In the challenging race for an HIV-1 vaccine, many researchers are considering genetic vaccines, either based on viral vectors or as naked DNA vaccines, due to their ability to induce protective immune responses in animal models 1. Although both modalities induce potent cellular and humoral responses in animals 23 there is a need to improve the potency of these types of vaccine 4. To improve and broaden immune responses against HIV-1, combinations of viral genes encoding structural, enzymatic and regulatory proteins of HIV-1 have been used 5678. The HIV-1 regulatory genes tat and nef are attractive as vaccine components since both are immunogenic and harbor several T and B cell epitopes. In addition, they are expressed early in the viral life cycle and are relatively well conserved. These features have led to several vaccine trials using the regulatory proteins, trials in which the proteins have been concluded immunogenic and safe in both humans and animals 9101112. However, the regulatory genes of HIV-1 have been shown to influence cellular functions and also to interact with several mechanisms of the host immune response. Nef has been shown to be important for viral replication, viral pathogenesis and progression to AIDS in infected persons 1314. The protein exhibits many intricate actions for helping the virus avoid the immune system, including a potent down-regulation of surface markers such as MHC class I and CD4 molecules (reviewed in 15). More recently, Nef was shown to redistribute the co-stimulatory molecules CD80 and CD86 away from the cell surface of antigen presenting cells 16. Furthermore, Nef is able to interact with the SH3 domain of Hck, a tyrosine kinase of the Src family. This interaction can lead to the transformation of fibroblasts and neuronal cells in vitro 17181920. In addition, experiments using transgenic mice constitutively expressing nef in cells of the kidney support the protein's contribution to the development of HIV-1 associated nephropathy (reviewed in 21). However, little is known about the long-term effects after immunization with Nef. Transgenic mice are well-suited for studying the role of Nef during HIV-1-infection, since the protein in these models is continuously expressed in the animals. They are, however, inappropriate for studying the effects after genetic immunization, in which the protein is expressed only transiently. To assess the potential long-term toxicity following nef-immunization in a model that allows for full necropsy and extensive investigation of the major organs, we immunized mice with a recombinant viral vector, MVA, encoding Nef. This construct has been evaluated for safety and immunogenicity in a clinical phase I trial 10. We chose to use BALB/c mice for evaluation of the long-term toxicity and tumorgenicity of the vaccine construct, since this strain of mice is less prone to spontaneous tumor development than other inbred mice 222324. Thus, the relative stability of BALB/c mice might better mimic the conditions in an outbred human population. The normal life span of female BALB/c mice is approximately twenty-four months 25. By evaluating the toxicity of the nef-encoding MVA construct in this mouse strain over one year, we believe we have a good model for assessing possible long-term side effects induced by Nef-vaccination.

In parallel with the long-term toxicology, we explored the long-term immune responses elicited either by repeated single modality immunization or by heterologous plasmid DNA prime MVA boost regimen.

The inability of plasmid-based vaccines to induce long-lasting immune responses in humans has resulted in the development of alternative strategies for delivering genetic immunogens. The concept of heterologous prime/boost immunization, where two or more vaccine modalities are combined, is now rather well established in experimental vaccinology 2728 and has also been evaluated in clinical trials 2930. One of the most promising prime/boost strategies is the combination of naked DNA with viral vectors. Several attenuated and modified viruses, including species within the families of Pox, Adeno and Alphaviruses, are being explored both as stand-alone vaccine vectors and as components of prime boost vaccine regimens.

We have used two constructs, plasmid DNA and Modified Vaccina virus Ankara (MVA), for delivering the HIV-1 nef gene. We assessed the long-term toxicology and immune responses following immunization with either vector alone or in a DNA prime MVA boost regimen. Both DNAnef and MVAnef have previously been shown to induce novel Nef-specific immune responses in HIV-1-infected humans 910. During the clinical evaluation of the MVAnef construct in HIV-1 infected persons, several parameters were assessed to ensure the safety of the construct. The vaccinees were monitored for standard hematologic and laboratory safety parameters and immunization with MVAnef was concluded to be safe 10. However, the ability of Nef to cause transformation in certain cell types in vitro 17181920 prompted us to perform an even more extensive toxicological evaluation of the vaccine, using a model permitting histopathological analysis of the major organs as well as analysis of the blood and bone marrow for the occurrence of tumors and other abnormalities. These types of extensive analysis are difficult to perform in a clinical trial since they require multiple organ biopsies.

We primed BALB/c mice with two injections of MVAnef or DNAnef and thirty-nine weeks later the animals received either a homologous or a heterologous boost. Three weeks after the boost, the animals were sacrificed and organs and blood were collected for toxicological as well as immunological analysis. The toxicological evaluation led to the conclusion that immunization of the mice with MVAnef did not cause pathological changes in the investigated organs or in the blood. Some abnormalities were noted, including accumulation of white blood cells in the lungs, a moderate decrease of segmented granulocytes in the blood and a slight to moderate degree of anisocytosis (abnormal size of the erythrocytes), poikilocytosis (irregularly shaped erythrocytes) and polychromasia of red blood cells. These observations were made in both immunized and control animals and were assigned to normal age-related pathology, not to effects of the vaccination. The absence of tumors and other adverse effects caused by the vaccine one year after the immunization is encouraging, since the animals were immunized with a massive dose of vaccine in relation to their body weight. We therefore conclude that immunizing with MVAnef does not induce any long-term toxicity.

