Hemagglutination by Dengue Viruses

NFECnON AND MMuNiTY, Feb. 1971, p. 193-199 Copyright 1971 American Society for Microbiology Vol. 3, No. 2 Printed in U.S.A. Nonspecific Factors in Monkey Tissues and Serum Causing nhibition of Plaque Formation and Hemagglutination by Dengue Viruses DONALD R. NASH,1 SCOT' B. HALSTEAD, ANDREW C. STENHOUSE,2 AND CAROLYN MCCUE Section of Tropical Medicine and Medical Microbiology, University of Hawaii School of Medicine, Leahi Hospital, Honolulu, Hawaii Received for publication 19 August 197 Normal monkey serum and the supernatant fluid from different triturated monkey tissues have been studied for the presence of nonspecific arbovirus hemagglutination and plaque forming inhibitors of dengue viruses types 1, 2, 3, and 4. Hemagglutination inhibition (H) activity was present in most tissue specimens and demonstrated a significant gradient of effectiveness starting with, respectively, serum, spleen, adrenal, and lung having a high degree of activity, whereas skin, heart, muscle, brain, and liver demonstrated low H titers. A slightly reversed gradient of effectiveness was obtained for the case of dengue virus inhibition of plaque formation with bile, liver, thymus, spleen, and adrenal giving high 5% plaque reduction titers and heart, muscle, serum, skin, and fat demonstrating little or no activity. Analysis by Sephadex G-2 chromatography and sucrose density gradient centrifugation suggests that H and plaque formation inhibition are independent activities of normal serum or tissue constituents or both. Also, in addition to the physical methods of characterization, chemical treatment by absorption with kaolin or acetone extraction indicate both phenomena to be the result of the action of lipids or lipoproteins. The hemagglutination of red cells by group B arboviruses may be affected nonspecifically in two ways by serum from normal animals of widely differing species. Firstly, normal bovine serum has been shown to contain natural agglutinins to goose red cells (14). Secondly, nonspecific serum inhibitors of viral hemagglutination (HA) have been demonstrated in human, bovine, rabbit, rat, chicken and other avian species, reptile and fish sera. The lipid nature of a human serum HA inhibitor was first perceived by Chanock and Sabin (2) and further studied by Porterfield and Rowe (11) who described the inhibitory activity in serum fractions containing choline, phosphates, and fatty acids. The lipoprotein nature of the inhibitor has been confirmed or inferred in other studies by removal after treatment with acetone or kaolin or both (1, 3, 4, 13-15). The isolation of dengue virus from tissues obtained at postmortem of patients dying with 1 Present address: W. H.. Reference Centre, Lausanne, Switzerland. 2 Present address: Nangrahar University Hospital, Jalalabad, Afghanistan. dengue hemorrhagic fever has been singularly unrewarding (1). Assuming virus to be present in these tissues, it seemed probable that the techniques used for their isolation might also release nonspecific neutralizing substances. The purpose of this paper is to describe the presence and relative importance of neutralizing activity and inhibitors of hemagglutination present in different tissues of normal rhesus monkeys. n addition, an attempt has been made to characterize and demonstrate differences between serum and tissue inhibitors of hemagglutination and nonspecific factors in normal tissues which inhibit plaque formation by dengue viruses. MATERALS AND METHODS Animals. Adult, female rhesus monkeys purchased from ndia were used after a 6-week quarantine period. The serum of each animal was tested for neutralizing activity against each of the four dengue virus types as described below. Specific antibody was also checked by hemagglutination inhibition (H) tests on the same sera after absorption with kaolin. The monkeys used in these experiments demonstrated no 193

