE(60) sorbitol tetraoleate - 4 37 aSurfactant concentration: 5 wt/vol%, incubation, 30 min. Abbrevia- tions as in Table 1. bSurfactant concentration: 0.0067 wt/vol%, incubation, 1 h. JAOCS, VoI. 70, no. 7 (July 1993)
`
`Ex. 2021-0004
`
`

`
`NONIONIC SURFACTANTS AND HEMOLYSIS Lecithin 50wt% Ethanol 50 wt% 0 /A\\\\\\ 0 50 100 wa*er P0E(20) oleyl ether 50wt% Ethanol 50 wt% (a) 683 Lecithin 50wt% Ethanol 50 wt% 0 / \\ 0 50 100 weber P0E(50) oleyl ether 50wt% Ethanol 50 wt% (b) Lecithin 50wt% Ethanol 50 wt% 0 I 00Y/~\\\\\ ,~\\\\XXXXXXXXk ~ • 0 0 50 100 Wster POE(60) hydr'ogenated castor oil 50wt% Ethanol 50 wt% (C) FIG. 2. Ternary phase diagrams of 50 wt% lecithin ethanol solu- tion/50 wt% nonionic surfactant ethanol solution/water system. Shaded area, isotropic solution region. (a} POE(10) oleyl ether system; (b) POEt50} oleyl ether system; and (c) POE(60) hydrogenated castor oil system. POE, polyoxyethylene. FIG. 3. Morphological changes of erythrocytes caused by the addi- tion of nonionic surfactants. (a) before contact with surfactant; (b) 15 s after contact with POE(10} oleyl ether; (c) 30 s after contact with POEI10} oleyl ether; and {d) 30 rain after contact with POE(60) hydrogenated castor oil. POE, polyoxyethylene. first biconcave-shaped erythrocytes gradually changed into spheroechinocytes after 30 min of contact with POE(60) hydrogenated castor oil. However, the subse- quent spheroechinocyte-spherocyte change did not occur in this system. The discocyte-echinocyte change indicated the adsorption of POE(60) hydrogenated castor oil onto the erythrocyte surface. POE(60) hydrogenated castor oil may not be able to penetrate the erythrocyte membrane because of its bulky structure and low solubilizing abil- ity with lecithin. POEt60) hydrogenated castor oil does JAOCS, VoI. 70, no. 7 (July 1993)
`
`Ex. 2021-0005
`
`

`
`684 M. OHNISHI AND H. SAGITANI not induce hemolysis but affects erythrocyte morphology. POE(10) oleyl ether, having linear alkyl chains, may easily penetrate into the lipophilic layer of the erythrocyte mem- brane and lead to uptake of water. Curvature of the erythrocyte membrane will change with the penetration of single alkyl chain nonionic surfactants because of the tendency for micellar aggregation. However, surfactants with polyalkyl chains in their molecules will not change the lamellar structure of the erythrocyte membrane, even though they penetrate the membrane (19). We conclude from these results that nonionic surfactants with several alkyl chains and high molecular weights are useful as solubilizers for intravenous drugs. REFERENCES 1. Thron, C.D., J. Pharmacol. 145:194 (1964). 2. Tragner, D., and A. Csordas, Biochem. J. 244:605 {1987}. 3. Miyajima, K., T. Baba and M. Nakagaki, Colloid Polymer Sci. 265".943 (1987). 4. Miyajima, K., T. Baba and M. Nakagaki, Ibid 267".201 {1989}. 5. Kondo, T., and M. Tomizawa, J. Pharm. ScL 57".1246 {1968}. 6. Azaz, E., R. Segal and I. Mile-Coldzweig, Biochim. Biophys. Acta 646:444 {1981). 7. Bonzal, R.W., and S. Hunt, Ibid. 249:266 (1971). 8. Physicians' Desk Reference, 45th edn., Medical Economics Co., Oredell, 1991, p. 1376. 9. Isomaa, B., H. Hagerstrand, G. Paatero and A.C. Engblom, Biochim. Biophys. Acta 86@.510 {1986}. 10. Fukuda, M., M. Koide and K. Ohbu, J. Jpn. Oil Chem. Soa 36:576 (1987). 11. Zaslavsky, B.Y., N.N. Ossipov, V.S. Krivich, L.P. Baholdina and S.V. Rogozhin, Ibid. 507".1 (1978). 12. Weltzien, H.V., Ibid. 311:6 (1973). 13. Lange, H., KoU. Z. Z. Polymere 201:131 {1965}. 14. Saito, H., and K. Shinoda, J. Colloid Interface Sci. 24:10 {1967}. 15. Handbook, Drug & Cosmetic Materials, edited by K. Hikime, Nikko Chemicals, Tokyo, 1977, pp. 880-893. 16. Kondo, M., M. Yoshimura and N. Okuyama, Seikagaku 44:849 (1972). 17. Lichtenberg, D., R.J. Robson and E.A. Dennis, Biochim. Biophys. Acta 737".285 {1983}. 18. Tanaka, Y., K. Inoue and S. Nojima, Ibid. 600:126 (1980). 19. Shinoda, K., and H. Sagitani, J. Phys. Chem. 87.'2018 (1983). [Received October 16, 1991; accepted January 6, 1992]
`
`Ex. 2021-0006