Preincubation of virus with UTIβ or HDL did not result in reduction of infectivity, indicating that they do not irreversibly neutralize viral infectivity. Similarly, preincubation of cells with HDL and UTIβ, unlike interferon, does not induce an antiviral state in the cells. Inhibition experiments at 4 and 37°C revealed that UTIβ inhibits virus by preventing virus attachment. At 4°C, the cell membrane is physiologically inert, and virus replication at 4°C does not proceed beyond initial attachment to the target cell. Comparable titers at 4 and 37°C, therefore, imply that an antiviral substance inhibits the attachment of virus to target cells. Significantly higher titers at 37 than at 4°C imply that the inhibitor acts at a post-attachment stage.

This type of experiment showed that HDL, unlike UTIβ, acts against most viruses at a post-attachment stage. Further investigations of the mechanism of viral inhibition by HDL indicated that HDL inhibits at an early stage of the multiplication cycle, between 0 and 1 h of initiation of infection, most likely by preventing penetration of cell surface by virus (Singh et al., 1999).

A possible defensive role of the narrowly active serum inhibitors has been reviewed by Kriznova and Rathova (1969). The mechanisms of action of these inhibitors are competitive inhibition of virus attachment to cells, and neutralization of viral infectivity by a mechanism unlike classical antibody. The various approaches undertaken to elucidate the possible role of these serum inhibitors in natural resistance were grouped into three categories. In the first category an effort was made to find a correlation between the susceptibility of experimental animals to a particular virus, and the natural inhibitory activity of their sera (Bang et al., 1950, Smorodintsev, 1957, Smorodintsev, 1960). The correlations showed that resistance of the experimental animals to several strains of influenza virus correlated with the presence of innate serum inhibitors against these viral strains.

In the second category, efforts were made to measure the effect of suppression or deletion of serum inhibitors on the susceptibility of the experimental animals to virus infection. In several experiments, it was shown that mice and hamsters fed ethionine (an ethyl analogue of methoinine) had appreciably reduced inhibitory activity in serum against influenza A1 and Newcastle disease viruses, compared to untreated controls. The ethionine-treated hamsters were more susceptible to influenza A1 virus as indicated by higher titers of virus in their lungs (Borecky et al., 1961, Borecky et al., 1962a, Borecky et al., 1962b). These observations point to a relationship in vivo between the level of serum inhibitors and influenza virus multiplication. Thus, they support a defensive role for the narrowly active, non-specific viral inhibitors.

The third category assessed the effect of administering inhibitors at different times during viral infection. Experimental animals were treated with an inhibitor preparation either before or immediately after virus inoculation. Using the gamma serum inhibitor, multiplication of the inhibitor-sensitive strain of influenza A₂ virus in mouse lungs was inhibited. This effect was seen only when the inhibitor was administered at a time before the virus had time to penetrate susceptible cells (Davoli and Bartolomei-Corsi, 1959, Cohen, 1960, Casazza et al., 1964, Link et al., 1964, Link et al., 1965). Thus, gamma inhibitor given intranasally 4–6 h post-infection or later was not protective. However, high doses of inhibitor given i.p., even at longer intervals after infection, showed an appreciable protective effect (Casazza et al., 1964). It appears, overall, that gamma inhibitor can function as a defense mechanism under certain conditions.

The intestinal samples were tested on a variety of size exclusion columns. The antiviral activity consistently emerged in the column volume for most supports, indicating that the active moiety was of very small size. On the Synchropak GPC peptide HPLC columns, on which it is possible to resolve small peptides, the neuramide, porcine mucosal, and mouse intestinal inhibitors all emerged with an apparent molecular size of <1000 Da. This is consistent with the results of dialysis experiments, where the antiviral activity was retained only by a 100 Da MWCO membrane. Based on the GPC Peptide column results, a tentative molecular weight of 600±400 Da was assigned to the antiviral moiety in these preparations.

The antiviral activity of all three gastrointestinal inhibitor preparations was broad in that five virus types were inhibited. The titers of the inhibitors were substantial, ranging from 24 to 64 U/ml. Broad antiviral activity was confirmed using partly purified fractions. Thus, all three gastrointestinal extracts contained significant titers of broadly antiviral substance(s).

