Login| Sign Up| Help| Contact|

Patent Searching and Data


Title:
IMMUNOGENIC COMPOSITIONS CONTAINING BACTERIAL OUTER MEMBRANE VESICLES AND THERAPEUTIC USES THEREOF
Document Type and Number:
WIPO Patent Application WO/2016/184860
Kind Code:
A1
Abstract:
The present invention provides means and products for the stimulation of an immune response against tumors in a subject in need thereof. More specifically the invention provides immunogenic compositions containing bacterial outer membrane vesicles loaded with tumor antigens, fusion proteins comprising a bacterial protein and a tumor antigen, and isolated bacterial outer membrane vesicles (OMVs) containing said fusion proteins. The fusion proteins, OMVs and immunogenic compositions according to the invention are used in the prevention and treatment of tumors.

Inventors:
GRANDI GUIDO (IT)
CAFARDI VALERIA (IT)
FANTAPPIE' LAURA (IT)
GRIFANTINI RENATA (IT)
GRANDI ALBERTO (IT)
Application Number:
PCT/EP2016/061031
Publication Date:
November 24, 2016
Filing Date:
May 17, 2016
Export Citation:
Click for automatic bibliography generation   Help
Assignee:
UNIVERSITA' DEGLI STUDI DI TRENTO (IT)
International Classes:
C07K14/705; C07K14/47; C12N15/62
Domestic Patent References:
WO1999057280A21999-11-11
WO2009013115A22009-01-29
WO2004032958A12004-04-22
WO2006024954A22006-03-09
WO2014106123A12014-07-03
WO2010086162A12010-08-05
WO2010086163A12010-08-05
WO2011051278A12011-05-05
WO2011051276A12011-05-05
WO2011051277A12011-05-05
WO2011051280A12011-05-05
WO2011051271A22011-05-05
WO2011135068A12011-11-03
WO2014198919A22014-12-18
WO2004019977A22004-03-11
WO2005004908A12005-01-20
WO2011036562A12011-03-31
Foreign References:
US20090124557A12009-05-14
EP2255826A22010-12-01
Other References:
BEVERIDGE TJ, J BACTERIOL., vol. 181, 1999, pages 4725 - 33
MAYRAND D; GRENIER D., CAN J MICROBIOL, vol. 35, 1989, pages 607 - 13
A. KULP; KUEHN M.J., ANNU.REV.MICROBIOL., vol. 64, 2010, pages 163 - 184
T.N. ELLIS; KUEHEN M.J., MICROBIOL.MOL.BIOL.REV., vol. 74, 2010, pages 81 - 94
A. KULP; KUEHN M.J., ANNU.REV.MICROBIOL, vol. 64, 2010, pages 163 - 184
T.N. ELLIS; KUEHEN M.J., MICRBIOL.MOL.BIOL.REV., vol. 74, 2010, pages 81 - 94
B.S. COLLINS, DISCOVERY MEDICINE, vol. 12, 2011, pages 7 - 15
J. HOIST ET AL., VACCINE, vol. 27S, 2009, pages B3 - B12
R.C. ALANIZ ET AL., J. IMMUNOL., vol. 179, 2007, pages 7692 - 7701
A. MOSHIRI ET AL., HUM. VACCINES.IMMUNOTHER, vol. 8, 2012, pages 953 - 955
T.N. ELLIS ET AL., INN.IMMUN, vol. 78, 2010, pages 3822 - 3831
M. KAPARAKIS ET AL., CELL.MIOCROBIOL, vol. 12, 2010, pages 372 - 385
D.J. CHEN ET AL., PNAS, vol. 107, 2010, pages 3099 - 3104
MALLEY R ET AL., INFECT IMMUN., vol. 74, 2006, pages 2187 - 95
ROY N. ET AL., IMMUNOL. CLINICAL MICROBIOL., vol. 60, 2010, pages 18 - 27
ROIER S. ET AL., INTJ.MED.MICROBIOL., vol. 303, 2013, pages 247 - 256
KIM OY ET AL., J. IMMUNOL., vol. 190, 2013, pages 4092 - 4102
N.C. KESTY; KUHEN M.J., J.BIOL.CHEM., vol. 279, 2004, pages 2069 - 2076
J. SCHROEDER; AEBISCHER T., VACCINE, vol. 27, 2009, pages 6748 - 6754
FANTAPPIE ET AL., JOURNAL OF EXTRACELLULAR VESICLES, vol. 3, 2014, pages 24015
FERRARI ET AL., PROTEOMICS, vol. 6, 2006, pages 1856 - 1866
BERNADAC A. ET AL., J. BACTERIOL., vol. 180, 1998, pages 4872 - 4878
DEATHERAGE B.L. ET AL., MOL. MICROBIOL, vol. 72, 2009, pages 1395 - 1407
MCBROOM A.J.; KUEHEN M.J., MOL. MICROBIOL., vol. 63, 2007, pages 545 - 558
FREDRIKSEN J.H. ET AL., NIPH ANN., vol. 14, 1991, pages 67 - 79
GRANOFF D., CLIN.INFECT.DIS., vol. 50, 2010, pages S54 - S65
BERLANDA SCORZA F. ET AL., PLOSONE, vol. 7, 2012, pages E35616
SNAPE M.D. ET AL., PEDIATR. INFECT. DIS. J., vol. 29, 2010, pages E71 - E79
FISSEHA M. ET AL., INFECT. IMMUN., vol. 73, 2005, pages 4070 - 4080
STEEGHS L. ET AL., INFECT. IMMUN., vol. 76, 2008, pages 3801 - 3807
LEE S-R ET AL., J. MICROB. BIOTECHNOL., vol. 19, 2009, pages 1271 - 1279
DONG H.L. ET AL., VACCINE, vol. 29, 2011, pages 8293 - 8301
NEEDHAM B.D. ET AL., PNAS, vol. 110, 2013, pages 1464 - 1469
F. CANTINI ET AL., JBIOL CHEM., vol. 281, 2006, pages 7220 - 7
LO H. ET AL., LANCET INFECT.DIS., vol. 9, 2009, pages 418 - 427
FLETCHER L. D. ET AL., INFECT. IMMUN., vol. 72, 2004, pages 2088 - 2100
MASIGNANI ET AL., JEM, vol. 6, 2003, pages 789 - 799
MASIGNANI V ET AL., JEM, vol. 6, 2003, pages 789 - 799
GIULIANI MM ET AL., INFECT IMMUN, vol. 2, 2005, pages 1151 - 1160
SCHNEIDER MC ET AL., NATURE, vol. 458, 2009, pages 890 - 893
FALERI A ET AL., FASEB J, vol. 4, 2014, pages 1644 - 53
M.P. BOS ET AL., ANN. REV. MICROBIOL., vol. 61, 2007, pages 191 - 214
TOKUDA H., BIOSCI. BIOTECHNOL. BIOCHEM., vol. 73, 2009, pages 465 - 73
L.D. FLETCHER ET AL., INF. IMMUN., vol. 72, 2004, pages 2008 - 2100
RICHMOND PC ET AL., THE LANCET INFECTIOUS DISEASES, vol. 12, 2012, pages 597 - 607
ESPOSITO, V. ET AL., JBIOL CHEM, vol. 286, no. 48, 2011, pages 4167 - 75
VAN WINKELHOFF, A.J.; J. SLOTS, PERIODONTOL 2000, vol. 20, 1999, pages 122 - 35
HENDERSON, B.; J.M. WARD; D. READY, PERIODONTOL 2000, vol. 54, no. 1, 2010, pages 78 - 105
KIESELBACH, T. ET AL., PLOS ONE,, vol. 10, no. 9, 2015, pages E0138591
ZIJNGE, V.; T. KIESELBACH; J. OSCARSSON, PLOS ONE, vol. 7, no. 7, 2012, pages E41662
SCHWARTZ ET AL., EUR. J. BIOCHEM., vol. 71, 1976, pages 167 - 170
SZMELCMAN ET AL., EUR. J. BIOCHEM., vol. 65, 1976, pages 13 - 19
WANDERSMAN ET AL., J. BACTERIOL, vol. 140, 1979, pages 1 - 13
ZUKIN, S., BIOCHEMISTRY, vol. 18, 1979, pages 2139 - 2145
FERENCI; RANDALL, J. BIOL. CHEM., vol. 254, 1979, pages 9979 - 9981
FERRARIO, M. ET AL., J. BACTERIOL., vol. 177, 1995, pages 103 - 1132
HALL, M. N.; SILHAVY, T. J., J. MOL. BIOL., vol. 151, 1981, pages 1 - 15
MIZUNO ET AL., PROC. NATL. ACAD. SCI. U. S. A., vol. 81, 1984, pages 1966 - 1970
CHEN, S.; ZHANG, A.; BLYN, L. B.; STORZ, G., J. BACTERIOL., vol. 186, 2004, pages 6689 - 66975
BASLE ET AL., PROTEIN ENGINEERING, DESIGN & SELECTION, vol. 9, 2004, pages 665 - 672
COLEY WB, AM. J. MED. SCI., vol. 105, 1893, pages 487 - 511
VAN; BOON, PNAS, vol. 79, 1982, pages 4718 - 4722
LU C. ET AL., EXP.OPIN. BIOL. THER., vol. 11, 2011, pages 99 - 108
RESTIFO ET AL., NATURE REV. IMMUNOL., vol. 12, 2012, pages 269 - 281
BERINSTEIN NL, VACCINE, vol. 25S, 2007, pages B72 - B88
MITTENDORFF EA ET AL.: "Annals of Oncology", vol. 25, 2014, pages: 1735 - 1742
DEL VECCHIO CA ET AL., EXPERT REV. VACCINES, vol. 11, 2012, pages 133 - 144
PORTER KR ET AL., NEURO. ONCOL., vol. 12, 2010, pages 520 - 527
GROSSMAN SA ET AL., SEMIN. ONCOL., vol. 31, 2004, pages 635 - 644
VREDENBURGH JJ ET AL., J. CLIN. ONCOL., vol. 25, 2007, pages 4722 - 4729
STUPP R ET AL., N. ENGL. J. MED., vol. 352, 2005, pages 987 - 996
SALOMON DS ET AL., CRIT. REV. ONCOL. HEMATOL., vol. 19, 1995, pages 183 - 232
WONG AJ ET AL., PROC. NATL ACAD. SCI. USA, vol. 89, 1992, pages 2965 - 2969
GORGOULIS V ET AL., ANTICANCER RES., vol. 12, 1992, pages 1183 - 1187
IRISH JC ET AL., LARYNGOSCOPE, vol. 103, 1993, pages 42 - 52
KORC M ET AL., PROC. NATL ACAD. SCI. USA, vol. 83, 1986, pages 5141 - 5144
MOORGHEN M ET AL., ANTICANCER RES., vol. 10, 1990, pages 605 - 611
ISHIKAWA J ET AL., INT. J. CANCER, vol. 45, 1990, pages 1018 - 1021
ZAJCHOWSKI D ET AL., CANCER RES., vol. 48, 1988, pages 7041 - 7047
MOSCATELLO DK ET AL., CANCER RES., vol. 55, 1995, pages 5536 - 5539
HUMPHREY ET AL., PNAS, vol. 87, 1990, pages 4207
DEL VECCHIO ET AL., EXP.REV., vol. 11, 2012, pages 133
DUNNE, J. ET AL., GENOMICS, vol. 30, 1995, pages 207 - 23
MOELLER, M.J. ET AL., THE EMBO JOURNAL, vol. 23, 2004, pages 3769 - 79
MORRIS, L.G.T. ET AL., NATURE GENETICS, vol. 45, 2013, pages 253 - 61
CHOSDOL, K. ET AL., BMC CANCER, vol. 9, 2009, pages 5
NAKAYA, K. ET AL., ONCOGENE, vol. 26, 2007, pages 5300 - 8
SETTAKORN, J. ET AL., JOURNAL OF CLINICAL PATHOLOGY, vol. 58, 2005, pages 1249 - 54
MORRIS, L.G.T. ET AL., CELL CYCLE (GEORGETOWN, TEX.), vol. 12, 2013, pages 1011 - 2
SETTAKORN, J. ET AL., JOURNAL OF CLINICAL PATHOLOGY,, vol. 58, 2005, pages 1249 - 54
KWAEPILA, N. ET AL., PATHOLOGY, vol. 38, 2006, pages 125 - 31
DE BOCK, C.E. ET AL., LEUKEMIA, vol. 26, 2012, pages 918 - 26
VALLETTA, D. ET AL., CARCINOGENESIS, vol. 35, 2014, pages 1407 - 15
HO SB ET AL., CANCER RES, vol. 53, 1993, pages 641 - 51
BALDUS SE ET AL., HISTOPATHOLOGY, vol. 40, 2002, pages 440 - 9
RABASSA ME ET AL., BMC CANCER, vol. 6, 2006, pages 253
BARND, D.L. ET AL., PROC. NATL. ACAD. SCI. USA, vol. 86, 1989, pages 7159 - 7163
JEROME, K.R. ET AL., CANCER RES., vol. 51, 1991, pages 2908 - 2916
IOANNIDES, C.G. ET AL., J. IMMUNOL., vol. 151, 1993, pages 3693 - 3703
VLAD, A.M. ET AL., J. EXP. MED., vol. 196, 2002, pages 1435 - 1446
VONMENSDORFF-POUILLY, S. ET AL., J. CLIN. ONCOL., vol. 18, 2000, pages 574 - 583
HAMANAKA, Y. ET AL., INT. J. CANCER, vol. 103, 2003, pages 97 - 100
PINHEIRO, S.P. ET AL., CANCER EPIDEMIOL. BIOMARKERS PREV., vol. 19, 2010, pages 1595 - 1601
KIMURA T.; FINN OJ, EXPERT OPIN. BIOL. THER., vol. 13, 2013, pages 35 - 49
KLOCK H.E.; LESLEY S.A, METHODS MOL. BIOL., vol. 498, 2009, pages 91 - 103
KLOCK H.E.; LESLEY S.A, METHODS MOL. BIOL, vol. 498, 2009, pages 91 - 103
Attorney, Agent or Firm:
BANFI, Paolo (Via Plinio 63, Milano, IT)
Download PDF:
Claims:
CLAIMS

1. A fusion protein comprising a bacterial protein and one or more copies of a tumor antigen protein, wherein the bacterial protein is selected from: Factor H Binding Protein (fHbp), Neisseria heparin binding antigen (NHBA), Maltose Binding Protein (MBP), Outer Membrane Protein-F (ompF) and Aggregatibacter actinomycetemcomitans Factor H binding protein (Aa-fHbp).

