Involvement of Interleukin-2 and Interferon-Gamma in the Immune Response Induced by Influenza Virus Iscoms

Splenocytes from mice primed with influenza virus envelope proteins incorporated in iscoms, as micelles or as infectious virus, were restimulated in vitro with the same antigen. Interleukin-2 (IL-2) and interferon-gamma (IFN-gamma) were assayed in the supernatants of such cultures. Influenza virus iscoms induced IL-2 and IFN-gamma responses in restimulation experiments that were antigen specific and significantly higher than those induced by micelles or infectious virus. Serum samples collected at the end of the experiments were analysed for the antibody response and profile. The antibody titres induced by iscoms were of a similar order of magnitude as those induced by infectious virus, and were about 18 times higher than the titres induced by micelles. In mice immunized with iscoms or infectious virus the most abundant antibodies were of the IgG1 and IgG2a isotype, and the IgE response was low. We conclude that immunization with iscoms stimulates the Th1-like subtype of murine T lymphocytes.     

Antigen-Specific Increases in the Number of Splenocytes Expressing MHC Class II Molecules Following Restimulation with Antigen in Various Physical Forms

To understand how a presentation system for antigens initiates an immune response and why it has a strong adjuvant activity, a number of parameters need to be analysed. In this study the frequency of spleen cells expressing MHC class II (Ia antigen) was determined after immunization of mice and restimulation of their spleen cells, in vitro, with influenza virus envelope proteins in different physical forms, namely iscoms, micelles and virus particles. All three forms of the antigen stimulated, in an antigen-specific manner, an increased proportion of spleen cells expressing MHC class II in the restimulation experiments. The induction of increased MHC class II expression was at least partly dependent on antigen-specific induction of IFN-gamma since an antibody to IFN-gamma partly inhibited the increase of MHC class II+ cells induced by iscom or by Concanavalin A. The iscom-borne antigens were superior to micelles to prime the immune response in vitro, indicating a capacity to induce memory cells. This primed immune response was readily recalled in vitro, as measured by IFN-gamma production and an increased number of MHC class II positive cells.     

Optimizing oral vaccines: induction of systemic and mucosal B-cell and antibody responses to tetanus toxoid by use of cholera toxin as an adjuvant.

Cholera toxin (CT) is an effective mucosal antigen and acts as an adjuvant when given orally with various antigens; however, few studies have compared the levels of antibody responses to CT and coadministered protein in systemic and mucosal tissues. In this study, we used tetanus toxoid (TT) for assessment of immune responses. Time course and dose-response studies established that 250 micrograms of TT given orally with 10 micrograms of CT three times at weekly intervals induced high serum and gastrointestinal tract anti-TT and anti-CT antibody responses. Oral immunization with TT alone induced no detectable mucosal immunoglobulin A (IgA) antibodies in fecal extracts and only weak serum IgG anti-TT responses. The coadministration of CT and TT induced peak serum IgG anti-TT responses following two oral doses that remained constant after the third oral immunization, while optimal mucosal IgA responses were seen after the third oral immunization. The serum anti-TT response obtained with CT and TT proved protective against TT challenge (100 minimum lethal doses), whereas mice orally given CT or TT alone died. Antigen-specific B-cell responses were assessed with an isotype-specific Elispot assay of isolated lymphoid cells from the spleen, Peyer's patches, and the small intestinal lamina propria. Interestingly, approximately fourfold-higher numbers of IgA anti-CT than of anti-TT antibody-producing (spot-forming) cells occurred in lymphocytes from the lamina propria of mice orally immunized with both TT and CT. The adjuvant CT did not induce polyclonal B-cell responses in mice given CT by the oral route, since no significant differences in total numbers of B cells producing IgA, IgG, or IgM were found compared with the numbers in mice given TT alone. The results clearly indicate that serum and mucosal antibody responses develop with different kinetics and that protective TT-specific antibody responses are generated in the systemic compartment when TT is administered with CT via the oral route.     