To determine the structural composition required for antiviral activity, the extracts were treated to degrade proteins, break carbohydrate linkages, oxidize carbohydrates, or extract lipids and then the residual antiviral activity was measured. Treatment of neuramide, porcine mucosal and mouse intestine preparations, with either proteinase K, carbohydrases, sodium periodate, or extraction by lipid solvents (butanol and ethyl ether), did not affect their antiviral activity. Since the carbohydrases do not degrade all possible sugar linkages, the inhibitors were also treated with NaIO₄ as a general oxidizing agent without inactivation of the antiviral activity. These findings with the intestinal preparations suggest that ordinary protein, carbohydrate, and lipid moieties may not be required for the antiviral activity of GI inhibitors.

To determine whether the intestinal preparations inactivated viruses directly, viruses were pre-treated with the inhibitor and then assayed for residual infectivity. Pretreatment of Sindbis virus, vesicular stomatitis virus and mengovirus with neuramide, porcine mucosal extract, and mouse intestine inhibitors did not reduce their infectivity. This indicated that the inhibitors do not bind irreversibly to virions to permanently neutralize their infectivity.

To determine whether the inhibitors acted during attachment, or later in the replication cycle, the inhibitory titers were compared at 4 and 37°C. The titers of neuramide and porcine mucosal inhibitors against three viruses at 37°C were significantly higher than those at 4°C. This indicates that neuramide and porcine mucosal inhibitors may act by blocking viral replication at a post-attachment stage. The mouse intestine inhibitor, on the other hand, is active at 4°C, and, therefore, inhibits virus replication by preventing attachment of virions to the target cells. These possible mechanisms of action must eventually be confirmed using fully purified inhibitors.

To determine if cells acquired resistance to virus infection after being pre-treated with inhibitor, cells were incubated with the inhibitor preparations (mouse intestine, neuramide and porcine mucosa) for 24 h and then washed before virus challenge. The pre-treated cells did not exhibit any resistance to subsequent infection, indicating the inhibitors do not induce a durable antiviral state in the target cells.

The resistance to degradation by the GI inhibitors could be of significance in vivo, since to remain active in the GI tract the inhibitors must resist the action of the various lipolytic, proteolytic and glycolytic activities of the digestive process. Enzyme-resistant peptides and carbohydrates have been reported previously (Witas et al., 1983, Alpers, 1986, Zijlstra et al., 1996, Day et al., 1998, Podolsky, 2000, Shen and Xu, 2000). Thus, the structure of the GI inhibitors remains undetermined.

Consistent with a possible defensive role for the GI tract inhibitors is their naturally high titer, broad antiviral activity and location in the GI tract and its lumen. To definitely assign defensive roles, future studies should show that (a) administration of these inhibitors is protective in vivo and (b) deletion of the inhibitors enhances infections.

The NS inhibitor in lamb brain tissue exhibited broad antiviral activity against DNA viruses and enveloped and non-enveloped RNA viruses (Singh et al., 1995, Baron et al., 1998). Human, bovine, ovine, porcine, lapine, murine and piscine brains all possessed similar levels of antiviral activity. In humans, this NS inhibitor is present in the brain, spinal cord, white and gray matter, and sciatic nerve, but not in the cerebrospinal fluid. Furthermore, the NS inhibitor can diffuse extracellularly from prerinsed brain tissue in vitro. The antiviral activity was not due to cell toxicity as no difference in Vero and CER cell growth over 48 h was detected in the absence or presence of up to 100 U of the inhibitor. Furthermore, transferrin and low-density lipoprotein, used as non-specific protein controls, were not antiviral.

Using sucrose-density gradient centrifugation (Sullivan et al., 1987, Singh et al., 1995), high levels of antiviral activity were detected in the 40 and 45% sucrose bands (Fig. 3 ). These results indicated that the antiviral factor from brain extracts was at least 4000 kDa.Fig. 3

Sedimentation rate of virus inhibitor from lamb brain tissue in a discontinuous sucrose gradient. Sucrose solutions in PBS, 5 mM MgC1₂, were used to form step gradients. Equivalent volumes of brain inhibitor preparation (0.5 ml) and 50% sucrose were mixed, layered on the discontinuous gradient, and centrifuged at 31 000 rev./min in a SW-55 rotor for 90 min. Collected fractions were examined for antiviral activity against Sindbis virus after extensive dialysis. Reprinted with permission from Antiviral Res., 1995, 27:380.