2. A fusion protein according to claim 1, wherein said fHbp is either the full length protein or the Domain A thereof (fHbpDomA).

3. A fusion protein according to claims 1-2, comprising from 1 to 20 copies of the tumor antigen protein, optionally spaced by a linker sequence.

4. A fusion protein according to claims 1-3, wherein said bacterial protein is selected from: fHbp, SEQ ID NO:l; fHbpDomA, SEQ ID NO:2; omp-F, SEQ ID NO:3; MBP, SEQ ID NO:4; NHBA, SEQ ID NO: 109; Aa-fHbp, SEQ ID NO:110

5. A fusion protein according to claims 1-4, wherein the tumor antigen is linked to the carboxyl terminus of the bacterial protein or, when the bacterial protein is OMP, it is inserted in or replaces an external loop of the protein.

6. A fusion protein according to claims 1-5, wherein the tumor antigen is selected from hEGFRvIII, hFAT-1 and hMUC-1, or an immunogenic fragment thereof.

7. A fusion protein according to claim 6, wherein the hEGFRvIII immunogenic fragment is LEEK GNYVVTDH (SEQ ID NO:5).

8. A fusion protein according to claim 6, wherein the hFAT-1 immunogenic fragment is IQVEATDKDLGPNGHVTYSIVTDTD (SEQ ID NO:6).

9. A fusion protein according to claim 6, wherein the hMUC-1 immunogenic fragment is GVTSAPDTRPAPGSTAPPAH (SEQ ID NO:7).

10. A fusion protein according to claims 1-6, which is selected from the group consisting of SEQ ID NO:8 through SEQ ID NO:25; SEQ ID NO: l 1 and SEQ ID NO:l 12.

11. An isolated bacterial outer membrane vesicle comprising a fusion protein according to any preceding claim, wherein said fusion protein carries one single tumor antigen or the outer membrane vesicle comprises a plurality of fusion proteins carrying different tumor antigens.

12. An isolated bacterial outer membrane vesicle, which is secreted by Escherichia coli. 13. An immunogenic composition comprising a bacterial outer membrane vesicle as per claims 11-12.

14. An immunogenic composition according to claim 13, comprising a mixture of outer membrane vesicles carrying different tumor antigens.

15. An immunogenic composition according to claims 13-14, further containing pharmaceutically acceptable adjuvants and excipients.

16. A cancer vaccine containing the immunogenic composition of claims 13-15.

17. An isolated bacterial outer membrane vesicle according to claims 11-12 or an immunogenic composition according to claims 13-15 or a cancer vaccine according to claim 16, for use in the stimulation of an immune response against tumor in a subject in need thereof.

18. An isolated bacterial outer membrane vesicle or an immunogenic composition or a cancer vaccine according to claim 17, for use in the prevention or treatment of tumors.

19. An isolated bacterial outer membrane vesicle or an immunogenic composition or a cancer vaccine according to claim 17, for use in the prevention or treatment of tumors selected from breast, brain, head-and-neck, non-small cell lung, renal, ovarian, kidney, stomach, prostate and colon cancer, oral cancer, astrocytoma, glioblastoma, ductal carcinoma, cholangiocarcinoma, hepatocarcinoma, acute myeloid leukemia, acute lymphoblastic leukemia, melanoma and pancreatic cancer.

Description:
IMMUNOGENIC COMPOSITIONS CONTAINING BACTERIAL OUTER

MEMBRANE VESICLES AND THERAPEUTIC USES THEREOF

The present invention provides means and products for the stimulation of an immune response against tumors in a subject in need thereof. More specifically the invention provides immunogenic compositions containing bacterial outer membrane vesicles loaded with tumor antigens, fusion proteins comprising a bacterial protein and a tumor antigen, and isolated bacterial outer membrane vesicles (OMVs) containing said fusion proteins. The fusion proteins, OMVs and immunogenic compositions according to the invention are used in the prevention and treatment of tumors.

BACKGROUND ART

Bacterial Outer Membrane Vesicles (OMVs)

More than 40 years ago, researchers made the observation that Gram-negative bacteria secrete Outer Membrane Vesicles (OMVs) (Beveridge TJ (1999) J Bacteriol. 181 :4725-33; Mayrand D & Grenier D. (1989) Can J Microbiol. 35:607-13). However, the last 15 to 20 years have brought a greater understanding of the regulation and function of vesiculation. OMVs are closed spheroid particles of a heterogeneous size, 20-300 nm in diameter, generated through a "budding out" of the bacterial outer membrane. Consistent with that, the majority of OMV components are represented by LPS, glycerophospholipids, outer membrane proteins and periplasmic proteins (A. Kulp and Kuehn M.J. (2010) Annu.Rev. Microbiol. 64, 163-184; T.N. Ellis and Kuehen M.J. (2010) Microbiol.Mol.Biol.Rev. 74, 81-94).

OMVs represent a distinct secretory pathway with a multitude of functions, including inter and intra species cell-to-cell cross-talk, bio film formation, genetic transformation, defense against host immune responses and toxin and virulence factor delivery to host cells (A. Kulp and Kuehn M.J. (2010) Annu.Rev. Microbiol. 64, 163-184). OMVs interaction to host cells can occur by endocytosis after binding to host cell receptors or lipid rafts. Alternatively, OMVs have been reported to fuse to host cell membrane, leading to the direct release of their content into the cytoplasm of the host cells (A. Kulp and Kuehn M.J. (2010) Annu. Rev. Microbiol. 64, 163-184; T.N. Ellis and Kuehen M.J. (2010) Micrbiol.Mol.Biol.Rev. 74, 81-94).

OMVs purified from several pathogens, including Neisseria, Salmonella, Pseudomonas, Vibrio cholerae Burkholderia, and E. coli, induce potent protective immune responses against the pathogens they derive from (B.S. Collins (2011) Discovery Medicine, 12, 7-15), and highly efficacious anti-Neisseria OMV-based vaccines are already available for human use (J. Hoist et al. (2009) Vaccine, 27S, B3-B12). Such remarkable protection is attributed to three key features of OMVs.

First, they are readily phagocytosed by professional antigen-presenting cells which get activated and present OMV-derived peptides in the context of MHC molecules.

Second, they carry the proper immunogenic antigens which, in extracellular pathogens, usually reside on the surface and therefore are naturally incorporated in OMVs. Indeed, OMV immunization induces potent antibody responses against the major membrane-associated antigens. However, OMV immunogenicity is not restricted to antibody responses. For instance, mice immunized with Salmonella OMVs develop robust Salmonella-specific B and T cell responses, and OMVs stimulate IFN-γ production by a large proportion of CD4+ T cells from mice previously infected with Salmonella, indicating that OMVs are an abundant source of antigens recognized by Salmonella-specific CD4+ T cells (R.C. Alaniz et al.,(2007) J. Immunol. 179, 7692-7701).

Third, and most importantly, OMVs carry most of the bacterial Pathogen- Associated-Molecular Patterns (PAMPs) which, by binding to pathogen recognition receptors (PRRs), play a key role in stimulating innate immunity and promoting adaptive immune responses. OMV-associated PAMPs include LPS which, in concert with MD-2 and CD14, binds TLR-4, lipoproteins whose acylpeptide derivatives interact with TLR-1/2 and 2/6 heterodimers, and peptidoglycan whose degradation products bind to intracellular NODI/2 (A. Moshiri et al, Hum. Vaccines. Immunother. (2012) 8, 953-955; T.N. Ellis et al, (2010) Inn.Immun. 78, 3822-3831; M. Kaparakis et al, (2010) Cell.Miocrobiol. 12, 372-385). The engagement of this group of PPRs results in the activation of transcription factors (NF-kB) and the consequent expression of specific cytokines. Interestingly, LPS, lipoproteins and peptidoglycan can work synergistically, thus potentiating the built-in adjuvanticity of OMVs (D.J. Chen et al, (2010) PNAS, 107, 3099-3104).

An additional interesting property of OMVs is their capacity to induce protection at the mucosal level. Protection at the mucosal sites is known to be at least partially mediated by the presence of pathogen-specific IgAs and Thl7 cells. In particular, a growing body of evidence suggests that Thl7 cells have evolved to mediate protective immunity against a variety of pathogens at different mucosal sites. Interestingly, Thl7 cells have recently also been shown to play a crucial role in the generation of vaccine-induced protective responses. For instance, it has been reported that in mice whole cell pertussis vaccines (Pw) induce Thl7 cells and neutralization of IL-17 after vaccination reduces protection against a pulmonary challenge with B. pertussis. Similarly, in a CD4+ T cell dependent, antibody-independent model of vaccine-induced protection following S. pneumoniae challenge, treatment with IL- 17-antibodies resulted in reduced immunity to pneumococcal colonization compared to the control serum treated mice (Malley R, et al. (2006) Infect Immun., 74:2187-95).

Elicitation of IgAs and Thl7 cells by OMVs has been well documented and this can explain mechanistically the good protective activities of OMVs against several mucosal pathogens. For instance, immunization with Vibrio cholerae-derived OMVs protects rabbits against Vibrio cholerae oral challenge (Roy N. et al. (2010) Immunol. Clinical Microbiol. 60, 18-27) and Pasteurella multocida-densQd and Mannheimia haemolytica- derived OMVs protect mice from oral challenge with P. multocida (Roier S. et al, (2013) IntJ.Med.Microbiol. 303, 247-256). In addition, intranasal immunization with Porphyromonas gingivalis OMVs elicits potent IgA production at both serum and mucosal level and immunization with Escherichia co/z-derived OMVs prevent bacteria-induced lethality. Protective effect of Escherichia co/z ' -derived OMVs is primarily mediated by OMV-specific, IFN-γ and IL-17 producing, T cells (Kim OY et al, (2013) J. Immunol. 190, 4092-4102).

Finally, a key feature of OMVs is the possibility to manipulate their protein content by genetic engineering. This feature was demonstrated for the first time by Kesty and Kuehn who showed that Yersinia enterocolitica outer membrane protein Ail assembled on OMVs surface when expressed in E. coli, and that the GFP fluorescence protein fused to the "twin arginine transport (Tat)" signal sequence was incorporated in the OMV lumen (N.C. Kesty and Kuhen M.J. (2004) J.Biol.Chem. 279, 2069-2076). Following the observation by Kesty and Kuehn, an increasing number of heterologous proteins have been successfully delivered to OMVs using a variety of strategies. For instance, heterologous antigens have been delivered to the surface of OMVs by fusing them to the β-barrel forming autotransporter AIDA and to hemolysin ClyA, two proteins that naturally compartmentalized into E. coli OMVs (J. Schroeder and Aebischer T. (2009) Vaccine, 27, 6748-6754; D.J. Chen et al, (2010) PNAS, 107, 3099-3104). Recently, heterologous antigens from Group A Streptococcus and Group B Streptococcus were delivered to the lumen of E.coli vesicles by fusing their coding sequences to the leader peptide of E. coli OmpA. Interestingly, when the recombinant vesicles were used to immunize mice, they elicited high titers of functional antibodies against the heterologous antigens, despite their luminal location (Fantappie et al, (2014) Journal of Extracellular Vesicles, 3, 24015).

Despite the many strategies successfully used to deliver heterologous antigens to the vesicle compartment, it has to be pointed out that a universal system working for any protein antigen has not been described yet. A strategy that is effective for one specific antigen in terms of level of expression and elicitation of immune responses can be inefficient with other antigens.

Therefore, the identification of novel strategies to deliver antigens to the OMV compartment is highly needed.