Lipid vesicle-entrapped influenza A antigen modulates the influenza A-specific human antibody response in immune reconstituted SCID-human mice

This study was designed to investigate the capacity of purified influenza antigen in the presence and absence of adjuvant to induce human antibody responses in human-PBL-SCID mice. Non-ionic surfactant vesicles (NISV) were used as adjuvant as they have been shown to promote the development of Th1 responses in mouse studies. Human peripheral blood lymphocyte-SCID mice were inoculated with either purified influenza antigen (A/Texas, H3N2) or influenza antigen entrapped in NISV. Both vaccinated groups produced significantly higher plasma levels of influenza-specific human IgG when individually compared with non-vaccinated controls. However, similar comparisons revealed that specific IgM levels were significantly higher only in the group challenged with purified antigen. Further analysis of IgG subclasses also demonstrated an adjuvant-dependent dichotomy in the responses of the vaccine groups when compared with non-vaccinated controls. Thus, only influenza-specific IgG1 antibodies (associated with Th1 responses in humans) were significantly increased above control levels using antigen with adjuvant, while both this subclass and antigen-specific IgG4 (Th2 associated) were significantly increased with antigen alone. These results illustrate the suitability of this model for use in human vaccination studies and demonstrates that influenza antigen applied with NISV selectively promotes only Th1 responses, unlike free antigen which also promotes Th2 responses in vivo.     

A synthetic multi-epitope antigen enhances hepatitis C virus-specific B- and T-cell responses

Combining results from previous studies, a multi-epitope antigen PCXZ against the hepatitis C virus was synthesized in this study. The antigenic specificity of PCXZ was determined by recognizing antibodies in serum samples from hepatitis C virus patients, but not from healthy subjects or subjects who had the hepatitis B virus. The characteristics of PCXZ immunogenicity were evaluated in BALB/c mice. Strong antibody responses were generated in mice immunized with either naked PCXZ or PCXZ in Freund's adjuvant. As for the T-cell responses, Freund's adjuvant significantly increased interferon-γ secretion and enhanced the lytic activity of cytotoxic T lymphocytes. The epitope Pa, one component of PCXZ, made the most significant contribution to specific CTL lysis; this epitope was also a B-cell epitope and was able to induce high IgG titers. In summary, PCXZ was found to be highly immunogenic, and elicited both humoral and cellular immune responses in mice.     

Improved systems for hydrophobic tagging of recombinant immunogens for efficient iscom incorporation

We have previously reported a strategy for production in Escherichia coli of recombinant immunogens fused to a hydrophobic tag to improve their capacity to associate with an adjuvant formulation [Andersson et al., J. Immunol. Methods 222 (1999) 171]. Here, we describe a further development of the previous strategy and present significant improvements. In the novel system, the target immunogen is produced with an N-terminal affinity tag suitable for affinity purification, and a C-terminal hydrophobic tag, which should enable association through hydrophobic interactions of the immunogen with an adjuvant system, here being immunostimulating complexes (iscoms). Two different hydrophobic tags were evaluated: (i) a tag denoted M, derived from the membrane-spanning region of Staphylococcus aureus protein A (SpA), and (ii) a tag denoted MI consisting of the transmembrane region of hemagglutinin from influenza A virus. Furthermore, two alternative affinity tags were evaluated; the serum albumin-binding protein ABP, derived from streptococcal protein G, and the divalent IgG-binding ZZ-domains derived from SpA. A malaria peptide M5, derived from the central repeat region of the Plasmodium falciparum blood-stage antigen Pf155/RESA, served as model immunogen in this study. Four different fusion proteins, ABP-M5-M, ABP-M5-MI, ZZ-M5-M and ZZ-M5-MI, were thus produced, affinity purified and evaluated in iscom-incorporation experiments. All of the fusion proteins were found in the iscom fractions in analytical ultracentrifugation, indicating iscom incorporation. This was further supported by electron microscopy analysis showing that iscoms were formed. In addition, these iscom preparations were demonstrated to induce M5-specific antibody responses upon immunisation of mice, confirming the successful incorporation into iscoms. The novel system for hydrophobic tagging of immunogens, with optional affinity and hydrophobic tags, gave expression levels that were increased ten to fifty-fold, as compared to the earlier reported system. We believe that the presented strategy would be a convenient way to achieve efficient adjuvant association for recombinant immunogens.     