Thermostability studies showed that the native brain tissue antiviral activity was stable at 100°C for at least 12 h, but not at 120°C. The inhibitory activity was also stable following treatment with 0.01 M DTT, n-butanol and ethyl ether. Viral inhibitory activity of the crude brain extract was abolished by proteolysis, enzymatic glycolysis, or periodate oxidation. This implies that the native inhibitor molecule contains both protein and carbohydrate structure, which are necessary to maintain antiviral activity. However, the lipid-extracted small inhibitor molecule was resistant to proteolysis and glycolysis and was more heat resistant (stable at 120°C for 15 min) than the native inhibitor. Taken together, the evidence indicates that the natural inhibitor moiety in brain, is composed of lipid, protein, and carbohydrate, and disruption of its structure by proteolysis or glycolysis inactivates its inhibitory activity.

Pretreatment of viruses with inhibitor for 2 h did not affect their infectivity, indicating that the inhibitor neither bound irreversibly to virions nor did it permanently neutralize infectivity. To determine whether the inhibitor acted during attachment or later in the replication cycle, inhibitory titers at 4 and 37°C were compared. In additional experiments, virus yields were determined after timed addition of an inhibitor preparation to cell monolayers. Taken together, the results indicated inhibition at the attachment stage of the virus multiplication.

Pretreatment of cells did not induce any resistance to subsequent infection with viruses, indicating that unlike IFN, the inhibitory material does not induce a durable antiviral state in the target cells.

The possible regulation of the NS antiviral activity levels in neural tissues in response to an ongoing infection was assessed using an encephalitic flavivirus infection of mice. Inhibitor titers were determined over 7 days of infection with the encephalitic Banzi virus (Singh et al., 1995). This timing was chosen since BZV can be detected in mouse brain from day 3 p.i. (Singh et al., 1989). No change in inhibitory titer was seen throughout this lethal infection (Fig. 4 ), indicating that the inhibitor, unlike IFN or antibody, is not specifically induced or decreased in response to viral infection.Fig. 4

Level of virus inhibitor in mouse brain extracts is unchanged during encephalitis. Three-week-old female outbred (ICR) mice were infected intraperitoneally with 3 LD₇₅ of BZV. Brains were harvested at days 3–6 postinfection, and homogenized. Virus inhibitory levels in the homogenates were titered against Sindbis virus, and are depicted in the bar graph (left Y axis). The rise in the virus titers in the brains and sera of infected mice (taken from Singh et al., 1989) are superimposed to illustrate the course of the BZV infections (○, virus load in brain; •, virus load per ml serum). Reprinted with permission from Antiviral Res., 1995, 27:384.

The possibility that this constitutive antiviral inhibitor in the nervous system could affect pathogenesis was studied by determining (Baron et al., 1998) whether the inhibitor was: (a) widely distributed in the human nervous system and released extracellularly, (b) present in effective concentrations and (c) protective in vivo (see below).

Tissue extracts from invertebrate crustaceans, crab, shrimp and crayfish are broadly antiviral against a variety of viruses including DNA and enveloped and non-enveloped RNA viruses. Characterization of the inhibitor(s) indicated a 440-kDa molecular size for the main inhibitor. The antiviral activity of the inhibitor preparation was reduced by lipid extraction with butanol, but was unaffected by treatments with DTT, urea, proteinase K and carbohydrases. The inhibitor preparation does not directly neutralize virus infectivity. The inhibitor appears to act during the virus attachment stage. In addition, the crab inhibitor preparation also induces a durable antiviral state in Vero cells against subsequent infection of Sindbis and vaccinia viruses that was not due to interferon. The non-specific, broadly antiviral substances that are present in crustaceans may be host defense substances and may contribute to survival of crustaceans, despite continuing exposure to extremely high dose of viruses (Fuhrman, 1999).

Viruses are the most common biological agents in the sea. They number ten billion per liter and probably infect many species (Fuhrman, 1999). Viral infections are common in crustaceans. For instance, penaeid shrimp are infected by at least 20 viruses (Johnson et al., 1989, Johnson et al., 1999), while blue crabs are infected naturally or experimentally by rotavirus, enterovirus, Newcastle disease virus, and poliovirus (McCumber and Clem, 1977, Seidel et al., 1983). In comparison with vertebrates, invertebrates do not produce specific antibodies (Schapiro, 1975), and therefore may rely on innate defense for protection against viruses (Schnapp et al., 1996). Little is known about possible innate antiviral substances in invertebrates. Thus, crustaceans, such as crabs, shrimp and crayfish, are appropriate invertebrate subjects to study innate immunity against viruses (Pan et al., 2000).