In general, the amount of OMV released by Gram-negative bacteria when grown under laboratory conditions is too low to allow the exploitation of OMVs in biotechno logical applications. However, it has been shown that under stress conditions, such as high temperature, OMV release is substantially increased. Furthermore, a number of mutations have been described, many of them affecting the composition of the periplasmic and membrane compartments, that result in abundant vesicle production. For instance, in Neisseria meningitidis, a mutation in the gna33 gene, encoding a glucosyltransferase, has been shown to drive the release of several milligrams of vesicles per liter in the culture supernatant (Ferrari et al., (2006) Proteomics, 6, 1856-1866). Similar quantities of vesicles are obtained from Escherichia coli strains carrying deletions in the genes encoding the Tol/Pal system (a protein complex involved in the connection of the inner membrane with the outer membrane) (Bernadac A. et al., (1998) J. Bacteriol. 180, 4872-4878) and in the ompA gene, encoding one of the major outer membrane proteins of E. coli (Fantappie et al, (2014) Journal of Extracellular Vesicles, 3, 24015). Such quantities make the production process of OMVs highly efficient and inexpensive thus making natural and engineered vesicles extremely attractive for vaccine purposes. A number of other mutations have been described that enhance the production of OMVs in several Gram negative bacteria, including Salmonella and E. coli (Deatherage B.L. et al. (2009) Mol. Microbiol. 72, 1395-1407; McBroom A.J. and Kuehen M.J. (2007) Mol. Microbiol. 63, 545-558), and such mutations are amenable to be exploited to develop scalable OMV production processes.

As far as the production of OMVs for industrial applications is concerned, a number of methods have been described. The first ones to be developed make use of mild detergents that promote the production of great yield of vesicles from biomass and decrease the level of toxicity by removing a substantial amount of LPS (Fredriksen J.H. et al, (1991) NIPH Ann. 14, 67-79). Although these processes have been proved to produce safe and effective vaccines designed to fight Meningococcal B epidemics (Granoff D. (2010), Clin.Infect.Dis. 50, S54-S65) their main drawback is that the detergent treatment favor bacterial cell lysis with the consequence that the OMV preparations are heavily contaminated with cytoplasmic proteins (Ferrari et al, (2006) Proteomics, 6, 1856-1866). This affect the immune response and can reduce the breath of protection. More recently, detergent-free methods for OMV production have been proposed. Such methods involve the separation of the bacterial culture supernatant from biomass and the purification of vesicles from the supernatant using tangential flow filtration (TFF) (Berlanda Scorza F. et al, (2012) PlosOne 7, e35616).

The yield of OMV production using centrifugation couple to TFF can exceed 100 mg/liter of culture and therefore the process is perfectly compatible with large scale production.

One of the potential issues encountered in using OMVs in vaccine applications is the presence of lipopolysaccharide (LPS), an endotoxin known to be reactogenic both in animals and humans. Possible strategies to reduce reactogenicity is to extract LPS from OMVs using mild detergents (Fredriksen J.H. et al, (1991) NIPH Ann. 14, 67-79) or to formulate OMVs with alum hydroxide which absorbs LPS and keeps it confined at the site of injection (Ferrari et al, (2006) Proteomics, 6, 1856-1866; Snape M.D. et al, (2010) Pediatr. Infect. Dis. J. 29, e71-e79). Another strategy is to genetically alter the LPS synthetic pathway of the strain used for OMV production so that the purified vesicles carry modified versions of LPS with reduced reactogenicity. For instance, in Neisseria meningitidis one promising mutant with attenuated endotoxin activity contains a deletion in the IpxLl gene (also referred to as the msbB gene) (Fisseha M. et al, (2005) Infect. Immun., 73:4070-4080). This mutation results in a LPS carrying a penta-acylated lipid A, which is poorly recognized by human Toll-like receptor 4 (Steeghs L. et al. (2008) Infect. Immun., 76:3801-3807), instead of the more toxic hexa-acylated lipid A, which is present in the LPS produced by wild-type strains. The inactivation of msbB gene to produce less toxigenic OMVs has also been reported for Shigella, Salmonella and E. coli (Berlanda Scorza F. et al, (2012) PlosOne 7, e35616; Lee S-R et al, (2009) J. Microb. Biotechnol. 19, 1271-1279; Dong H.L. et al, (2011) Vaccine, 29, 8293-8301). In E. coli an additional mutation in the pagP gene has been described that, when combined with msbB mutation, results in the production of LPS with a fully penta-acylated lipid A which has a low reactogenicity property (Dong H.L. et al, (2011) Vaccine, 29, 8293-8301). Finally, by using Synthetic Biology, Needham and co-workers (Needham B.D. et al, (2013) PNAS, 110, 1464-1469) have created a collection of novel LPS synthetic pathways which lead to the synthesis of LPS carrying different modifications, each displaying distinct TLR4 agonist activities, cytokine induction and reactogenicity properties. OMVs purified fromE. coli carrying such engineered LPS pathways have high potential for the design of vaccines with ad hoc modulated immunogenicity and adjuvanticity properties.

Neisseria meningitidis Factor H Binding Protein

Factor H binding protein (fHbp) is a 28 kDa surface-exposed lipoprotein of Neisseria meningitidis (F. Cantini et al, (2006), J Biol Chem., 281, 7220-7). fHbp is able to bind factor H (fH), the central regulator of the alternative complement pathway, and in this way it down-regulates complement activation (Lo H. et al, (2009) Lancet Infect.Dis. 9: 418-427) and impairs complement-mediated bacterial lysis by human plasma. fHbp is present on the surface of most meningococcal strains (Fletcher L. D. et al., (2004) Infect. Immun., 72: 2088-2100) and high levels of fHbp expression have been found in hyper- virulent meningococcal strains (Masignani et al, (2003) JEM 6: 789-799). Sequence analysis has classified fHbp in 3 main variant groups (varl, var2 and var3) (Masignani V et al (2003) JEM 6: 789-799). To study the immunogenic and functional properties of the protein, fHbp variant 1 was initially divided for convenience into three regions, named "domain" A, B, and C (Giuliani MM et al (2005) Infect Immun. 2: 1151-1160). Domain A encompasses amino acids 27 to 119 of the lipoprotein unprocessed precursor, domain B starts from amino acid 120 and ends at residue 183, and finally domain C spans from amino acid 184 to the end (amino acid 274) (F. Cantini et al, (2006), J Biol Chem., 281, 7220-7). More recent structural studies revealed that fHbp folds to form two β-barrels, with the amino -terminal barrel consisting of the A and part of the B regions and the carboxy- terminal barrel composed of the rest of the B and the C regions (Schneider MC et al, (2009) Nature, 458: 890-893; Faleri A et al, (2014) FASEB J. 4: 1644-53).

In the E. coli model Gram-negative bacterium, in which the lipoprotein sorting process has been well characterized, all outer membrane lipoproteins face the periplasmic space. They reach their final destination in two major steps (M.P. Bos et al. (2007) Ann. Rev. Microbiol., 6J_: 191-214). First, the protein is synthesized as a precursor carrying at its N- terminal a leader peptide (LP) including a lipobox. This LP is recognized by Sec pathway and the protein crosses the cytoplasmic membrane. The first amino acid of mature lipoprotein is a cysteine. Processing into mature form takes place on the periplasmic side of the inner membrane, where the thiol group on the side chain of the cysteine residue is modified by the covalent attachment of a diacylglycerol moiety and an amide-linked acyl group is attached to the N-terminus. Both the diacylglyceryl group and the amino -terminal acyl group participate to the anchorage of the lipoprotein to the membrane. The second step consists in the translocation of the acylated protein from the inner membrane to the inner leaflet of the outer membrane. Translocation is mediated by the Lol system (Tokuda H. (2009) Biosci. Biotechnol. Biochem., 73, 465-73). In N. meningitidis and other Gram- negative species, cell- surface exposed lipoproteins have been identified. The molecular mechanism responsible for the flipping out of fHbp and other surface lipoproteins from the inner to the outer leaflet of the outer membrane is not known.

fHbp is efficiently incorporated into OMVs released by Neisseria meningitidis and fHbp-containing OMVs have been shown to induce potent bactericidal antibodies against Neisseria meningitidis. Therefore, such engineered OMVs have been proposed as highly efficacious vaccines (EP 22 55 826 A2).

fHbp has been successfully expressed in E. coli and the protein accumulates in the membranes fraction and is efficiently lipidated. Whether or not the protein, when expressed in E. coli, is exposed on the surface as it is the case in Neisseria meningitidis has not been reported. If a specific flippase is involved in fHbp translocation, when expressed in E. coli the protein is expected to be anchored to the inner leaflet of the outer membrane but not surface-exposed. Lipidated fHbp has been shown to induce high titers of bactericidal antibodies and an anti-meningococcus B vaccine based on lipidated fHbp purified from a recombinant E. coli strain is close to registration in USA and Europe (L.D. Fletcher et al, (2004) Inf. Immun. 72, 2008 - 2100; Richmond PC, et al, (2012), The Lancet Infectious Diseases, 12, 597-607).

Neisseria meningitidis NHBA (Neisseria heparin binding antigen)

Neisseria heparin binding antigen (NHBA) is a Neisseria meningitidis surface-exposed lipoproteins which play important roles in pathogenesis. NHBA carries an arginine-rich region responsible for heparin binding, which correlates with an increased survival of N.meningtidis in human serum. NHBA fold consists of an 8-strand β -barrel that closely resembles the C-terminal domain of Nm-fHbp suggesting that they derive from a common ancestor (Esposito, V. et al. J Biol Chem, 2011. 286(48). 4167-75.

Consistent with its biological role NHBA extend out of the bacterial surface and is accessible to antibodies with bactericidal activity.

Aggregatibacter actinomycetemcomitans Factor H binding protein (Aa-fHbp) A. actinomycetemcomitans is a human pathogen associated with aggressive periodontitis and endocarditis (van Winkelhoff, A.J. and J. Slots, Periodontal 2000, 1999. 20,122-35; Henderson, B., J.M. Ward, and D. Ready, Periodontal 2000, 2010. 54(1), 78-105).

The genome sequence of the pathogen revealed the presence of an Open Reading Frame (ORF) encoding a lipoprotein (Aa-fHbp) sharing 41% amino acid identity with Nm- fHbp var.l (Figure 24A). Recent proteomic analyses showed that Aa-fHbp localized in the culture supernatants and/or in OMVs of A. actinomycetemcomitans (Kieselbach, T., et al, PLoS One, 2015. 10(9), e0138591; Zijnge, V., T. Kieselbach, and J. Oscarsson, PLoS One, 2012. 7(7), e41662). Although no direct evidence has been reported so far on the localization of Aa-fHbp, its predicted Factor H binding capacity strongly suggests that the protein protrudes out of the outer membrane of A. actinomycetemcomitans.

Escherichia coli Maltose Binding Protein

The Maltose Binding Protein (MBP) is a periplasmic protein that specifically binds maltose and maltodextrins, with highest affinity for maltotriose (Schwartz et al., (1976) Eur. J. Biochem. 71, 167-170; Szmelcman et al, (1976) Eur. J. Biochem., 65, 13-19; Wandersman et aL, (1979) J. Bacteriol, 140, 1-13). Studies on the binding specificity have suggested that the binding site recognizes the glycosidic bond linking the glucose moities of maltose. MBP can be purified in dimeric form from bacteria. Maltose induces the conversion of the protein dimers to monomers and upon binding of the substrate, the monomer undergoes a conformational change (Szmelcman et al., (1976) Eur. J. Biochem., 65, 13-19; Zukin, S. (1979) Biochemistry 18,2139-2145).

MBP is synthesized in large amounts. A fully induced cell may contain up to 40,000 monomers. Expression of the malE gene is activated by the MalT protein in the presence of maltose or maltodextrim and by the cyclic AMP receptor protein.

Like most outer membrane and periplasmic proteins, MBP is synthesized initially with an amino -terminal signal peptide which interacts with the cell secretory apparatus and is cleaved during the export process. The MBP precursor is active in binding maltose (Ferenci and Randall (1979) J. Biol. Chem. 254, 9979-9981). Interestingly, MBP is essential for the energy-dependent translocation of maltose and maltodextrins through the cytoplasmic membrane.

MBP is extensively used to express proteins in E. coli and to increase their solubility.

Proteins of interest are usually expressed as MBP-fusions, and this has been shown to prevent aggregation through a mechanism that is not fully understood. MBP can itself be used as an affinity tag for purification of recombinant proteins. The fusion protein binds to amylose columns while all other proteins flow through. The MBP-protein fusion can be purified by eluting the column with maltose. Once the fusion protein is obtained in purified form, the protein of interest is often cleaved from MBP with a specific protease and separated from MBP by ion exchange chromatography.

Since MBP is a periplasmic protein, chimeric MBP carrying a heterologous protein at its C-terminus also accumulates in this bacterial compartment. However, the delivering of MBP chimera to OMVs and their exploitation as immunogens has never been tested.

OmpF

Escherichia coli expresses three major Outer Membrane Proteins (OMPs) OmpA, OmpC and OmpF. OmpA is noncovalently anchored to peptidoglycan, and is constituted by eight β- barrel strands connected by four long loops at the outer membrane surface and three short periplasmic turns. The other two OMPs, OmpF and OmpC, are porins which serve as passive diffusion pores across the outer membrane. Expression of these two proteins is reciprocally regulated by medium osmolarity. OmpF is preferentially produced at low osmolarity, whereas OmpC is almost exclusively produced at high osmolarity. It has been proposed that the larger OmpF pore size is important for efficient nutrient uptake from nutritionally poor media, whereas the smaller OmpC pore size is important to exclude the passage oftoxic bile salts across the outer membrane (Ferrario, M. et al, (1995) J. Bacteriol. 177. 103-1132).