The iscom: an immunostimulating system

To make purified antigens highly immunogenic, they have to be presented in several copies in the form of a microscopic or submicroscopic particle. This is the case, regardless of whether the antigens are obtained by isolation from conventional microorganisms, or from gene-manipulated cells, or synthesized. In the iscom, the antigens are attached as multimers to a 40-nm cage-like particle with a built-in adjuvant. The antigens in iscoms are rapidly transported from the injection site to the draining lymphatic organ. Iscom-borne antigens induced a 10-fold higher antibody response than the same amount of antigen in micelle form. One intranasal immunization with influenza virus iscoms induced protection to intranasal challenge infection in mice. Besides a strong antibody response in all Ig classes and isotypes, cytotoxic T cells were induced. With iscoms containing gp160 of HIV-1, cytotoxic T cells (CD8+ CD4-) were induced under restriction of class I MHC antigen. Iscoms containing the fusion protein of measles virus induced T cell clones in mice whereof one, after adoptive transfer, protected mice against intracerebral challenge infection. Protective immunity against Epstein-Barr virus (EBV)-induced tumor formation by iscoms containing gp350 of EBV has been elicited in cotton-top Tamerin monkeys. Protective immunity has also been induced against several virus infections including feline leukemia virus and against parasites, i.e., Trypanosoma cruzi, in mice.     

Protein and modified vaccinia virus Ankara‐based influenza virus nucleoprotein vaccines are differentially immunogenic in BALB/c mice

Because of the high variability of seasonal influenza viruses and the eminent threat of influenza viruses with pandemic potential, there is great interest in the development of vaccines that induce broadly protective immunity. Most probably, broadly protective influenza vaccines are based on conserved proteins, such as nucleoprotein (NP). NP is a vaccine target of interest as it has been shown to induce cross‐reactive antibody and T cell responses. Here we tested and compared various NP‐based vaccine preparations for their capacity to induce humoral and cellular immune responses to influenza virus NP. The immunogenicity of protein‐based vaccine preparations with Matrix‐M? adjuvant as well as recombinant viral vaccine vector modified Vaccinia virus Ankara (MVA) expressing the influenza virus NP gene, with or without modifications that aim at optimization of CD8⁺ T cell responses, was addressed in BALB/c mice. Addition of Matrix‐M? adjuvant to NP wild‐type protein‐based vaccines significantly improved T cell responses. Furthermore, recombinant MVA expressing the influenza virus NP induced strong antibody and CD8⁺ T cell responses, which could not be improved further by modifications of NP to increase antigen processing and presentation.     

Characterization of immunostimulating complexes (ISCOMS) of HIV-1

HIV-1, strain HTLV-III, propagated in H9 cells and purified by sucrose gradient centrifugation, was used as native antigen source for the preparation of immunostimulating complexes, HIV-iscoms. The major antigen detected in the iscom was the cell-derived HLA-DR, which readily could be removed from the virus lysate by immunosorbent. In the iscoms the HIV structural proteins MA p17, p55 and TM gp41 were identified; SU gp120 was present in only minute amounts in the virus lysate. The iscom particles appeared well preserved after freeze drying with a round shape, approximately 35 nm in diameter, comprising morphological subunits, assembled with icosahedral symmetry. Immunization experiments in mice reflected the antigen content of the iscoms. High antibody response was induced to HLA-DR in non-depleted iscoms. Major humoral responses were observed to the viral structural proteins MA p17, CA p24, p55, and also to TM gp41. A low or negligible antibody response to SU gp120 was induced by the HIV-iscoms. The negligible response was, however, overcome by the addition of recombinant gp160 to the virus lysate prior to formation of iscoms, resulting in a preparation evoking a clear serum antibody to gp160.     

IgG isotype distribution of local and systemic immune responses induced by influenza virus infection

The IgG isotype profile of the influenza virus-specific immune response was studied by quantitation of serum antibody (Ab) levels in correlation with the enumeration of antibody-secreting cells (ASC) detected in the lung, spleen, mediastinal lymph nodes (MLN), Peyer's patches and bone marrow (BM). Distincst isotypic patterns for serum Ab and Ab produced by cells present at or close to the site of infection were found after primary or repeated infections. An elevated number of IgM ASC was found after primary challenge in the spleen, lung and MLN. In contrast, the site of IgA and IgG production is restricted to the lung and lymph nodes draining the site of infection. In these organs IgA, IgG2a and IgG1 ASC are found as a result of primary virus infection while viral challenge induces mostly activation of IgA-producing cells and secretion of IgA to the lung lavage. In contrast, the majority (80-90%) of Ab detected in the serum belong to the IgG2a subclass and their serum level is maintained at a high level during the whole period of the response. The relative level of virus-specific serum IgG2a in correlation with the production of IgG2a Ab found predominantly in MLN and lung is highly dependent on the viral dose used for priming or challenge. As IgG2a ASC can be detected at relatively low numbers in the spleen and BM these results suggest that the production of the dominant IgG2a isotype of serum Ab occurs close to the viral challenge site. These data, however, point to distinct isotypic regulation in systemic versus local virus-specific Ab responses.     