The inhibitor extract from crabs is broadly inhibitory for the six types of DNA and RNA viruses tested. The titers of the inhibitor varied moderately with the different viruses. The inhibitory activity of tissue extracts from two other crustaceans, shrimp and crawfish, were also found to be broadly active against the same viruses with titers comparable to that of crab inhibitor preparations. No inhibitory effect of the inhibitor preparation was detected on either cell-free HIV or HIV-infected MT-2 cells.

The crab inhibitor preparation was fractionated by size-exclusion chromatography on a Superdex 200 26/60 FPLC column, and individual fractions were tested for the presence of antiviral activity. The antiviral peak (44 U/ml) corresponded to a molecular mass of approximately 440 kDa. In addition, a secondary peak of relatively low antiviral activity (17 U/ml) was found in fractions corresponding to a molecular size of about 15 kDa.

Lipid extraction with butanol significantly decreased the antiviral activity in the residual aqueous phase of the inhibitor preparation. Lipid extraction also resulted in the reduction of the size of the antiviral substance as evidenced by the filterability of the residual antiviral activity through a 10 K MWCO ultrafiltration membrane (Filtron, Northborough, MA). In contrast, neither insoluble proteinase K, nor a cocktail of glycosidases, significantly inactivated the antiviral activity of the inhibitor preparation. In addition, a sulfhydryl reagent (DTT) and a chaotropic reagent (urea) were used to treat the inhibitor preparation, and no loss of antiviral activity was found. Thus, the inhibitor is resistant to many chemical degradations.

The crab inhibitor preparation did not irreversibly inactivate viruses. To determine whether the inhibitor preparation acts during virus attachment or thereafter, antiviral activity at 4 and 37°C was compared, and the equal activity at 4 and 37°C indicated that one action of the crab inhibitor is to prevent virus attachment to target cells.

To determine whether the inhibitor preparation also induced durable antiviral activity in host cells, the inhibitor preparation was preincubated with Vero cells overnight and then washed before virus challenge. The inhibitor preparation-pre-treated Vero cells were resistant to subsequent infections by Sindbis and vaccinia viruses, indicating that the crab inhibitor preparation also induces a durable antiviral state in Vero cells against subsequent infection by Sindbis and vaccinia viruses. In contrast, vesicular stomatitis virus was not inhibited by the same pretreatment of cells, indicating a reversible and probably different mechanism of inhibition and the possibility of two different inhibitors.

The oral cavity is unique among mucosal surfaces in having a defensive barrier to transmission of HIV and probably to the related human T-cell leukemia virus (Mohor, 1973, Phillips and Bourinbaiar, 1992, Milman and Sharma, 1994, Bomsel, 1997, Zacharopoulos and Phillips, 1997). The main defense mechanism appears to be inactivation of the transmitting leukocytes by the hypotonicity of saliva. Additional protection may be contributed by the other less active salivary inhibitors (Coppenhaver et al., 1984, Fultz, 1986, Fox et al., 1988, Fox et al., 1989, Archibald and Cole, 1990, Yeh et al., 1992, Bergey et al., 1994, McNeely et al., 1995, McNeely et al., 1997, Wahl and Orenstein, 1997, Crombie et al., 1998). This salivary defense appears to explain the rarity of casual HIV transmission orally. However, the normally effective oral barrier can be overcome by the solutes in seminal fluid, milk, and colostrums (Mayer and DeGruttola, 1987, Lifson et al., 1990, Quatro et al., 1990, Lane et al., 1991, Ryder et al., 1991, Dunn et al., 1992, Keet et al., 1992, Bertolli et al., 1996, Black, 1996, Berrey and Shea, 1997, Cao et al., 1997).

Of special importance for medical application is that the salivary defense has directed attention to the HIV-infected leukocyte as the main mucosal transmitting agent. Consequently, anticellular solutions that are more potent than saliva's hypotonicity could be applied topically at the vulnerable mucosal sites (e.g. vagina and rectum) to inactivate infected leukocytes in seminal fluid. To prevent the vaginal and rectal transmission, studies of candidate topical anticellular substances (Baron, 2000), other than the reportedly ineffective nonoxynol-9 (Mohor, 1973, Pauza et al., 1993, Pauza et al., 1994), are being done. These candidate preventives include disinfectants for blood products, bile detergents, and over-the-counter (OTC) commercial vaginal products, which are effective in vitro. Some of these OTC products are already established to be safe for use by FDA standards.