E. coli has a highly sophisticated regulatory system to modulate the reciprocal expression of ompF and ompC. Both are controlled at the transcriptional level by the histidine kinase EnvZ, a transmembrane osmosensor, and the response regulator OmpR, a transcriptional factor (Hall, M. N., and Silhavy, T. J. (1981) J. Mol. Biol. 151, 1-15), and also at the translational level by antisense R As: micF for ompF mR A (Mizuno, et al. (1984) Proc. Natl. Acad. Sci. U. S. A. 81, 1966-1970) and micC for ompC mRNA (Chen, S., Zhang, A., Blyn, L. B., and Storz, G. (2004) J. Bacteriol. 186, 6689-66975).

When grown at 37°C in rich media such as LB or in minimal media, such as M9 medium, supplemented with glycerol, E. coli BL21(D3) mostly produces OmpA and OmpF. The two proteins accumulate in the outer membrane vesicles (OMVs) released by the strain and when the ompA gene is inactivated OmpF constitutes more than 50% of total OMV proteins (Fantappie et al, (2014) Journal of Extracellular Vesicles, 3, 24015).

Structurally, OmpF is characterized by the trimeric assembly of monomeric 16- stranded β-barrels, each containing its own hydrophilic pore. Each barrel essentially spans the thickness of the membrane. Except for the third loop that folds inward and constricts the channel opening, seven long loops connect adjacent pairs of β-strands on the extracellular side. Recently each OmpF loop has been systematically deleted with the aim of dissecting their role in pH and voltage sensitivity. Despite the fact that the study did not allow to completely clarify the contribution of the loops in OmpF activity, it nicely demonstrated that the protein can be mutilated in any of its external loop without appreciably effecting the level of expression, membrane integration and overall protein function (Basle et al, (2004) Protein engineering, design & selection, 9, 665-672).

No experimental evidence that OmpF-derived chimeras can be generated have been reported so far.

Cancer vaccines

The notion that the immune system can recognize and mount a response against tumors was postulated in the late nineteenth century by Co ley who demonstrated that attenuated bacteria or bacterial products injected into tumor-bearing patients in some cases resulted in tumor regression (Coley WB (1893) Am. J. Med. Sci. 105: 487-511). Nearly a century later, it was demonstrated that immunization of mice with mutated tumor cells could induce a protective anti-tumor immune response against non-immunogenic tumor (Van and Boon, (1982) PNAS, 79, 4718-4722). Together, these studies set a foundation for cancer immunotherapy research and demonstrated the therapeutic potential of strategies targeting immune modulation for tumor eradication and protection against tumor recurrences. Therefore, the development of cancer vaccines capable of generating an active tumor-specific immune response serves as a promising venue for cancer therapy.

Probably the best example that illustrates the potential of the immune system to fight cancer is given by Sipuleucel-T, the recently approved vaccine for prostate cancer patients. The vaccine is produced by isolating an individual patient's CD54+ white cells via leukapheresis, exposing the isolated cells ex vivo to PA2024, a protein antigen expressed in over 95% of prostate cancers, and infusing the vaccine back into the patient. Sipuleucel-T therefore consists of personalized primed APCs and of a mixed cell suspension containing also monocytes, macrophages, B and T cells, exposed to activated APCs (Lu C. et al. (2011) Exp.Opin. Biol. Ther. Π, 99-108). Although complicated and expensive to produce the vaccine clearly indicates that, if properly stimulated, the immune system can control tumor growth and progression. Other promising cell-based vaccines are being developed by collecting Tumor Infiltrating Lymphocytes (TILs) from freshly dissected tumors, expanding them upon stimulation with tumor antigens (total tumor extracts or selected tumor antigens) and infusing TILs back into the patients (Restifo et al, (2012) Nature Rev. Immunol. 12, 269- 281). Also this approach has shown to reduce tumor growth and to prolong overall survival.

A more practical way to develop cancer vaccines is to stimulate patient's immune system by injecting into patients specific cancer antigens formulated with proper adjuvants/immune potentiators (Berinstein NL (2007) Vaccine 25S, B72-B88). This approach has the great advantage to avoid the complication of collecting immune cells from each patient and of re-injecting them back after activation and/or amplification.

Several trials are ongoing exploiting this strategy. Among the most promising ones are two peptide-based vaccines, Her2-E75 (Nelipepimut-S) (Mittendorff EA et al, (2014) Annals of Oncology 25: 1735-1742) and EGFRvIII (Rindopepimut) (Del Vecchio CA et al. (2012), Expert Rev. Vaccines H, 133-144) against Her2 -positive breast cancer and glioblastoma, respectively. These vaccines, which are formulated with the immune stimulator GM-CSF, appear to have different mechanisms of action. The first primarily induces cytotoxic CD8+ T cells while the other mostly elicits humoral response.

However, despite demonstrated efficacy in various murine models, cancer vaccines have found little success in the clinic. Although several factors may contribute to the failure of therapeutic cancer vaccines in the clinic, the most important ones are i) the weak immunogenicity of Tumor Associated Antigens (TAAs), ii) central and peripheral immune tolerance to self TAAs, and iii) various immune evasion mechanisms employed by the progressing tumor.

Therefore, the success of therapeutic cancer vaccines may require formulations that induce potent immune responses that overcome immune tolerance to TAAs as well as reverse or inhibit tumor-mediated immune evasion mechanisms.

Human EGFRvIII

Glioblastoma multiforme (GBM) accounts for over 50% of primary brain tumors (Porter KR et al, (2010) Neuro. Oncol. 12, 520-527). Although gliomas often respond to radiotherapy, recurrence occurs in most of the patients with a median time to progression of 7 months (Grossman SA et al, (2004) Semin. Oncol. 31, 635-644). Tumor recurrence almost inevitably leads to death, the median overall survival of patients being only 15-16 months with surgical resection and combination chemo/radiation therapy (Vredenburgh JJ et al. (2007) J. Clin. Oncol. 25, 4722-4729; Stupp R et al, (2005) N. Engl. J. Med. 352, 987-996).

Abnormal cell signaling by EGF receptor has been implicated in numerous cancers. Physiologically, EGF binds to the monomeric form of its receptor and this leads to receptor dimerization and autophosphorylation, which in turn triggers the downstream signal cascades (Salomon DS et al, (1995) Crit. Rev. Oncol. Hematol. 19, 183-232). In the majority of solid tumors, including breast, brain, head-and-neck, non-small-cell lung, renal, ovarian, prostate and colon cancer EGFR is overexpressed (Wong AJ et al, (1992) Proc. Natl Acad. Sci. USA 89, 2965-2969; Gorgoulis V et al. (1992) Anticancer Res. 12, 1 183- 1187; Irish JC et al. (1993) Laryngoscope 103, 42-52; Korc M et al. (1986) Proc. Natl Acad. Sci. USA 83, 5141-5144; Moorghen M et al. (1990) Anticancer Res. 10, 605-611; Ishikawa J et al, (1990) Int. J. Cancer 45, 1018-1021; Zajchowski D et al, (1988) Cancer Res. 48, 7041-7047). EGFR overexpression leads to the enhancement of downstream signaling pathways stimulating growth and invasiveness of cancer cells. Furthermore, cell motility and VEGF-mediated angiogenesis is increased, while cell adhesion requirements are reduced.

In addition to overexpression, there is a naturally occurring variant of the EGF receptor called EGFRvIII. This variant was originally identified in GBM, and is now known to occur in up to 60% of primary GBM. EGFRvIII expression mainly occurs as a consequence of gene rearrangement and amplification even though alternative splicing event has also been implicated. Both gene rearrangement and alternative splicing result in an in- frame 801 base pair deletion of exons 2-7. This deletion gives rise to a truncated receptor that maintains its signal peptide, transmembrane, intracellular kinase and autophosphorylation domains, but lacks a significant portion of the extracellular ligand- binding domain, thus rendering EGFRvIII ligand independent and constitutively active. Other tumors have also been shown to express this variant, including lung, breast, ovarian and prostate cancer, but EGFRvIII is only rarely expressed in normal tissue (Moscatello DK et al, (1995) Cancer Res. 55, 5536-5539).

The in- frame deletion of the extracellular domain of EGFR creates a novel antigenic epitope which is exquisitely tumor-specific (Humphrey et al, (1990) PNAS, 87, 4207). Therefore, the newly generated epitope can be exploited in active and passive immunization. Indeed, a vaccine has been developed (Rindopepimut) which is based on a 14-amino acid peptide (LEEKKGNYVVTDHC) spanning the new epitope conjugated to keyhole limpet hemocyanin (KLH) and formulated with GM-CSF. A number of Phase II clinical trials in EGRFvIII-positive GBM patients have demonstrated that vaccination with Rindopepimut resulted in significantly higher progression- free and overall survival times (Del Vecchio et al, (2012) Exp. Rev. U, 133). The vaccine is highly promising, but it has margins for improvement. For instance, the immune responses could be optimized. In one clinical trial GMTs were below 910 ng/ml with 6 responders out of 14 patients, and in a second trial all 22 enrolled patients responded but after as many as eight vaccine injections. Furthermore, the vaccine production process is relatively complex (three components with a chemical conjugation). Therefore, the development of an easy-to-produce vaccine with enhanced immunogenicity properties is highly desirable.

Human FAT-1

Human FAT gene family is a subclass of the cadherin superfamily, composed of four giant proteins (FAT 1-4) of 500-600 kDa sharing structural similarities from invertebrates to mammals. Human FAT1 is a type 1 transmembrane protein carrying 34 cadherin repeats, five EGF-like repeats, a laminin A-G domain in the extracellular region and a cytoplasmic tail (Dunne, J. et al, (1995) Genomics 30, 207-23). The protein undergoes a proteolytic cleavage by Furin and is predicted to be further cleaved by γ secretase so that its intracellular domain (ICD) can translocate into the nucleus and directly activate cell signaling. FATl ICD also interacts with Ena/VAPS and Scribble, promotes actin-mediated cell migration and inhibits YAPl-mediated cell proliferation (Moeller, M.J. et al, (2004) The EMBO journal, 23, 3769-79). In addition, FATl ICD also interacts with β-catenin and prevents its translocation to the nucleus (Morris, L.G.T. et al., (2013) Nature Genetics 45, 253-61).

Alteration of FATl expression and function has been clearly associated to several human cancers however its role in tumors is controversial.

Several publications provide evidence that in many tumors including oral cancer, astrocytoma, glioblastoma and ductal carcinoma FATl acts as a tumor suppressor (Chosdol, K. et al, (2009) BMC Cancer, 9, 5; Nakaya, K. et al, (2007) Oncogene, 26, 5300-8; Settakorn, J. et al, (2005) Journal of Clinical Pathology, 58, 1249-54). Indeed, it has been shown that FATl depletion leads to a significant stimulation of cell growth and proliferation, while its expression robustly suppresses tumor growth (Morris, L.G.T. et al, (2013) Cell Cycle (Georgetown, Tex.), 12, 1011-2; Morris, L.G.T. et al, (2013) Nature Genetics 45, 253-61). Furthermore, FATl gene falls within the genomic locus 4q35, a highly prevalent region of deletion in many types of human cancers. FATl deletion or loss of heterozygosity have been described in oral cancer (Nakaya, K. et al, (2007) Oncogene, 26, 5300-8), astrocytoma and glioblastoma (Chosdol, K. et al, (2009) BMC Cancer, 9, 5). Finally, in cholangiocarcinoma FATl shows a reduced plasma membrane localization (Settakorn, J. et al, (2005) Journal of Clinical Pathology, 58, 1249-54) and in invasive breast cancer is preferentially down-regulated (Kwaepila, N. et al, (2006) Pathology, 38, 125-31).

By contrast, in other cancers such as acute myeloid leukemia (AML), pre-B acute lymphoblastic leukemia (ALL) and T-ALL (De Bock, C.E. et al, (2012) Leukemia, 26, 918-26) FATl has been described to act as tumor promoter. FATl appears to be upregulated in hepatocarcinoma (Valletta, D. et al, (2014) Carcinogenesis, 35, 1407-15) and leukemia (de Bock et al. 2012). Furthermore, FATl up-regulation is an unfavorable prognostic factor for precursor B-cell acute lymphoblastic leukemia patients (De Bock, C.E. et al., (2012) Leukemia, 26, 918-26). Finally, recent studies in melanoma and pancreatic cancer have demonstrated that FAT1 undergoes an aberrant processing and an altered localization compared to normal cells (De Bock, C.E. et al., (2012) Leukemia, 26, 918-26).

Most recently, (PCT/EP2014/062419(Wo2014/198919) it was discovered that FAT1 is expressed in a large fraction of early and late stage CRCs. Moreover, a murine monoclonal antibody (mAbl98.3) was isolated that selectively binds the surface of different FAT 1 -positive colon cancer cell lines and, upon binding, it is efficiently internalized. mAb 198.3 also showed an intrinsic tumor- inhibiting activity in mouse models of human colon cancer xenografts.

mAb 198.3 was shown to recognize an epitope present on cadherin domain 8 (D8) and cadherin domain 12 (D12), and antibody binding was efficiently abrogated in the presence of the synthetic peptide IQVEATDKDLGPNGHVTYSIVTDTD designed on the basis of the amino acid sequence of D8 domain.