194 NASH ET AL. NFEC. MMUN. significant antibody activity against any of the dengue viruses studied. Hemagglutination inhibition. All H tests were carried out in microtiter plates employing a modification of the method of Clarke and Casals (3). Serial twofold dilutions of tissue suspensions or serum were incubated with to hemagglutinating units of antigen at 4 C for 1 hr after which 2 drops of a.4% suspension of fresh washed goose red cells were added. No agglutination of red blood cells (RBC) is indicative of the presence of inhibitor. The viral antigen was prepared from infectious mouse brain suspensions by the sucrose acetone extraction method described by Work (19). Several group B and two group A arboviruses were used: dengue 1 (Hawaii), dengue 2 (New Guinea B), dengue 3 (H7), dengue 4 (H1), Zika, Tembusu, Jap B (Nakayama), Sindbis (Bg-Ar339), and Chikungunya (Ross). Plaque formation inhibition. The ability of tissue extracts to inhibit plaque formation by infectious dengue virus of either type was determined by using the plaque reduction test in 1-oz prescription bottles (). Briefly,.3 ml of three different concentrations of a tissue extract was added to.3 ml of virus suspension and allowed to incubate in a water bath at 37 C for 1 hr after mixing on a vortex mixer. After this initial incubation,.2 ml of virus-extract mixture was added to 1. ml of diluent, and.2 ml of this mixture was added to each of three plaque bottles containing MK2 cell monolayers. The concentrations of each tissue extract before mixing with the virus suspension varied somewhat depending on tissue size. Generally, the three dilutions used ranged as follows: (i) 1:5-1:2, (ii) 1:5-1:4, (iii) 1:5-1:4,. Each plaque bottle receiving the diluted virus-extract mixture was then incubated at 37 C for 1.5 hr with periodic rotation of the bottle. At the end of the second incubation period, the fluid mixture was poured off, and the one stage agar medium overlay was added to the cell sheet (6). n order to maintain a standard for each virus studied, the number of plaque forming units added to each 1 -oz bottle was kept at between 4 and 65. Tissue extracts. Normal rhesus monkeys known to be free of specific serum neutralizing or hemagglutinating antibodies against the four dengue virus types were anesthetized with 1 M phenobarbital and exsanguinated by cardiac puncture. Each organ to be tested (Table 1) was removed under sterile conditions. A variable quantity of tissue from each organ was then weighed and added to a predetermined volume of diluting fluid which consisted of cold sterile phosphate buffered saline, ph 7.9, 2% agamma calf serum, penicillin, and streptomycin. Trituration was done in a cold sterile mortar and pestle. The suspensions were subsequently centrifuged in stecile tubes, at 2,5 rev/min for 3 min at 4 C. Each supernatant fluid was divided into two samples one of which was stored at -7 C and the other used directly to test for plaque formation and HA inhibition. All tests were performed within hr during which time samples were kept at 4 C. When necessary, repeat experiments were done on samples that had previously been stored at -7 C. Chromatography. Serum or liver extracts were eluted from a Sephadex G-2 column with a tris- (hydroxymethyl) aminomethane (Tris) -hydrochloride buffer. Five-ml fractions were collected, and the relative protein concentration was determined by measuring optical density at 2mm with a model DB spectrophotometer (Beckman nstruments, nc., Fullerton, Calif.). On the basis of the shape of the protein curves (Fig. 1 and 2), individual 5-ml fractions were pooled and concentrated to 1 ml by negative pressure. Sucrose density gradient ultracentrifugation. Samples of monkey liver, adrenal, and lung extracts (.4 ml) or serum (.25 ml) were layered on top of a 4 to 1%/' sucrose gradient. The gradient was formed by overlaying.6 ml of sucrose solution decreasing in concentration by increments of 5%C/D (4, 35, 3, 25, etc.). Each tube contained a sucrose solution volume of 4.2 ml. Centrifugation was carried out at 35, rev/min, at 4 C for hr in a model L-2 preparative ultracentrifuge (Beckman) and SW-5 rotor. Fractions consisting of 23 drops each (approximately.5 ml) were collected after puncturing the bottom of the nitrocellulose centrifuge tube. Each fraction was dialyzed against phosphate buffered saline, ph 7, and its volume was adjusted to.5 to 1 ml. Other characterization procedures. n addition to Sephadex chromatography and density gradient centrifugation, liver extracts or whole serum or both were treated as follows: kaolin absorption; ether extraction; acetone extraction (7); exposure to. to.