One issue when using MVA as a vehicle for gene delivery is the anti-vector immunity that inevitably will be induced after immunization. Anti-vector immunity may neutralize the viral vector in subsequent vaccinations, possibly leading to a reduced capacity to boost the immune responses directed against the antigen expressed by the vector 31. Strong cellular reactivity as well as high titers of vaccinia neutralizing antibodies were detected in animals immunized repeatedly with MVA (Figure 5). However, the immune responses against Nef could clearly be boosted by a third injection of MVAnef despite potent anti-vector reactivity (Figure 5b). The ability to boost immune responses despite the presence of neutralizing antibodies directed against the viral vector has been observed by others 32 and is of importance for clinical vaccine regimens where multiple MVA injections are given. The matter of anti-MVA reactivity is also of interest since the virus has been proposed as a candidate smallpox vaccine and there are results indicating that MVA is a potent inducer of immunity to smallpox 3334. The immunity to smallpox induced by the MVA would of course be an additional beneficial effect of using this particular viral vector for delivering other vaccine antigens.

The humoral and cellular immune responses against Nef were monitored during the forty weeks between the second and third immunizations. The two initial MVAnef immunizations induced prominent and stable long-lasting humoral responses against Nef. These results demonstrate the potential of MVA to induce a long-lasting B-cell memory to the vectored antigen, which is highly desirable for a vaccine vector. Intriguingly, when the animals primed with DNAnef were boosted with MVAnef, the humoral response to Nef increased to levels similar to those seen in the animals immunized repeatedly with MVAnef. This increase occurred despite undetectable antibody levels following the two priming DNAnef injections, demonstrating the capacity of plasmid-encoded Nef to prime humoral immune responses.

Interestingly, no significant difference in Nef-specific cellular responses could be detected when comparing the animals immunized repeatedly with either MVAnef or DNAnef or the prime/boost combination of the two vectors. Importantly, as discussed above for the Nef-specific humoral responses, the nef-specific cellular responses could also be readily boosted (Figure 4) despite the presence of anti-vector immunity (Figure 5).

The Nef-specific cellular responses that were detected two weeks after the second immunization, with either DNAnef or MVAnef, decreased over time (Figure 3) but were readily boosted forty weeks later (Figure 4), demonstrating the induction of a long-lasting cellular Nef-specific immunological memory. Importantly, a late boost with DNAnef could significantly enhance immune responses induced with the same DNA construct. These results suggest that the same plasmid vaccine can be used for late boosting and not only, as is common today, for priming of an immune response. This finding could have a great impact on upcoming DNA vaccine trials, as there are several advantages of naked DNA over viral vectors, including the lack of preexisting immunity to the plasmid vector DNA.

Ex-vivo depletion of CD8⁺ cells from splenocytes revealed that only animals immunized with plasmid DNAnef three times over one year displayed a prominent CD8⁺ dependent response. In contrast, animals immunized with MVAnef primarily developed CD4⁺ responses to nef, a finding that was further confirmed by intracellular staining of cryo-preserved splenocytes. These findings are analogous to observations in human trials 10. Interestingly, animals primed with plasmid DNA and subsequently boosted with MVA also predominantly exhibited CD4⁺ and humoral responses, suggesting that MVAnef primarily induce a CD4⁺ T cell memory, while repeated DNA immunization instead amplify the CD8 T cell response.

To conclude, this study shows i) that the nef gene vectored by MVA does not induce malignancies or other adverse effects in mice and may safely be used as a vaccine in humans; ii) that a DNAnef prime followed by an MVAnef boost induced a strong and robust CD4⁺ T cellular response balanced with high titers of Nef-specific antibodies; iii) that immunizing repeatedly with DNAnef induced a strong CD8⁺ T cellular response, while low or no humoral or CD4⁺ T cell responses could be detected; and iv) that it is possible to boost the Nef-specific responses using MVAnef despite the presence of neutralizing anti-vector antibodies.

Our findings have important implications for human immunizations since we demonstrate the possibility to induce a long-lasting immune memory, which can be readily boosted by an additional injection of DNA or recombinant MVA vaccine. Moreover, we demonstrate that it is possible to focus the immune response to either a CD4+ or CD8+ response by selecting the appropriate boosting vector.

ABråve, LG, KH, ER, ABoberg and GE all helped with the ELISpot, the ELISA and the vaccinia neutralization and interpreted the results. SR performed and interpreted the histopathological examinations. GG, WK, AC, and ID developed the MVA construct and performed the intracellular cytokine staining, besides helping to draft the manuscript. ABråve, JH, BW and VE all participated in planning the study, as well as in drafting and finalizing the manuscript. All authors read and approved the final manuscript.