Overall, this study provides the proof of concept that mAb 198.3 could be exploited as novel tools for the treatment of CRC and paves the way to the use of FAT 1 as an anti- CRC vaccine component.

Human MUC-1

Human mucin 1 (MUCl) is a large glycoprotein expressed at relative low level on the apical surface of healthy ductal epithelial cells. A peculiar characteristic of MUCl is the presence of a region of variable number of tandem repeats (VNTR) in its extracellular domain. In particular, the VNTR region is constituted by a 20-amino acid sequence (PDTRPAPGSTAPPAHGVTSA) that repeats an average 20-150 times. In healthy epithelia, the VNTR is highly glycosylated on serines and threonines with long and branched O-linked carbohydrates.

During tumorigenesis, MUC 1 loses polarity and its expression is highly upregulated.

Furthermore, in tumor cells MUCl is aberrantly hypoglycosylated. This is due to the downregulation of glycosyltransferases and the upregulation of sialyltransferases in tumor cells resulting in premature termination of glycosylation and expression of novel carbohydrate structures. This aberrant MUCl expression is found in diverse kinds of human cancers including colon, lung, pancreas, breast, ovarian, prostate, kidney, stomach and head and neck (Ho SB et al, (1993) Cancer Res 53, 641-51; Baldus SE et al, (2002) Histopathology 40, 440-9; Rabassa ME et al. (2006) BMC Cancer 6, 253).

Abnormal expression and reduced glycosylation make MUCl immunogenic.

Cytotoxic T lymphocytes (CTLs) specific for VNTR epitopes have been found in pancreatic, breast and ovarian cancer patients Barnd, D.L. et al. (1989) Proc. Natl. Acad. Sci. USA 86, 7159-7163; Jerome, K.R. et al. (1991) Cancer Res. 5J_: 2908-2916; Ioannides, C.G. et al. (1993) J. Immunol. 151 : 3693-3703). Interestingly, it has been found that peptide epitopes as well as novel truncated glycopeptide epitopes are presented by MHCI and induce MHC-restricted, MUCl -specific CTLs (Vlad, A.M. et al. (2002) J. Exp. Med. 196: 1435-1446). Furthermore, MUCl-specific antibodies have been found in cancer patients and have been associated to a survival benefit in breast, pancreatic, and ovarian cancer patients (vonMensdorff-Pouilly, S. et al. (2000) J. Clin. Oncol. 18: 574-583; Hamanaka, Y. et al. (2003) Int. J. Cancer. 103: 97-100; Pinheiro, S.P. et al. (2010) Cancer Epidemiol. Biomarkers Prev. 19: 1595-1601).

Since its first characterization as a human tumor-associated antigen recognized by T cells MUCl has been studied as a highly promising antigen for passive (adoptive transfer of antibodies or T cells) and active (vaccines) immunotherapy of multiple human cancers (Kimura T. and Finn OJ (2013) Expert Opin. Biol. Ther. 13: 35 -49). However, the numerous clinical trials conducted over the last twenty years to demonstrate the potential of MUCl in immunotherapy were somehow disappointing. Some protection was clearly observed but below the expectations and a MUCl -based cancer immunotherapy still needs substantial optimization before being considered as a valid therapeutic option for the treatment of cancer.

DISCLOSURE OF THE INVENTION

The inventors have found that bacterial Outer Membrane Vesicles (OMVs) loaded with tumor antigens fused to suitable bacterial proteins, are able to induce potent antigen- specific immune response. In particular it was observed that the following bacterial proteins: Factor H Binding Protein (fHbp), Maltose Binding Protein (MBP) and Outer Membrane Protein-F (Omp-F), are able to deliver the tumor antigen to the outer membrane compartment and correctly expose the antigen on the vesicular surface.

Accordingly, the invention provides a fusion protein comprising a bacterial protein selected from Factor H Binding Protein (fHbp), Maltose Binding Protein (MBP), Outer Membrane Protein-F (Omp-F), Neisseria heparin binding antigen (NHBA) and Aggregatibacter actmomycetemcomitans Factor H binding protein (Aa-fHbp), which is fused to one or more copies of a tumor antigen protein.

Preferably, in the fusion protein, the carboxyl end of the bacterial protein is linked to the amino terminus of the tumor antigen, directly or by interposition of a peptide linker. When the bacterial protein is OmpF, the tumor antigen can be inserted in - or replace - any of its external loops.

The fusion protein can contain one or more copies, preferably up to 10 copies, of tumor antigens, optionally separated by a peptide linker. As herein intended, the term "linker" designates any peptide sequence containing from 1 to 20 aa and preferably 2 to 4 aa.

The Factor H Binding Protein (fHbp) according to the invention includes the full- length protein or its domain A.

In preferred embodiments, the bacterial proteins are selected from the group consisting of: fHbp, SEQ ID NO:l; fHbpDomA, SEQ ID NO:2; OMP, SEQ ID NO:3 and MBP, SEQ ID NO:4; NHBA, SEQ ID NO: 109; Aa-fHbp, SEQ ID NO: 110.

Any tumor antigen can be potentially used to construct the fusion protein according to the invention and particularly the following:

(a) cancer-test is antigens including NY-ESO-1, SSX2, SCP1 as well as RAGE,

BAGE, GAGE and MAGE family polypeptides, for example, GAGE-1, GAGE-2, MAGE- 1, MAGE-2, MAGE-3, MAGE -4, MAGE-5, MAGE-6, and MAGE- 12, which can be used, for example, to address melanoma, lung, head and neck, NSCLC, breast, gastrointestinal. and bladder tumours; (b) mutated antigens, including p53, associated with various solid tumours, e.g., colorectal, lung, head and neck cancer; p2 1 ' Ras associated with, e.g., melanoma, pancreatic cancer and colorectal cancer; CDK4, associated with, e.g., melanoma; MUM1 associated with, e.g., melanoma; caspase-8 associated with, e.g., head and neck cancer; CIA 0205 associated with, e.g., bladder cancer; H LA-A2-R 1 701 , beta catenin associated with, e.g., melanoma; TCR associated with, e.g., T-cell non-Hodgkin lymphoma; BCR-abl associated with, e.g., chronic myelogenous leukemia; triosephosphate isomerase; K1A 0205; CDC-27, and LDLR-FUT; (c) over-expressed antigens, including, Galectin 4 associated with, e.g.. colorectal cancer; Galect in 9 associated with, e.g., Hodgkin's disease; proteinase 3 associated with, e.g., chronic myelogenous leukemia; WT 1 associated with, e.g., various leukemias; carbonic anhydrase associated with, e.g., renal cancer; aldolase A associated with, e.g., lung cancer; FRAME associated with, e.g., melanoma; HER-2/neu associated ith, e.g., breast, colon, lung and ovarian cancer; mammaglobin, alpha- fetoprotein associated with, e.g., hepatoma; KSA associated with, e.g., colorectal cancer; gastrin associated with, e.g., pancreatic and gastric cancer; telomerase catalytic protein, MUC-1 associated with, e.g., breast and ovarian cancer; G-250 associated with, e.g., renal cell carcinoma; p53 associated with, e.g., breast, colon cancer; and carcinoembryonic ant igen associated with, e.g., breast cancer, lung cancer, and cancers of the gastrointestinal trac t such as colorectal cancer; (d) shared antigens, including m cl anom a- me I anocyt e differentiation ant igens such as MART-l Melan A; gplOO; MO R; m e 1 a noc t e-st i m u 1 at i ng hormone receptor; tyrosinase; tyrosinase related protein- 1/TRPl and tyrosinase related pt tein-2 TRP2 associated with, e.g., melanoma; (e) prostate associated antigens including PAP, PSA, PSMA, PSl l-P l , PSM-P1 , PSM-P2, associated with e.g., prostate cancer; (f) immunoglobulin idiotypes associated with myeloma and B cell lymphomas. In certain embodiments, the one or more TAA can be selected from pi 5, Horn Nlel-40, II- Ras, E2A- PRL, H4-RET, IGH-IGK, MYL-RAR, Epstein Barr virus antigens, EBNA, human papillomavirus ( HPV ) antigens, including E6 and E7, hepatitis B and C virus antigens, human T-celi iymphotropic viras antigens, TSP-180, p!85erbB2, pi 80erbB-3, c-met, mn- 23H1, TAG-72-4, CA 19-9, CA 72-4, CAM 17.1 , NuMa, K-ras, pi 6, TAGE, PSCA, CT7, 43-9F, 5T4, 791 Tgp72, beta-HCG, BCA225, BTAA, CA 125, CA 15-3 (CA 27.29\BCAA), CA 195, CA 242, CA-50, CAM43, CD68\KP1, CO-029, FGF-5, Ga733 (EpCAM), HTgp- 175, M344, MA-50, MG7-Ag, MOV18, NB/70K, NY-CO- 1 , RCAS1, SDCCAG16, TA- 90 (Mac-2 binding protein/cyclophiiin C-associated protein), TAAL6, TAG72, TLP, TPS.

Preferably the tumor antigen is the entire sequence, a portion of it, or specific immunogenic epitopes of one of the following human proteins: TCTNl (Gene ID: ENSG00000204852), TCTN2 (Gene ID: ENSG00000168778), TCTN3 (Gene ID: ENSGOOOOOl 19977), HIGD2A (Gene ID: ENSG00000146066), HIGD2B (Gene ID: ENSG00000175202), C40RF32 (Gene ID: ENSG00000174749), FAM62A (E-SYTl, Gene ID: ENSG00000139641 ), COLEC11 (Gene ID: ENSGOOOOOl 18004), FSTL5 (Gene ID: ENSG00000168843), FAM82A2 (Gene ID: ENSG00000137824), SCARA5 (Gene ID: ENSG00000168079), VSTM1 (Gene ID: ENSG00000189068), RNF5 (Gene ID: ENSG00000183574), UNQ6126 (Gene ID: gi| 169216088), DPY19L3 (Gene ID: ENSG00000178904), SLC39A10 (gene ID: ENSG00000196950), GPR107 (Gene ID: ENSG00000148358), COL20A1 (Gene ID: ENSG00000101203), GLT25D2 (Gene ID: ENSG00000198756), SYTL3 (Gene ID: ENSG00000164674), DENND1B (Gene ID: ENSG00000162701), C6orf98 (Gene ID: EG: 387079), FAM69B (Gene ID: ENSG00000165716), EMID1 (Gene ID: OTTHUMG00000030824), KLRG2 (GENE ID: ENSG00000188883), ERMP1 (GENE ID: ENSG00000099219), VMOl (Gene ID: ENSG00000182853), C9orf46 (Gene ID: ENSG00000107020), FLJ37107 (Gene ID: ENSG00000177990), YIPF2 (Gene ID: ENSG00000130733), TRYX3 (PRSS58, ENSG00000258223.2), C14orfl35 (Gene ID: ENSG00000126773), ANGPTL7 (Gene ID: ENSG00000171819), TPCN2 (Gene ID: ENSG00000162341), C18orfl9 (Gene ID: ENSG00000177150), OLFML1 (Gene ID: ENSG00000183801), LYPD4 (Gene ID: ENSG00000101203), MEGF8 (Gene ID: ENSG00000105429), FLJ42986 (Gene ID: ENSG00000196460), SLC46A1 (Gene ID: ENSG00000076351), FAM180A (Gene ID: ENSGOOOOOl 89320), CRISP-3 (GENE ID: ENSG00000096006) These tumor antigens are disclosed in WO2010/086162, WO2010/086163, WO2011/051278, WO2011/051276, WO2011/051277, WO2011/051280, WO2011/051271, WO2011/135068, WO2014/198919, all in the applicant's name, the content of which is herein incorporated by reference.

More preferably the tumor antigen is selected from hEGFRvIII, hFAT- 1 and hMUC- 1 , or an immunogenic fragment thereof. In preferred embodiments, the immunogenic fragments are selected from the following peptide sequences: LEEKKG YVVTDH (EGFRvIII), IQVEATDKDLGPNGHVTYSIVTDTD (hFAT-1), GVTSAPDTRPAPGSTAPPAH (hMUC-1).

In preferred invention embodiments the fusion protein is selected from the group consisting of NOs:8 through 25, SEQ ID NOs: l 11 and 112.

The invention also provides an isolated bacterial outer membrane vesicle (OMV) loaded with a fusion protein as above defined. The isolated OMV can contain a fusion protein carrying one species of tumor antigen or a plurality of fusion proteins carrying different tumor antigens.

The OMV can be isolated and purified from Gram-negative bacteria, including species from any of genera Escherichia, Shigella, Neisseria, Moraxella, Bordetella, Borrelia, Brucella, Chlamydia, Haemophilus, Legionella, Pseudomonas, Yersinia, Helicobacter. Salmonella, Vibrio, etc. For example, the vesicles may be from Bordetclla pertussis, Borrelia burgdorferi, Brucella melitensis, Brucella ovis, Chlamydia psittaci. Chlamydia trachomatis, Moraxella catarrhalis, Escherichia coli (including extraintest inal pathogenic strains), Haemophilus influenzae (including non-typeable stains), Legionella pneumophila, Neisseria gonorrhoeae, Neisseria meningitidis, Neisseria lactamica, Pseudomonas aeruginosa, Yersinia enterocoiitica, Helicobacter pylori, Salmonella enterica ( including serovar typhi and typhimurium ). Vibrio ch lerae. Shigella dysenteriae, Shigella flexneri, Shigella boydii or Shigella sonnei.