% trypsin at 22 C for 1 hr; precipitation with 33.3 or 5%, final concentration, (NH4)2SO4; dialysis against distilled water; heating at 37, 56,, and 1 C (Table 4); and filtration through filters (Millipore Corp., Bedford, Mass.) of different sizes. RESULTS Tissue inhibitors of hemagglutination. The extracts of 14 different tissues from two monkeys as well as whole serum and bile were tested for their ability to inhibit hemagglutination of goose red cells by dengue (D1, D2, D3, and D4), Zika, Tembusu, Jap B, Sindbis, and Chikungunya antigens. Preliminary work indicated that the tissue extracts used did not contain any natural agglutinins to the goose red cells. The results of one experiment are recorded in Table 1. Subsequent work with tissues from another monkey gave similar results relative to each tissue and the virus antigen used, but, as might be expected, the quantitative titers were not exactly the same. The highly reproducible qualitative results between tissues and different viruses indicate a gradient of strength of HA inhibitor from serum (highest) to liver (lowest) and that within the four dengue types, dengue 4 is the most susceptible, whereas dengue 1 is least susceptible to the action of any natural nonspecific inhibitors of hemagglutination. Zika virus is highly susceptible to the inhibitory action of tissues and serum and resembles those results obtained with dengue 4. Group A

VOL. 3, 1971 TSSUE AND SERUM NHBTORS OF DENGUE VRUSES 195 TABLE 1. Tissue extract Serum... Spleen... Adrenal... Lung... Jejunum... Aorta... Bile... Kidney... Mesenteric lymph nodes. Thymus... Skin... Axillary lymph nodes... Heart... Muscle... Brain... Liver... Dengue 1 Dengue 2 1 Dengue 3 Dengue 4 Zika Tembusu 192 192 192 192 1 192 54 Jap B a Reciprocal of maximum dilution of 1 g of tissue which inhibits hemagglutination. TABLE 2. 2 Sindbus 4 < nhibition of hemagglutinationz by sevent group B alnd two group A arboviruses by using the superniatatt flu(id of triturated normal montkey tissue- Chikungunya < < Fifty per cenit plaque redllctiont liters of den1glue viruises after iiicubationt with n2ormal monkey bile, serum, or the superniatant fluid from differenit triturated tissue samplesa Tissue extract Bile... Liver... Thymus... Spleen... Adrenal... Lung... Kidney... Brain... Axillary lymph nodes... Mesenteric lymph nodes. Heart... Muscle... Skin... Serum... Fat... Dengue 1 Dengue 2 Dengue 3.~~~~~ 1:3 >1:5 1:3 1:19 1:2 1:25 1: 1:65 1: 1:4 1:45 1:1 1:9 1:4 1:3 1:1, 1:2 NTb 1:1 1:1 1:1 1:5 1:3 NT 1:5 >1:5 1: 1:6 1:9 1:7 1: 1:6 1:55 1:3 1:6 1:2 <1:5 Dengue 4 1:25 1:6 1:9 1:25 1:9 1:4 1:45 1:3 1:2 <1:2 <1:5 a Values recorded were calculated on the basis of a reduction in virus plaques when different dilutions of each tissue extract were mixed with a known number of plaque forming units. bnot tested. arboviruses (Sindbis and Chikungunya) are less affected by these natural inhibitors even though slight inhibition can be demonstrated with some tissue extracts. Tissue inhibitors of infectivity. The ability of different tissue extracts, serum and bile to inhibit growth of infectious dengue virus of each type in tissue culture monolayers is recorded in Table 2. Again, as with H, relative abilities to inhibit dengue viruses were noted for each sample tested. The tissues which were very high in their apparent neutralizing ability were bile, liver, thymus, spleen, adrenal, and lung. Those tissues demonstrating very little or no activity included serum, skin, and fat. Most lymphoid tissue tested (axillary lymph nodes, mesenteric lymph nodes), brain, and kidney demonstrated similar intermediate abilities to inhibit growth of all dengue viruses. nhibitory activity of Sephadex G-2 fractions. Monkey liver extract, serum, and human serum were eluted from a Sephadex G-2 column. The elution patterns and results of H tests on the different fractions of monkey and human sera are illustrated in Fig. 1. t can be noted that the major