Immune response after adjuvant mucosal immunization of mice with inactivated influenza virus

Satisfactory mucosal immunity in the respiratory tract is very important for protection against influenza. It can be achieved only by mucosal immunization. Mucosal vaccination with inactivated influenza virus may not be sufficiently effective and suitable adjuvants are therefore sought. We tested intratracheal immunization of mice with inactivate B type influenza virus in a mixture with formolized G+ bacterium Bacillus firmus, whose adjuvant effects have previously been documented in another system. The treatment resulted in a marked increase of both systemic and mucosal antibody response in IgG and IgA classes. Stimulation of T lymphocytes after adjuvant immunization was very mild, no proliferation taking place after specific stimulation with antigen in vitro. However, slightly increased systemic (spleen) and local (lungs) production of cytokines without perceptible Th1/Th2 polarization was determined. B. firmus is an efficient adjuvant in respiratory tract immunization while with subcutaneous immunization it lowers the antibody response.     

Th1 adjuvant N-acetyl-D-glucosamine polymer up-regulates Th1 immunity but down-regulates Th2 immunity against a mycobacterial protein (MPB-59) in interleukin-10-knockout and wild-type mice.

Treatment of mice with heat-killed (HK) Mycobacterium bovis BCG or 1- to 10-microm chitin particles (nonantigenic N-acetyl-D-glucosamine polymers) is known to induce innate immune responses, including gamma interferon (IFN-gamma) production, which plays a Th1 adjuvant role. However, HK BCG further induces prostaglandin E2-releasing spleen macrophages (Mphi) (PGE2-Mphi), which potentially inhibit Th1 adjuvant activities. We found that chitin particles did not induce PGE2-Mphi formation. To further assess whether chitin has Th1 adjuvant effects, interleukin-10 (IL-10)-knockout (KO) mice and their wild-type (WT, C57BL/6) controls were immunized with a 30-kDa MPB-59 mycobacterial protein mixed with chitin. Immunization with MPB-59 alone induced Th2 responses, characterized by increases in total serum immunoglobulin E (IgE) and specific serum IgG1 levels and spleen Th2 cells producing IL-4, IL-5, and IL-10. No IFN-gamma-producing spleen Th1 cells, specific serum IgG2a, or delayed-type hypersensitivity (DTH) footpad reactions were detected. On the other hand, chitin-MPB-59 immunization significantly increased spleen Th1 responses, DTH reaction, and serum IgG2a levels along with decreases of Th2 responses. The magnitude of these Th1 adjuvant effects was greater in IL-10-KO mice than in WT mice. In contrast, immunization with HK BCG-MPB-59 showed little or no Th1 adjuvant effect. These data indicate that chitin has a unique Th1 adjuvant effect on the development of Th1 immunity against a mycobacterial antigen. IL-10 down-regulates the adjuvant effect of chitin.     

General expression vectors for production of hydrophobically tagged immunogens for direct iscom incorporation

A new general strategy for the production of recombinant protein immunogens has been investigated. The rationale involves the production of a recombinant immunogen as fused to a composite tag comprising one domain suitable for affinity purification and a hydrophobic tag designed for direct incorporation through hydrophobic interaction of the affinity-purified immunogen into an adjuvant system, in this case immunostimulating complexes (iscoms). Three different hydrophobic tags were evaluated: (i) a tag denoted IW containing stretches of hydrophobic isoleucine (I) and tryptophan (W) residues; (ii) a tag denoted MI consisting of the transmembrane region of hemagglutinin from influenza A virus; and (iii) a tag denoted PD designed to be pH-dependent in such a way that an amphiphatic alpha-helix would be formed at low pH. As an affinity tag, an IgG-binding domain Z derived from Staphylococcus aureus protein A (SpA) was used, and a malaria peptide M5, derived from the central repeat region of the Plasmodium falciparum blood-stage antigen Pf155/RESA, served as a model immunogen in this study. Three different fusion proteins, IW-Z-M5, MI-Z-M5 and PD-Z-M5, were produced in Escherichia coli, and after affinity purification these were evaluated in iscom-incorporation experiments. Two of the fusion proteins, IW-Z-M5 and MI-Z-M5 were found in the iscom fraction following preparative ultracentrifugation, indicating iscom incorporation. This was further supported by electron microscopy analysis showing that iscoms were formed. Furthermore, these iscom preparations were demonstrated to induce efficient M5-specific antibody responses upon immunization of mice, confirming successful incorporation into iscoms. The implications of these results for the design and production of subunit vaccines are discussed.     