When the HIV epidemic began, an initial concern was that HIV might be transmitted casually from the oral secretions of the millions of HIV carriers during kissing, dental treatment, biting, and aerosolization. Later it became clear that casual transmission is actually a rare event, even when infectious HIV is shed into the oral cavity by bleeding and exudation (Ho et al., 1985, Levy and Greenspan, 1988, Rogers et al., 1990, Gooch et al., 1993, Moore et al., 1993, Richman and Rickman, 1993, Yeung et al., 1993, Piazza et al., 1994). Saliva of infected individuals usually contains only non-infectious components of HIV, including viral DNA, indicating that the saliva may cause a breakdown of infected leukocytes. The infectivity of these salivas in vitro is only 1%, although all carriers have infectious virus in their blood (Goto et al., 1991, Barr et al., 1992, Moore et al., 1993, Coppenhaver et al., 1994, Qureshi et al., 1995, Melvin et al., 1997, Baron et al., 1999). In comparison, in seminal fluid the percentage of patients shedding HIV, mainly as infected leukocytes, is much higher — approximately 20% (Milman and Sharma, 1994, Vernazza et al., 1996, Zhu et al., 1996). Thus the rarity of casual oral transmission correlates with the low infectivity of saliva compared to the high infectivity of seminal fluid, which is associated with 27% vaginal transmission (European Study Group, 1989, Johnson et al., 1989, Lazzarin et al., 1991, European Study Group on Heterosexual Transmission of HIV, 1992, Royce et al., 1997). It is, therefore, important to determine the salivary mechanisms of protection.

There are many inhibitors of HIV in saliva (Coppenhaver et al., 1984, Fultz, 1986, Fox et al., 1988, Fox et al., 1989, Archibald and Cole, 1990, Yeh et al., 1992, Bergey et al., 1994, McNeely et al., 1995, McNeely et al., 1997, Wahl and Orenstein, 1997, Crombie et al., 1998, Shugars, 1999, Baron et al., 1999, Baron et al., 2000). An important question is ‘which of the many reported salivary inhibitors of HIV best accounts for saliva's protection against HIV-infected leukocytes?’ These salivary inhibitors include HIV-1-specific antibodies, lysozyme, peroxidases, complement, cystatins, lactoferrin, defensins, mucins, amylase, statherin, proline-rich peptides, thrombospondin-1, histatins, Rnases (Lee-Huang et al., 1999) and secretory leukocyte protease inhibitor (Coppenhaver et al., 1984, Fultz, 1986, Fox et al., 1988, Fox et al., 1989, Archibald and Cole, 1990, Sun et al., 1990, Yeh et al., 1992, Robinovitch et al., 1993, Bergey et al., 1994, McNeely et al., 1995, McNeely et al., 1997, Wahl et al., 1997, Shugars, 1999).

The existence of an undiscovered salivary inhibitor was suggested because these known salivary inhibitors did not inhibit HIV sufficiently to explain the rarity of oral transmission (Coppenhaver et al., 1984, Fultz, 1986, Fox et al., 1988, Fox et al., 1989, Archibald and Cole, 1990, Sun et al., 1990, Yeh et al., 1992, Robinovitch et al., 1993, Bergey et al., 1994, McNeely et al., 1995, McNeely et al., 1997, Wahl et al., 1997, Shugars, 1999). A new inhibitor also was suggested by the unique hypotonicity of salivas, i.e. one-seventh the tonicity of normal interstitial fluids (Ben-Aryeh et al., 1990, Edgar, 1992). This hypotonicity was shown to disrupt the crucial infected leukocytes and thereby render them incapable of supporting virus multiplication and cell-to-cell transmission of HIV (Fig. 5 ). Specifically, there was a 10 000-fold inhibition of the multiplication of HIV by the hypotonicity of saliva compared to the only 3- to 5-fold inhibition by the macromolecular inhibitors in saliva (Baron et al., 1999, Baron et al., 2000). Also, reconstitution of the isotonicity of saliva reversed its inhibition (Baron et al., 1999). Thus, hypotonic disruption appears to be a major mechanism by which saliva kills the transmitting mononuclear leukocytes and thereby prevents their attachment to mucosal epithelial cells and their production of infectious HIV.Fig. 5

Saliva interrupts the multiplication of human immunodeficiency virus (HIV) in infected human mononuclear leukocytes. *P<0.05 by Student's t-test. Reprinted with permission from Arch. Intern. Med., 1999, 159:306.