1 NASH ET AL. NFEC. MMUN. NORMAL SERUM ELUTED FROM SEPHADEX G-2 11111m E v m Fraction n a 4 z 3 2z z - 1 1s t.. '11 TUBE FG. 1. Hemagglutination inhibition (H) titers offractions of normal human and monkey serum after elution from a Sephadex G-2 column. Each fraction (roman numerals) was concentrated to ml and dialyzed against borate buffered saline, ph 6., before testing for H activity. inhibitory activity is limited to the first two fractions which represent the immunoglobulin (19S) or macroglobulin peak. No H activity was noted in those fractions having a molecular size range below 2,. As can readily be seen (Fig. 1), no significant differences were apparent between the serum of the two species tested. The results of H activity for fractions of liver extract are illustrated in Fig. 2. nhibition was noted in the first fraction which represents the beginning of the ascending portion of the first peak and which also represents the fractions having the lowest protein concentration. Some activity was also present in fractions 3 (D, D4), (D1, 2, 3, and 4), and 1 (1, 2, 3, and 4). Neutralizing activity of G-2 fractions of liver extract are also illustrated in Fig. 2. Virtually all the activity is in the 2nd fraction (first peak). t will also be noted from the results (Fig. 2) that the H activity and plaque formation inhibition activity of liver can be separated. nhibitory activity of fractions obtained from sucrose density gradient centrifugation. Hemagglutination inhibitory activities for the different fractions of liver and serum obtained by centrifugation in a sucrose density gradient are recorded in Table 3. The major H activity for all samples * ėhuman serum o...--.o monkey serum H titers human serum 2 a H titers monkey serum is located in the last fraction which represents the top of the gradient. Neutralizing studies on the fractions of serum were not successful due to the very low activity of whole serum. Although extracts of adrenal gland, thymus, and liver were centrifuged in a sucrose gradient, 5% plaque reduction titers were obtained only with the liver fractions. n this instance, plaque formation inhibition activity was concentrated in the 5th, 6th, and 7th fractions, representing an area near the top of the gradient but slightly lower than the area of maximum hemagglutination-inhibition activity. Effects of different physico chemical treatment. The effects of several different methods of treatment on the ability of liver extracts and serum to inhibit hemagglutination are listed in Table 4. With regards to liver extract, treatment with trypsin enhances inhibition, whereas absorption with kaolin removes H activity. The inhibitor in serum was lost after treatment with kaolin and acetone. The HA inhibitor of serum was partially removed by passage through a.65-,um membrane filter and destroyed by exposure to (NH2)4SO4. n addition, it was found that after dialysis against distilled water, the inhibitor remained in the pseudoglobulin fraction indicating its solubility in water. U.4 n 2 J 1-4 3i

VOL. 3, 1971 7 6 TSSUE AND SERUM NHBTORS OF DENGUE VRUSES 197 LVER EXTRACT ELUTED FROM SEPHADEX G-2 aci 5 4 z 3.- z z Oo 2 z.-. 1 TUBE a 19 H titer * neutralizing titer 9 Z n r, A4 lo 5 = ẓ 4-4 3 : U FG. 2. Hemagglutination inhibition (H) titers and 5% plaque reductioni titers (PRNTso) of fractions of normal monkey liver extract after elution from a Sephadex G-2 column. Each fractiont (roman numerals) was concentrated to ml and dialyzed against borate buffer, ph 7, before testing. Note separation ofh and neutralizing activity. The plaque forming inhibitor present in liver extracts was removed by absorption with kaolin and extraction with acetone. No data were obtained with the remaining parameters for the inhibitors of plaque formation. DSCUSSON The data reported here concerning hemagglutination inhibitors in normal monkey serum are in agreement with similar work described for the case of rabbit serum (9), bovine serum (5), human serum (1, 11, 1), as well as serum from fish (1), rodents (1), and birds (1). The fact that arbovirus agglutination of goose RBC can be inhibited with the macroglobulin fraction of serum obtained by filtration on a Sephadex G-2 column and in the low density fraction after sucrose density gradient ultracentrifugation demonstrates the lipid nature of this inhibitor. The interesting results obtained with the triturated tissue