N. meningitidis OMVs have a proven safety record in humans and so they are a preferred choice. Another useful choice is E. coli vesicles, for example the BL21(DE3) strain (sec Methods).

OMVs are prepared artificially from bacteria, and may be prepared using detergent treatment (e.g. with deoxycholate), or by non-detergent means (e.g. see WO2004 019977). Techniques for forming OMVs include treating bacteria with a bile acid salt detergent (e.g. salts of lithocholic acid, chenodeoxycho! ic acid, ursodeoxycholic acid, deoxycholic acid, cholic acid, ursocholic acid, etc.) at a pH sufficiently high not to precipitate the detergent ( WOO l 91788). Other techniques may be performed substantially in the absence of detergent (WO2004 019977) using techniques such as sonication, homogenisation, microfluidisation, cavitation, osmotic shock, grinding, French press, blending, etc. Methods using no or low detergent can retain useful antigens ( WO2004 019977). Thus a method may use an OMV extraction buffer ith about 0.5% deoxycholate or lower e.g. about 0.2%, about 0.1%, <0.05% or zero.

Bacterial vesicles can conveniently be separated from whole bacterial culture by filtration e.g. through a 0.22 πτ filter. Bacterial filtrates may be clarified by centrifugation, for example high-speed centrifugation {e.g. 20,000 x g for about 2 hours). Another useful process for OMV preparation is described in WO2005 004908 and involves ultrafiltration on crude OMVs, instead of high-speed centrifugation. The process may involve a step of u Itracentrif ugat ion after the ultrafiltration takes place. A simple process for purifying bacterial vesicles is described in WO2011/036562, comprising: (i) a first filtration step in which the vesicles are separated from the bacteria based on their different sizes, with the vesicles passing into the filtrate e.g. using a function (Basle et al., (2004) Protein engineering, design & selection, 9, 665-672).

In a further embodiment, the invention provides an immunogenic composition comprising a bacterial outer membrane vesicle as herein disclosed, together with pharmaceutical acceptable vehicles and excipients. The immunogenic composition may contain a mixture of outer membrane vesicles differing from each other for the type of tumor antigen, for the bacterial protein, or both.

The compositions of the invention for administration to subjects are preferably vaccine composit ions. Vaccines according to the invent ion may either be prophylact ic { e.g. to prevent cancer) or therapeutic {e.g. to treat cancer). Pharmaceutical compositions used as vaccines comprise an immunological ly effective amount of ant igen(s), as well as any other components, as needed. By 'immunologically effective amount', it is meant that the administration of that amount to an individual, either in a single dose or as part of a series, is effective for treatment or prevention. This amount varies depending upon the health and physical condition of the indiv idual to be treated, age, the taxonomic group of indiv idual to be treated {e.g. non-human primate, primate, etc.), the capacity of the individual's immune system to stimulate antibody production, the degree of protection desired, the formulation of the vaccine, the treating doctor's assessment of the medical situation, and other relevant factors. It is expected that the amount will fall in a relatively broad range that can be determined through routine trials. The antigen content of composit ions of the inv ent ion will generally be expressed in terms of the amount of protein per dose. The amount of OMVs in composit ions of the inv ent ion may generally be between 10 and 500 ug, preferably between 25 and 200 tig, and more preferably about 0 ,ug or about 1 00 iig.

Compositions of the inv ent ion may be prepared in various l iquid forms. For example, the composit ions may be prepared as injectables. either as solutions or suspensions. The composition may be prepared for pulmonary administration e.g. by an inhaler, using a fine spray. The composit ion may he prepared for nasal, aural or ocular administration e.g. as spray or drops, and intranasal v esicle vaccines are known in the art. Injectables for intramuscular administration are typical. Injection may be via a needle (e.g. a hypodermic needle), but needle- free injection may alternat ively be used.

The OMVs and the immunogenic compositions according to the invention are conveniently used for the stimulation of an immune response against tumor in a subject in need thereof. Particularly they can be used for the prevention or treatment of different types of tumor, including but not limited to

bronchogenic carcinoma, nasopharyngeal carcinoma, laryngeal carcinoma, small cell and non-smal l cell lung carcinoma, lung adenocarcinoma, hepatocarcinoma, pancreatic carcinoma, bladder carcinoma, colon carcinoma, breast carcinoma, cervical carcinoma, ov arian carcinoma, or lymphocyt ic leukaemias, prostate cancer.

In a preferred embodiment, the isolated bacterial outer membrane vesicles or the immunogenic composition are used in the prevention or treatment of tumors selected from breast, brain, head-and-neck, non-small cell lung, renal, ovarian, kidney, stomach, prostate and colon cancer, oral cancer, astrocytoma, glioblastoma, ductal carcinoma, cholangiocarcinoma, hepatocarcinoma, acute myeloid leukemia, acute lymphoblastic leukemia, melanoma, pancreatic cancer and prostate cancer.

DISCLOSURE OF PREFERRED EMBODIMENTS

In one embodiment, the DNA coding sequence of a selected tumor antigen is fused to the 3 ' end of the gene coding for the full length Neisseria meningitidis Factor H Binding Protein (fHbp) (SEQ ID NO: 85) and such gene fusion is inserted into an appropriate plasmid expression vector in order to drive the expression of a fusion protein constituted by fHbp carrying the foreign antigen at its carboxyl terminus. Because of the presence of the fHbp leader sequence and lipidation site the fusion protein is delivered to the bacterial outer membrane and subsequently incorporated into OMVs. It was surprisingly found that tumor antigens fused to the C-terminus of fHbp are exposed to the surface of E. coli cells and of its derived OMVs.

In another embodiment, the DNA coding sequence of a selected tumor antigen is fused to the 3' end of the gene coding for the domain A of fHbp (fHbpDomA, SEQ ID NO: 86) and such gene fusion is inserted into an appropriate plasmid expression vector in order to drive the expression of a fusion protein constituted by fHbp carrying the foreign antigen at its carboxyl terminus. Because of the presence of the fHbp leader sequence and lipidation site the fusion protein is delivered to the bacterial outer membrane and subsequently incorporated into OMVs. It has been surprisingly found that when expressed in E. coli, fHbpDomA is delivered to the outer membrane and tumor antigens fused to the C-terminus of fHbpDomA are exposed to the surface of E. coli cells and of its derived OMVs. In another embodiment, the DNA coding sequence of a selected tumor antigen is fused to the 3' end of the full length gene coding for the Neisseria meningitidis NHBA (SEQ ID NO: l 13) and such gene fusion is inserted into an appropriate plasmid expression vector in order to drive the expression of a fusion protein constituted by NHBA carrying the foreign antigen at its carboxyl terminus. Because of the presence of the NHBA leader sequence and lipidation site the fusion protein is delivered to the bacterial outer membrane and subsequently incorporated into OMVs. It has been surprisingly found that when expressed in E. coli, NHBA is delivered to the outer membrane and tumor antigens fused to the C-terminus of NHBA are exposed to the surface of E. coli cells and of its derived OMVs.

In another embodiment, the DNA coding sequence of a tumor antigen of interest is fused to the 3' end of the gene coding for the E. coli Maltose Binding Protein (MBP, SEQ ID NO: 100). Such gene fusion is inserted into an appropriate plasmid expression vector in order to drive the expression of a fusion protein constituted by MBP carrying the foreign antigen at its carboxyl terminus. It has surprisingly been found that such fusion protein can be efficiently incorporated into OMVs.

In another embodiment, the DNA coding sequence of a selected tumor antigen is inserted into coding sequence of ompF in correspondence of DNA regions coding for the external loops of the protein. Such gene fusions are inserted into appropriate plasmid expression vectors in order to drive the expression of engineered OmpF in which any of the external loop is replaced with the foreign antigen. It has surprisingly been found that the engineered OmpF are efficiently integrated into the E. coli outer membrane and the tumor antigen is efficiently exposed on the bacterial surface, and that the engineered OmpF is also efficiently incorporated into OMVs.

In a further embodiment, the tumor antigen is the EGFRvIII peptide

LEEK GNYVVTDH. The DNA sequence coding for one or more copies of the EGFRvIII peptide is ligated at the 3' end of gene encoding the full length fHbp in order to generate gene chimeras encoding fHbp with one or more copies of EGFRvIII peptide fused at their C-terminus. A relevant aspect of this invention is that when the gene chimera are expressed in E. coli the hybrid proteins are efficiently delivered to the outer membrane and incorporated into OMVs. Remarkably, OMVs carrying fHbp-EGFRvIII peptide fusions induce potent anti-EGFRvIII peptide immune responses.

In another embodiment, the DNA sequence coding for one or more copies of the

EGFRvIII peptide is ligated at the 3' end of gene encoding fHbpDomA (SEQ ID NO:86) in order to generate gene chimera encoding fHbp with one or more copies of EGFRvIII peptide fused at its C-terminus. A relevant aspect of this invention is that when the gene chimera are expressed in E. coli the hybrid proteins are efficiently delivered to the outer membrane and incorporated into OMVs. Remarkably, OMVs carrying fHbpDomA- EGFRvIII peptide fusions induce potent anti-EGFRvIII peptide immune responses.

In another embodiment, the DNA sequence coding for one or more copies of the EGFRvIII peptide is ligated at the 3' end of gene encoding the full length NHBA (SEQ ID NO: 1 13) in order to generate gene chimeras encoding NHBA with one or more copies of EGFRvIII peptide fused at their C-terminus. A relevant aspect of this invention is that when the gene chimera are expressed in E. coli the hybrid proteins are efficiently delivered to the outer membrane and incorporated into OMVs.

In yet another embodiment, the DNA sequence coding for one copy of the EGFRvIII peptide is ligated at the 3 ' end of gene encoding MBP (SEQ ID NO : 100) in order to generate a gene chimera encoding MBP with one copy of EGFRvIII peptide fused at its C-terminus. A relevant aspect of this invention is that when the gene chimera is expressed in E. coli the hybrid protein is efficiently incorporated into OMVs. Remarkably, OMVs carrying MBP- EGFRvIII peptide fusion induce potent anti-EGFRvIII peptide immune responses.

In a further embodiment, the DNA sequence coding for one or more copies of EGFRvIII peptide is inserted into coding sequence of ompF in correspondence of DNA regions coding for the external loops of the proteins. A relevant aspect of this invention is that when the gene chimeras are expressed in E. coli the hybrid proteins are efficiently delivered to the outer membrane and incorporated into OMVs. Remarkably, OMVs carrying OmpF-EGFRvIII peptide fusions expose the EGFRvIII peptide to the surface of OMVs and are accessible to anti-EGFRvIII antibody binding.

In another embodiment, the foreign antigen is the FAT 1 -derived peptide IQ VEATDKDLGPNGHVT YSI VTDTD . The DNA sequence coding for one or more copies of FAT 1 peptide is ligated at the 3' end of gene encoding the full length fHbp in order to generate gene chimeras encoding fHbp with one or more copies of FAT 1 peptide fused at their C-terminus. A relevant aspect of this invention is that when the gene chimeras are expressed in E. coli the hybrid proteins are efficiently delivered to the outer membrane and incorporated into OMVs. Remarkably, OMVs carrying fHbp-FATl peptide fusions induce potent anti-FATl peptide immune responses.

In yet another embodiment, the DNA sequence coding for one or more copies of the FAT1 peptide is ligated at the 3' end of gene encoding fHbpDomA (SEQ ID NO: 86) in order to generate gene chimeras encoding fHbp with one or more copies of FAT 1 peptide fused at their C-terminus. A relevant aspect of this invention is that when the gene chimeras are expressed in E. coli the hybrid proteins are efficiently delivered to the outer membrane and incorporated into OMVs. Remarkably, OMVs carrying fHbpDomA- F ATI peptide fusions induce potent anti-FATl peptide immune responses.

In another embodiment, the DNA sequence coding for one or more copies of the FAT1 peptide is ligated at the 3' end of the gene encoding MBP (SEQ ID NO: 100) in order to generate gene chimeras encoding fHbp with one or more copies of FAT 1 peptide fused at their C-terminus. A relevant aspect of this invention is that when the gene chimeras are expressed in E. coli the hybrid proteins are efficiently delivered to the periplasm and incorporated into OMVs. Remarkably, OMVs carrying MBP-FAT1 peptide fusions induce potent anti-EGFRvIII peptide immune responses.

In a further embodiment, the foreign antigen is the MUCl -derived peptide

GVTSAPDTRPAPGSTAPPAH. The DNA sequence coding for one or more copies of MUCl peptide is ligated at the 3' end of gene encoding the full length fHbp (SEQ ID NO: 85) in order to generate gene chimeras encoding fHbp with one or more copies of MUCl peptide fused at their C-terminus. A relevant aspect of this invention is that when the gene chimeras are expressed in E. coli the hybrid proteins are efficiently delivered to the outer membrane and incorporated into OMVs. Remarkably, OMVs carrying fHbp-MUCl peptide fusions induce potent anti-MUCl peptide immune responses.