Anti-DNP antibody responses to DNP-histone H1 in mice.

Anti-DNP antibody responses to DNP-histone H1 in C3H/HeN mice were not suppressed by in vivo treatment with carrageenan in which many phagocytic macrophages were presumed to be impaired. Rather higher antibody responses to this antigen were observed in athymic nude mice than in heterozygous nude mice. Further, non-adherent spleen-cell and T-cell depleted spleen cells induced in vitro anti-DNP antibody responses to DNP-histone H1 to the same extent as normal spleen cells. These results suggest that anti-DNP antibody responses to DNP-histone H1 are macrophage- and thymus-independent. It was also observed that IgG-type anti-DNP antibodies to DNP-histone H1 were produced although most thymus-independent antigens were shown to induce predominantly IgM type antibodies but little, if any, IgG type antibodies. Furthermore, histone H1 did not show any polyclonal B-cell activator activities in contrast to many other thymus-independent antigens which act as polyclonal B-cell activators.     

Antibody response in the female rabbit reproductive tract to influenza haemagglutinin encoded by a recombinant myxoma virus

The antibody response in serum and the reproductive tract of female rabbits to a model antigen, influenza virus haemagglutinin (HA), encoded by a recombinant myxoma virus was investigated. Strong and lasting IgG antibody responses to HA were induced in serum following intradermal, intranasal, and intravaginal immunisations. HA IgG was also detected in reproductive tract fluids but was only about 1% the titer of that in serum. HA IgA was not detected in serum of any infected groups and was occasionally detected in reproductive tract fluids at a low titer only after infections through mucosal sites. HA IgM was also detected only in some of the reproductive tract fluids at very low levels. Induction of ovulation did not change these patterns and B cell homing to the reproductive tract was not profound. In contrast, HA IgG and IgM titers in ovarian follicular fluids were comparable to that in serum. These data suggest that if this virus is used to deliver an immunocontraceptive vaccine that requires a high-level antibody response, the target antigen needs to be accessible to serum antibody or in the ovary.     

Antibody Response of Hamsters to A2/Hong Kong Virus Vaccine after Priming by Heterotypic Virus Infection

Hamsters previously infected with influenza virus A1/FM/1/47 produced serum hemagglutination inhibition (HI) antibody in response to 1/100 the antigenic dose of inactivated influenza virus A2/Hong Kong vaccine necessary to induce antibody in normal animals. This priming effect was believed to be due to the virus infection which caused an immune response to a virus antigen common to both the infecting virus and the virus vaccine; this antigen acted as a carrier for the specific vaccine virus hemagglutinin and potentiated the immune response to the new antigen. This theory, which has been established in other immune systems, was tested, and the results obtained did not contradict the conditions imposed in the above explanation. Thus, the priming effect could be transferred to normal hamsters by inoculation of spleen cells from virus-infected animals, and the HI antibody response to the virus vaccine was characteristic of a secondary response. The theory also required that the new antigen be coupled to the carrier protein; however, primed hamsters produced serum HI antibody after inoculation with ether-Tween-split virus vaccine, but there was no proof that this vaccine was completely dissociated.     

Regional T- and B-cell responses in influenza-infected ferrets.

Ferrets were infected with A/Port Chalmers/72 influenza virus and the T- and B-cell responses in the spleen, in lymph nodes draining the upper and lower respiratory tract, and in lung washings were examined in vitro. Lymphocyte responses were measured by using a hemolytic plaque assay for B cells and a proliferation assay for T cells. Virus and antibody levels were measured in respiratory tract washings, and antibody titers were measured in sera from infected animals. Individual B cells secreting specific antibody to A/Port Chalmers/72 virus were detected in regional lymph node and spleen preparations as early as 3 days and as late as 43 days after infection. T-cell assays showed an in vitro response of lymph node cells to A/Port Chalmers/73 virus from day 6 to day 43. Virus was isolated from the respiratory tract up to 7 days after infection. Serum hemagglutination-inhibiting antibody was first detectable on day 6, with maximum titers reached by day 10. These results demonstrated that antibody production and a cellular immune responses were detectable at regional sites at a time when virus was still present and before serum antibody was measured.     