extracts reveal that tissues also contain hemagglutination inhibitors and that the concentration is quite variable for each tissue type giving a gradient of inhibition starting with serum, spleen, adrenal, and lung and ending with muscle, brain, and liver, respectively. The role of serum within the tissue extracts cannot be wholly discounted as a source of this inhibitor. The fact that many tissue extracts inhibited hemagglutination by dengue 4 antigen at greater titers than for other dengue antigens, although serum did not, suggests that tissue and serum inhibitors of HA may be different. However, results of H tests on liver and serum fractions after Sephadex filtration indicate that, in this instance, liver HA inhibitors have some properties similar to those present in serum. Tissue extracts but not normal serum readily inhibit virus growth in MK2 cell monolayers under agar. As might be expected from the work of Sunaga et al. (), Theiler (17), and Hardy et al. (), bile and liver demonstrated the greatest ability to neutralize dengue virus. The high concentrations of bile salts in these samples is probably the reason for this activity. The fact that other organs known to be low or devoid of bile salts (lung, adrenal, muscle, etc.) also inhibit virus infectivity suggests that substances other than bile 1

19 NASH ET AL. NFEC. MMUN. TABLE 3. Relative hemaggluttiniationt inlhibitioni titers of fractions of niormal montkey liver extracts, monkey serum, antd humani serum after sucrose density gradient ultracenitrifiugationt Prepn idenigue 1 Dengue 21Deingue 3 Dengue 4 Whole liver extract Fraction 1l Fraction 2 Fraction 3 Fraction 4 Fraction 5 Fraction 6 Fraction 7 Fraction Fraction 9 Whole monkey serum Fraction 1 Fraction 2 Fraction 3 Fraction 4 Fraction 5 Fraction 6 Fraction 7 Fraction Fraction 9 Whole human serum Fraction 1 Fraction 2 Fraction 3 Fraction 4 Fraction 5 Fraction 6 Fraction 7 Fraction Fraction 9 1 4 4 2 2 >5 a Fraction numbers start at the bottom of the gradient. salts are important. The work described here further suggests the involvement of low density molecules resembling those natural inhibitors of hemagglutination t is apparent, however, that inhibition of infectivity (plaque formation) and HA is not mediated by the same substances. Fractionation of liver extract on a G-2 column resulted in isolating the activity in the first peak (2nd fraction) having a high protein concentration. nterestingly, this fraction was devoid of H activity which was notable in that portion of the curve eluted just prior to this peak (Fig. 2). A separation of plaque forming inhibitors and hemagglutination inhibitors from liver extract was also noted after centrifugation in a sucrose gradient. n this case, the former demonstrated a slightly higher density than the latter. Recent work by Yuill (2) has revealed the presence of inhibitors of plaque formation by extracts of adult mosquitoes. Although the number of viral plaques was reduced in vitro, viral extract mixtures did not appear to lose infectivity in vivo. n vivo studies with monkey tissue extracts and infectious virus have not been attempted. t is clear, however, that virus recovery, particularly group B arbovirus, from triturated tissue samples can be affected by natural inhibitors especially in a plaque system. Less clear are the mechanisms by which plaque formation is inhibited. Two main possibilities exist: first, that TABLE 4. Hemagglutiniationi inihibitiont by monikey tissue extracts anid serum after different plhysicoclhemical mantipulationts Prepn Dengue Dengue Dengue Dengue 1 2 3 4 Liver extract -7 C, 13 days 37 C, 3 hr. 56 C, 3 min C, 15 min 1 C, 5 min Trypsin Ether. Kaolin. Serum -7 C, 13 days. 5-,m membrane filter.. 2-,um membrane filter. 1 -Am membrane filter....65 Am membrane filter. 33.3%/ supernatantc.. 33.3%76 precipitatec. 5% supernatantc 5% precipitate,.' l1a Euglobulin. Pseudoglobulin. Acetone Kaolin.. ",R-eciprocal- of a fin-al 1,2 1,2 b 5 <1 <1 dilution of liver demonstrating complete inhibition of units of hemagglutinating antigen. b Reciprocal of a final dilution of serum demonstrating complete inhibition of units of hemagglutinating antigen. c Final (NH4)2SO4 concentration used to precipitate serum.