In a further embodiment, the DNA sequence coding for one or more copies of the

MUC1 peptide is ligated at the 3' end of gene encoding fHbpDomA (SEQ ID NO: 86) in order to generate gene chimeras encoding fHbpDomA with one or more copies of MUC1 peptide fused at their C-terminus. A relevant aspect of this invention is that when the gene chimeras are expressed in E. coli the hybrid proteins are efficiently delivered to the outer membrane and incorporated into OMVs. Remarkably, OMVs carrying fHbpDomA-MUC 1 peptide fusions induce potent anti-MUCl peptide immune responses.

In another embodiment, the DNA sequence coding for one or more copies of MUC1 peptide is ligated at the 3 ' end of gene encoding MBP (SEQ ID NO : 100) in order to generate gene chimeras encoding MBP with one or more copies of MUC1 peptide fused at their C- terminus. A relevant aspect of this invention is that when the gene chimeras are expressed in E. coli the hybrid proteins are efficiently delivered to the periplasm and incorporated into OMVs. Remarkably, OMVs carrying MBP-MUC1 peptide fusions induce potent anti- MUCl peptide immune responses.

In another embodiment, the DNA sequence encoding for the Aggregatibacter actinomycetemcomitans factor H binding protein (Aa-fHbp, SEQ ID NO: 116) gene was cloned in E.coli and used to decorate OMVs. A relevant aspect of this invention is that when the gene is cloned in E. coli the protein is efficiently expressed, delivered to the external side of the outer membrane and incorporated into OMVs.

DESCRIPTION OF THE FIGURES

Figure 1. Strategies for decorating bacterial OMVs with the EGFRvlll peptide.

(A) A peptide containing one or three copies of EGFRvlll is fused to the C-terminus of factor H binding protein (fHbp) from N. meningitidis MC58. (B) A peptide containing one or three copies of EGFRvlll is fused to the C-terminus of a truncated form of fHbp, lacking domains B and C (fHbpDoniA). (C) A peptide containing one or three copies of EGFRvIII replaces one of OmpF external loops. (D) A peptide containing one copy of EGFRvIII is fused to the C-terminus of Maltose Binding Protein (MBP) from E. coli K12-MG1655. (E) A peptide containing one or three copies of EGFRvIII is fused to the C-terminus of neisseria heparin binding antigen (NHBA) fromN. meningitidis MC58.

Figure 2. Amino acid and nucleotide sequences of single and triple copy EGFRvIII. (A) Amino acid and nucleotide sequences of single copy EGFRvIII peptide (vIII). (B) Amino acid and nucleotide sequences of triple copy EGFRvIII peptide (vlllx3). Three copies of EGFRvIII are separated and flanked by short linker sequences. To minimize the possibility of recombination between EGFRvIII coding DNA fragments, three different vIII nucleotide sequences were generated, taking advantage of codon degeneracy but considering E.coli BL21 codon usage.

Figure 3. Cloning strategy used to fuse one copy of the EGRFvIII peptide to fHbp full length (A), fHbpDoniA (B) and NHBA (C). To generate pET-fHbp, pET- fHbpDomA and pET-NHBA plasmids the sequence coding for fHbp full length, fHbpDomA or NHBA was amplified by PCR from N. meningitidis MC58 genomic DNA using primers fHbp-ss-F/fHbp R, fHbp-ss-F/fHbp A rev and NHBA-F/NHBA-R, respectively to generate extremities complementary to pET21 expression vector linear DNA, amplified with primers petrev/nohisflag (Table 1), using the polymerase incomplete primer extension (PIPE) cloning method (Klock et al, 2009). To clone one copy of the EGFRvIII peptide in translational fusion to fHbp, fHbpDomA and NHBA, pET-fHbp, pET- fHbpDomA and pET-NHBA plasmids were PCR amplified using primers vlll-single fh for/ vlll-single fh-wt rev, vlll-single fh for/ vlll-single fh-domA rev and NHBA VIII 1XF/ NHBA VIII IX R, respectively. Each couple of primers carries partially complementary 5 ' tails which, when annealed, reconstitute the nucleotide sequence coding for the EGFRvIII peptide. PCR-amplification followed by E. coli HK-100 transformation generated pET-fHbpvIII, pET-fHbpDomAvIII and pET-NHBAvIII plasmids encoding chimeric proteins carrying one copy of EGFRvIII peptide fused to the C-terminus of fHbp, fHbpDomA and NHBA, respectively. LS: leader sequence; LP: lipobox; A: fHbp domain A; B: fHbp domain B; C: fHbp domain C.

Figure 4. Cloning strategy to fuse three copies of the EGRFvIII peptide to fHbp full length, fHbpDomA and NHBA. (A) The DNA fragment coding for the tripeptide vlllx3 was subcloned in pUC plasmid, generating plasmid pUC-vIIIx3. To fuse three copies of EGFRvIII to fHbp full length (B), fHbpDomA (B) and NHBA (C), pET-fHbp, pET- fHbpDomA and pET-NHBA plasmids were PCR-amplified using primers nohisflag/ fHbp Pv2, nohisflag/ fHbp A rev2 and NHBA-vIII-3x-v-f/ NHBA-vIII-3x-v-r, respectively (Table 1), while the vlllx3 insert was PCR-amplified from pUC-vIIIx3 using primers vIII- triple fh-wt for/ vlll-triple rev, vlll-triple fh-domA for /vlll-triple rev and NHBA-vIII-3x- i-f/ NHBA-vIII-3x-i-r , respectively. Finally, the PCR products were used to transform E. coli HK100 cells to allow the recombination of complementary ends, obtaining plasmids pET-fHbpvIIIx3, pET-fHbpDomAvIIIx3 and pET-NHBAvIIIx3. LS: leader sequence; LP: lipobox; A: fHbp domain A; B: fHbp domain B; C: fHbp domain C.

Figure 5. Analysis of EGFRvIII expression in total lysates of

BL21(DE3)/AompA strain transformed with pET-fHbpvIII, pET-fHbpDomAvIII and pET-NHBAvIII. Total extracts from recombinant clones expressing fHbpvIII, fHbpDomAvIII and pET-NHBAvIII in single (A) or in triple (B) copy were separated by SDS-PAGE and analysed by Western blot. Proteins were transferred from the gel to nitrocellulose membrane and analyzed with anti-EGFRvIII rabbit polyclonal antibody. Strains transformed with pET empty vector and pET-fHbp were used as negative controls.

Figure 6. Western blot analysis of fHbpvIII and fHbpDomAvIII expression in OMVs. OMVs were purified by ultrafiltration and ultracentrifugation from the supernatants of recombinant strains transformed with pET-fHbpvIII (A), pET-fHbpvIIIx3 (B), pET-fHbpDomAvIII (C) and pET-fHbpDomAvIIIx3 (D) constructs. OMVs were collected from cultures induced with 1 mM IPTG for 2 hours. Vesicles were subjected to SDS-PAGE and Western Blot analysis using anti-EGFRvIII rabbit polyclonal antibody.

Figure 7. Analysis of surface exposition of EGFRvIII in fHbpvIII, fHbpvIIIx3 (A), NHBAvIII and NHBAvIIIx3 (B) recombinant strains, evaluated by Fluorescence Activating Cell Sorting. EGFRvIII surface expression was evaluated on bacterial cells after 2 h induction with 1 mM IPTG. Cells were stained with anti-vIII polyclonal antibody followed by anti-rabbit-FITC secondary antibody. BL21(DE3)/AompA strain earring pET21 empty plasmid was used as a negative control.

Figure 8. Analysis of vlll-specific IgG induced in mice immunized with fHbpvIIIx3 and fHbpDomAvIIIx3 OMVs. Antigen-specific IgG were measured by ELISA in sera from mice immunized with two (post2) or three (post3) doses of OMVs. As a control, antibody titers from mice immunized with "empty" OMVs were tested. Anti- mouse IgGs conjugated to alkaline phosphatase were used as secondary antibody. OD405 was measured for each serum dilution.

Figure 9. Analysis of antigen expression in total lysates of BL21(DE3)/AompA strain transformed with pET-OmpFvIII constructs. Total extracts from recombinant clones expressing OmpFvIII were separated by SDS-PAGE and analyzed by Coomassie blue staining. Protein expression was induced by addition of 1 mM IPTG to the culture supematants and extracts were prepared by collecting bacteria after 3 h induction.

Figure 10. Analysis of antigen expression in total lysates of BL21(DE3)/AompA strain transformed with pET-OmpFvIIIx3 constructs. Total extracts from recombinant clones expressing OmpFvIIIx3 were separated by SDS-PAGE and analysed by Coomassie blue staining (A) and Western blot (B). Protein expression was induced by addition of 1 mM IPTG to the culture supematants and extracts were prepared by collecting bacteria after 3 h induction. Strain expressing OmpF wt was used as a negative control. A few strains expressing vIII in single copy were used to compare induction levels.

Figure 11. Western blot analysis of OmpFvIII expression in OMVs. OMVs were purified by ultrafiltration and ultracentrifugation from the supematants of recombinant strains transformed with pET-OmpFvIII constructs. OMVs were collected from cultures induced with 1 mM IPTG for 2 hours. OMVs were loaded on SDS-polyacrylamide gel and analyzed by Western blot using rabbit polyclonal antibody against purified synthetic EGFRvIII peptide. Strains transformed with empty pET vector and with pET-OmpF wt plasmid were used as negative controls.

Figure 12. SDS-PAGE analysis of OmpFvIIIx3 expression in OMVs. OMVs were purified by ultrafiltration and ultracentrifugation from the supernatants of recombinant strains transformed with pET-OmpFvIIIx3 constructs. OMVs were collected from cultures induced with 1 mM IPTG for 2 hours. OMVs were loaded on SDS- polyacrylamide gel and analyzed by Coomassie blue staining. Strains transformed with empty pET vector and with pET-OmpF wt plasmid were used as negative controls.

Figure 13. Analysis of surface exposition of vIII in OmpFvIII recombinant strains. Surface expression was evaluated on bacterial cells after 3 h induction with 1 mM IPTG. Cells were stained with 50 μg/ml anti-vIII polyclonal antibody followed by anti- rabbit-FITC secondary antibody. BL21(DE3)/AompA strain overexpressing OmpF wt was used as a negative control. (A) Analysis of vIII surface exposition in strains overexpressing OmpFvIII_L2, OmpFvIII_L4 and OmpFvIII_L6. (B) Analysis of vIII surface exposition in strains overexpressing OmpFvIIIx3_Ll and OmpFvIIIx3_L4.

Figure 14. Cloning strategy used to fuse one copy of the EGRFvIII peptide to MBP. The DNA sequence coding for MBP was amplified by PCR from E. coli K12- MG1655 genomic DNA using primers Ndl-MalEf/Rl-MalEr. The forward primer (Ndl- MalEf) anneals to the 5' end of malE; the reverse primer (Rl-MalEr) anneals to the 3' end of malE (excluding the stop codon) and its 5' tail contains nucleotides 1-17 of the vIII sequence. Then, in order to complete the vIII coding sequence at the 3' of malE, a second and a third polymerase chain reactions were performed using the same forward primer (Ndl-MalEf) but different reverse primers. The reverse primer used in the second amplification step (R2-MalEr) anneals to a region containing the 3' end of malE and nucleotides 1-17 of the vIII sequence; its 5' tail contains nucleotides 18-31 of the vIII sequence. The reverse primer used in the third amplification step (XhI-R3-MalEr) anneals to nucleotides 2-31 of the vIII sequence; its 5' tail contains nucleotides 32-39 of the vIII sequence and the TAA stop codon. With the aim of cloning the fusion gene coding for MBPvIII in pET21 expression vector the PIPE cloning method was used (Klock and Lesley, 2009). Briefly, pET21 was amplified using primers petrev/nohisflag and the fusion gene coding for MBPvIII was reamplified using primers pET21-MalEf/pET21-R3-MalEr, which have 5 ' tails that generate extremities complementary to pET21 expression vector linear DNA. E. coli HK100 transformation leads to the generation of pET-MBPvIII plasmid.

Figure 15. Analysis of antigen expression in total lysates of BL21(DE3)/AompA strain transformed with pET-MBPvIII. Total extracts from bacterial cells after induction with ImM IPTG were separated by SDS-PAGE and analysed by Coomassie blue staining (A) and by Western blot (B). For Western Blot analysis, proteins were transferred from the gel to nitrocellulose membrane and analyzed with anti-MBP mouse monoclonal antibody and anti-EGFRvIII rabbit polyclonal antibody raised against purified synthetic vIII peptide.

Figure 16. Analysis of MBPvIII expression in OMVs. (A) OMVs were purified by ultrafiltration and ultracentrifugation from the supematants of a recombinant strain transformed with pET-MBPvIII. OMVs were collected from cultures induced with 1 mM IPTG for 2 hours. 13 μg of OMVs were loaded on SDS-polyacrylamide gel and analyzed by Coomassie blue staining. (B) The same OMVs preparations were analyzed by Western blot using rabbit polyclonal antibody against purified synthetic vIII peptide. TL, total lysates; OMVs, outer membrane vesicles.

Figure 17. Analysis of EGFRvIII-specific IgG induced in mice immunized with OMVs expressing MBPvIII. Antigen-specific IgGs were measured by ELISA in sera from mice immunized with three doses of OMVs expressing MBPvIII. Anti-mouse IgGs conjugated to alkaline phosphatase were used as secondary antibody. OD405 was measured for each serum dilution.