Mucosal immunogenicity of a recombinant Salmonella typhimurium-cloned heterologous antigen in the absence or presence of coexpressed cholera toxin A2 and B subunits.

An avirulent Salmonella typhimurium vaccine strain expressing a streptococcal protein adhesin and a similar clone which produces the same streptococcal antigen linked to the cholera toxin (CT) A2 and B subunits (CTA2/B) were compared for the ability to induce antibody responses to the expressed heterologous antigen after oral or intranasal immunization of mice. Expression of cloned immunogens in these systems is temperature regulated, being optimal at 37 degrees C, and the two clones under comparison were shown to produce similar levels of the streptococcal antigen. Both clones were found to stimulate high levels of serum immunoglobulin G (IgG) and mucosal IgA antibodies to the cloned immunogen. A consistent trend was observed toward higher mucosal IgA but lower serum IgG responses in the case of the S. typhimurium vector that coexpressed CTA2/B, a potential mucosal adjuvant, regardless of the route of administration. Also noteworthy was the capacity of these antigen delivery systems to induce anamnestic systemic and secretory responses to the cloned immunogen 15 weeks after the primary immunization, despite preexisting immunity to the Salmonella vectors. These antibody responses were sustained for at least 7 months following the booster immunization, at which time the secretory IgA antibody levels were significantly higher in mice given the Salmonella clone that coexpressed CTA2/B. Although the serum IgG response against the Salmonella vector was characterized by a high IgG2a/IgG1 ratio (indicative of the T helper type 1 [Th1]/Th2 profile), a mixed IgG1 and IgG2a pattern was observed for the carried heterologous antigen, which displayed a dominant IgG1 response when administered as a purified immunogen. Our findings indicate that the recombinant streptococcal antigen and CTA2/B are strong immunogens when expressed by the antigen delivery system used in this study and suggest that CTA2/B may have an additional immunoenhancing activity in the mucosal compartment besides its ability to target antigen uptake into the mucosal inductive sites. CTA2/B may thus be useful as an S. typhimurium-cloned adjuvant for coexpressed protein antigens.     

The immune response to influenza vaccines

Specific immunity to influenza is associated with a systemic immune response (serum haemagglutination inhibition antibody), local respiratory immune response (virus-specific local IgA and IgG antibodies in nasal wash), and with the cell-mediated immune response. Both inactivated and live influenza vaccines induce virus-specific serum antibody which can protect against infection with influenza virus possessing the same antigenic specificity. In the absence of serum antibodies, local antibodies in nasal wash are a major determinant of resistance to infection with influenza virus. In comparative studies in humans it was shown that nasal secretory IgA develops chiefly after immunization with live cold-adapted (CA) vaccine, but persistent nasal secretory IgG was detected in both CA live and inactivated vaccines. The origin of nasal wash haemagglutination inhibition (HI) antibodies is not completely known. Recently it was found that cytotoxic T-cells (CTL) play an important role in immunity against influenza and in clearance of influenza virus from the body. In primed humans, inactivated influenza vaccine stimulates a cross-reactive T-cell response, whereas the ability of inactivated vaccine to stimulate such immunity in unprimed humans has not been determined. Data on the T-cell response to live vaccine in humans are limited to the development of secondary T-cell responses in primed individuals vaccinated with a host-range (HR) attenuated vaccine. The data obtained have shown that immunity induced by inactivated influenza vaccines is presumably dependent on the stimulation of serum antibody. Live CA vaccines not only stimulate a durable serum antibody response, but also induce long-lasting local respiratory tract IgA antibody that plays an important role in host protection.     

Monocyte and iscom enhancement of cell-mediated response to cytomegalovirus

Cell-mediated and humoral responses to cytomegalovirus were studied in a monkey model. Repeated low doses of virus antigen gave poor reactivities in both respects. High antigen doses gave a good humoral IgG response. When autologous monocytes were incubated with the CMV antigen as the immunizing injection, the specific cellular response to CMV antigen increased. The monocytes themselves did not contribute to the in vitro specific proliferation response. When iscoms were the carrier particles for CMV antigens, cellular response was even more strongly enhanced. In immunization schedules where specific cellular responses are important, we suggest that autologous monocytes or iscoms may be employed as antigen carriers.