VOL. 3, 1971 TSSUE AND SERUM NHBTORS OF DENGUE VRUSES 199 natural tissue components bind to the virion in such a way as to prevent virus attachment to cells or second that these substances bind to receptor sites on the cell. Pathogenesis studies for dengue virus in monkeys, recently completed in this laboratory, resulted in higher plaque counts for dilute rather than concentrated triturated tissue solutions. t was also noted in that study that the use of cocultivation techniques greatly enhanced the amount of recoverable virus from infected tissue. n the former case, inhibitor is apparently removed by dilution, whereas the latter case precludes contact of virus or cell monolayers with any natural inhibitor of plaque formation present in triturated tissue specimens. ACKNOWLEDGMENTS We thank E. Nishimura and N. Palumbo for their help in various phases of this work. Also, the excellent technical assistance of Ellen Hattis is gratefully acknowledged. LTERATURE CTED 1. Bidwell, D. E., and G. L. Mills. 1. Serum non-specific inhibitors of arbovirus hemagglutination. J. Comp. Pathol. 7:469-479. 2. Chanock, R. M., and A. B. Sabin. 1954. The hemagglutinin of Western equine encephalomyelitis virus: recovery, properties and use for diagnosis. J. mmunol. 73:33-343. 3. Clarke, D. H., and J. Casals. 195. Techniques for hemagglutination and hemagglutination-inhibition with arthropod-borne viruses. Amer. J. Trop. Med. Hyg. 7:561-573. 4. Daneyarov,. A. 15. Content of non-specific hemagglutination inhibitors, induced by arboviruses, in the sera of different animals and techniques for removing them. Nauch. Tr. Virusal. (Transl.) 15, p. 1-5. 5. Gresikova, M., and M. Sekeyova. 17. Non-specific inhibitors of arboviruses in animal sera. Acta Virol. 11:1-113. 6. Halstead, S. B., S. Udomsakdi, P. Simasthien, P. Singharaj, P. Sukhavachana, and A. Nisalak. 197. Observations related to pathogenesis of dengue hemorrhagic fever.. Experience with classifications of dengue viruses. Yale J. Biol. Med. 42:261-275. 7. Hammon, W. M., and T. H. Work. 19. Arbovirus infection in man. hli E. H. Lennette and N. J. Schmidt (ed.), Procedures for viral and rickettsial diseases. American Public Health Association, nc., New York.. Hardy, J. L., W. F. Scherer, and J. B, Carey, Jr. 15. Differential inactivation of arthropod borne animal viruses by bile salts in plasma or serum. Amer. J. Epidemiol. 2:73-4. 9. Mekler, L. B., L. G. Khaltayeva, and S. Y. Gaidamovich. 16. Some properties of non-specific inhibitors of arbovirus hemagglutination. Acta Virol. 1:343-347. 1. Nisalak, A., S. B. Halstead, P. Singharaj, S. Udomsakdi, S. W. Nye, and K. Vinjchaikul. 197. Observations related to pathogenesis of dengue hemorrhagic fever.. Virologic studies of fatal disease. Yale J. Biol. Med. 42:293-31. 11. Porterfield, J. S., and C. E. Rowe. 1. Hemagglutination with arthropod-borne viruses and its inhibition by certain phospholipids. Virology 11:765-77.. Russel, P. K., A. Nisalak, P. Sukhavachanu, and S. Vivona. 17. A plaque reduction neutralization test for dengue virus neutralizing antibodies. J. mmunol. 99:25-29. 13. Salminen, A. 12. Chemistry of non-specific inhibitors of hemagglutination by arthropod-borne viruses. Virology :21-2. 14. Sanderson, C. J. 1. A study of arbovirus non-specific inhibitors and natural aggliitinins in bovine serum. Res. Vet. Sci. 9:4-47. 15. Semenov, B. F., S. P. Chunikin, A. M. Butenko, V. V. Brezin, and J. Woodall. 1. Non-specific inhibitors of hemagglutinating activity of arboviruses in sera of rodents and birds. Vop. Virusol. (Transl.) 13:2-22.. Sunaga, H., R. M. Taylor, and J. R. Henderson. 1. Comparative sensitivity of viruses to treatment with diethyl ether and sodium desoxycholate. Amer. J. Trop. Med. Hyg. 9:419-4. 17. Theiler, M. 1957. Action of sodium desoxycholate on arthropod-borne viruses. Proc. Soc. Exp. Biol. Med. :3-32. 1. Verani, P., and M. Gresikova. 16. Study on non-specific inhibitors of arboviruses in human sera. Acta Virol. 1: 1-5. 19. Work, T. H. 19. solation and identification of arthropod borne virus, p. 3. n E. H. Lennette and N. J. Schmidt (ed.), Procedures for viral and rickettsial diseases. American Public Health Association, nc., New York. 2. Yuill, T. M. 1. Reduction of arbovirus plaque numbers by suspensions of normal mosquitoes. Amer. J. Trop. Med Hyg. 1:69-6l3.