Figure 18. SDS-PAGE and Western Blot analyses of proteins preparations from EL21(DE3)AompA recombinant strains expressing fHbp-FATl, fHbpDomA- FAT1 and MBP-FAT1 fusions proteins. A) Recombinant clones were grown in LB at 37 C and when the cultures reached OD600 = 0.6, the expression of the fusion proteins was induced by addition of 1 mM IPTG. After 2 hour growth, the equivalent in volume of 1 OD 6 oo of each bacterial culture was collected, centrifuged at 13,000 X g for 5 minutes and pellets were lysed in 200 μΐ of Bacterial protein Extraction Reagent (BPer) (Thermo Scientific, Cat. Number 78266), Lysozime lmg/ml, DNAase 10 U/ml and 0.1 mM MgCb for 30 minutes. The samples were centrifuged at 13.000 x g for 20 minutes to separate the supernatants (soluble fraction) from the pellets (insoluble fraction). The soluble fractions were collected (200 μΐ) and diluted with 100 μΐ of 4x SDS-PAGE loading buffer while the pellets were re-suspended in 300 μΐ of 2X loading buffer. 20 μΐ of each sample were analyzed by SDS-PAGE. B) For Western blot analysis 5 μΐ of the samples prepared as described in A) were loaded onto a 4-12% polyacrilamide gel, transferred to a nitrocellulose filter and fusion proteins carrying FATl peptide visualized by using the FATl -specific mAb 198.3.

Figure 19. SDS-PAGE and Western Blot analyses of OMVs preparations purified from EL21(DE3)AompA recombinant strains expressing fHbp-FATl, fHbpDoniA-FATl and MBP-FAT1 fusions proteins. A) BL2l(OE3)/ ompA (pET- fHbp-FATl), ΒΕ2\(ΌΕ3)Ι AompA (pET-fHbpDomA-F AT 1 ) and BL2\(DE3)/AompA (pET-MBP-FATl) strains were grown in LB and when the cultures reached an OD 6 oo=0.6 ImM IPTG was added. OMVs were purified from the culture supernatants by using ultrafiltration coupled to ultracentrifugation. 10 μg of total proteins of each OMV preparation were analyzed by SDS-PAGE. B) For Western blot analysis the samples prepared as described in A) were loaded onto a 4-12% polyacrilamide gel, transferred to a nitrocellulose filter and fusion proteins carrying FATl peptide visualized by using the FATl -specific mAb 198.3.

Figure 20. Analysis of fHBP-FATl and fHBPDoniA-FATl expression on the surface of E.coli BL21(DE3) AompA strain by FACS.

BL2\(DE3)/AompA (pET-fHbp-F AT 1 ) and BL21(DE3> 1 AompA (pET-fHbpDomA-

FAT1) E.coli strains were grown in 10 ml LB medium at 37°C and when the cultures reached OD 6 oo = 0.6, the expression of the fusion proteins was induced by addition of 1 mM IPTG. After 2 hours growth bacteria were collected by centrifugation and incubated with 50 μΐ of an appropriate dilution of anti-FATl mAbl98.3 or, as negative controls, with PBS containing 1% BSA or with an unrelated mAb. After 1 hour, bacterial cells were washed with PBS containing 1% BSA and subsequently incubated for 30 minutes on ice with goat anti-mouse antibodies added at a final dilution of 1 :200. Finally, after 2 wash steps, pellets were re-suspended in 200 μΐ of PBS and analyzed with FACS CANTOII. Collected data were analyzed with Flow Jo software.

Figure 21. SDS-PAGE analysis of protein preparations from BL21(DE3)AompA recombinant strains expressing fHbpDomA-MUCl and MBP- MUC1 fusion proteins. Recombinant clones were grown in LB at 37°C and when the cultures reached OD600 = 0.6, the expression of the fusion proteins was induced by addition of 1 mM IPTG. After 2 hours growth, the equivalent in volume of 1 OD600 of each bacterial culture was collected, centrifuged at 13,000 X g for 5 minutes and pellets were lysed in 200 μΐ of Bacterial Protein Extraction Reagent (Life Technologies), Lysozime lmg/ml, DNAase 10 U/ml and 0.1 mM MgC12 for 30 minutes. The samples were centrifuged at 13,000 x g for 20 minutes to separate the supernatants (soluble fraction) from the pellets (insoluble fraction). The soluble fractions were collected (200 μΐ) and diluted with 100 μΐ of 4x SDS-PAGE loading buffer while the pellets were re-suspended in 300 μΐ of 2x loading buffer. 20 μΐ of each sample were analyzed by SDS-PAGE. As a negative control, soluble and insoluble protein fractions were also prepared from BL21(DE3)AompA strain carrying pET21 cloning vector.

Figure 22. SDS-PAGE analysis of OMV preparations purified from BL21(DE3)AompA recombinant strains expressing fHbpDomA-MUCl and MBP- MUC1 fusions proteins. BL21(DE3)/AompA (pET-fHbpDomA-MUC 1) and BL21(DE3)/AompA (pET-MBP-MUCl) strains were grown in LB and when the cultures reached an OD600=0.6 ImM IPTG was added. OMVs were purified from the culture supernatants by using ultrafiltration coupled to ultracentrifugation. 10 μg of total proteins of each OMV preparation were analyzed by SDS-PAGE. As a negative control, OMVs were also prepared from BL21(DE3)AompA strain carrying pET21 cloning vector. Figure 23. Antibody titers elicited in mice immunized with engineered OMVs carrying FAT1 and MUC1 fusion proteins. Engineered OMVs (20 μg) were used to i.p. immunize CD1 mice (5 mice per group) three times at two-week intervals and after two weeks from the third immunization, sera were collected and pooled to analyze anti-MUC 1 and anti-FATl antibody titers by ELISA. Plates were coated with the synthetic MUC1 peptide GVTSAPDTRPAPGSTAPPAH (A) and synthetic FAT1 peptide IQVEATDKDLGPNGHVTYSIVTDTD (B) and different dilutions of pooled sera were incubated at 37°C for 2 hours. After three washes in PBST, 100 μΐ of goat anti-mouse antibodies conjugated to alkaline phosphatase (SouthernBiotech, Cat. 1030-04, 1 :2.000 dilution) were added to each well and incubated at 37°C for 1 hour. Finally, after three washes, the phosphatase substrate (4-Nitrophenyl phosphate disodium salt) was added to each well at a concentration of 1 mg/ml (100 μΐ/well) and after 30 minutes incubation at room temperature in the dark, substrate hydrolysis was measured spectrophotometrically at 405 nm.

Figure 24. Expression, compartimentalization and surface localization of Aa- fHbp in BL21 AompA and derived OMVs.

(A) Protein sequence alignment of Nm-fHbp and Aa-fHbp. (B) SDS-PAGE analysis of total cell extracts (TL) and OMVs isolated from BL21 UompA (pET Aa-fHbp). (C) Western blot analysis of OMVs purified from BL21 UompA (pET Aa-fHbp). (D) Assessment of Aa-fHbp localization by FACS analysis. Aa-fHbp detection was carried out using anti-His-tag antibodies. (E) Assessment of Aa-fHbp localization by proteinase K surface shaving

DETAILED DESCRIPTION OF THE INVENTION

Engineered OMVs expressing EGFRvIII peptide

Figure 1 schematizes the different approaches used to decorate bacterial OMVs with the EGFRvIII peptide. Three different strategies were used. Four strategies were designed to deliver the vIII peptide to the membrane compartment of OMVs. The rationale was that since fHbp, NHBA and OmpF are efficiently incorporated into OMVs they could serve as chaperones for the vIII peptide. To fuse the peptide to fHbp and NHBA, a DNA fragment encoding one copy (vIII) or three copies (vlllx3) of LEEKKGNYVVTDH vIII peptide (Fig. 2) was cloned at the 3 ' end of the full length fHbp gene, full length NHBA gene and at the 3' end of the sequence coding for fHbp lacking domains B and C (fHbpDomA), thus generating chimeric proteins carrying the vIII peptide at their C-terminus (Figure 1A, IB, IE). To fuse the vIII peptide to OmpF, the DNA coding for OmpF extracellular loops was replaced with synthetic DNA coding for one or three copies of the vIII peptide (Figure 1C). Finally, the fifth strategy was designed to deliver the vIII peptide into the lumen of OMVs. To this aim, the synthetic DNA coding for one copy of the vIII peptide was fused to the 3 ' end of malE, the gene coding for MBP, to create an in frame C-terminal fusion (Figure ID).

The detailed description of the construction of the protein chimeras and the preparation of the engineered OMVs decorated with the different protein fusions are reported below.

OMV engineering with fHBPvIII

Construction of pET-flibpvIII and pET-fHbpDomAvIII plasmids

In an attempt to express and deliver the EGFRvIII peptide to the membrane compartment of E. coli OMVs the Neisseria meningitidis fHbp lipoprotein was used as a carrier. To this purpose one or three copies of the EGFRvIII peptide were fused to the 3' end of either the full length fHbp or fHbpDomA. The first step to achieve this was to amplify the sequence coding for full length fHbp (fHbpFL) and fHbpDomA and to clone the amplified sequences into pET21 plasmid. fHbpFL and fHbpDomA coding sequences (SEQ ID NOs:85 and 86) were amplified by PCR from N meningitidis MC58 genomic DNA using primers fHbp-ss-F/fHbp R and fHbp-ss-F/fHbp A rev, respectively (Table 1 and Figure 3), to generate extremities complementary to pET21 expression vector linear DNA, amplified with primers petrev/nohisflag (Table 1), using the polymerase incomplete primer extension (PIPE) cloning method (Klock H.E. and Lesley S.A (2009) Methods Mol. Biol. 498, 91-103). Primer fHbp-ss-F was designed to include at the 5' end of the amplified products the sequence coding for the leader peptide for secretion and the lipobox. PCR products were then mixed together and used to transform E. coli HK-100 strain, generating plasmids pET-fHbp and pET-fHbpDomA. The correctness of the cloning of fHbp and fHbpDomA was verified by sequence analysis (SEQ ID NOs:85 and 86). To clone the EGFRvIII peptide as translational fusion to the C-terminus of fHbp and fHbpDomA, the polymerase incomplete primer extension (PIPE) cloning method was used. In particular, to fuse one copy of the EGFRvIII peptide, pET-fHbp and pET-fHbpDomA plasmids were PCR amplified using primers vlll-single fh for/ vlll-single fh-wt rev and vlll-single fh for/ vlll-single fh-domA rev respectively (Table 1). Each couple of primers carries partially complementary 5 ' tails which when annealed reconstitute the nucleotide sequence coding for the EGFRvIII peptide. PCR-amplification followed by E. coli HK-100 transformation generated pET-fHbpvIII and pET-fHbpDomAvIII plasmids encoding chimeric proteins carrying one copy of EGFRvIII peptide fused to the C-terminus of fHbp and fHbpDomA, respectively (Figure 3). The correctness of the cloning of fHbp-vIII and fHbpDomA-vIII fusions was verified by sequence analysis (SEQ ID NOs: 87 and 88). To fuse three copies of the EGFRvIII peptide to the C-termini of full length fHbp and fHbpDomA, the strategy schematized in Figure 4 was used. In brief, a DNA fragment, named vlllx3, coding for three copies of vIII separated by the GlySer dipeptide and carrying single stranded 3' protruding ends complementary to the protruding single stranded 3' ends generated by EcoRI and BamHI restriction sites was chemically synthesized (Figure 2) and cloned in pUC plasmid cut with EcoRI and BamHI. The synthetic DNA and the linear pUC were in vitro ligated and the ligation mixture was used to transform E. coli HK-100, thus generating plasmid pUC-vIIIx3. Subsequently, pET-fHbp and pET-fHbpDomA plasmids were PCR amplified using primers nohisflag/ fHbp R2 and nohisflag/ fHbp A rev2, respectively (Table 1), while the vlllx3 insert was PCR-amp lifted from pUC-vIIIx3 using primers vIII- triple fh-wt for/ vlll-triple rev and vlll-triple fh-domA for /vlll-triple rev, respectively (Table 1). Finally, the PCR products were mixed together and used to transform HK-100 competent cells, obtaining plasmids pET-fHbpvIIIx3 and pET-fHbpDomAvIIIx3. The correctness of the cloning of fHbp-vIIIx3 and fHbpDomA-vIIIx3 fusions was verified by sequence analysis (SEQ ID NOs: 89 and 90).

Expression of the fHbpvIII and fHbpDomAvIII heterologous proteins in E. coli BL21(DE3)/AompA strain

The four recombinant plasmids encoding fHbp and fHbpDoniA fused to one copy and three copies of the EGFRvIII peptide were used to transform E. coli strain BL2l(OE3)/ AompA. Four recombinant strains were obtained: BL21 (DE3)/zfom/¾4(pETfHbpFL-vIII), BL2 l(DE3)/zfom/¾4(pETfHbpFLvIIIx3),

BL21(DE3)/Jom/^(pETfHbpDomAvIII) and BL2\(DE3)/AompA

(pETfHbpDomAvIIIx3). Each strain was grown in LB medium and when cultures reached an OD 6 oo value of 0.6, IPTG was added at 1 mM final concentration. After two additional hours of growth at 37°C, cells were collected and total protein extracts were analyzed by Western Blot. To this aim, 25 μg of total proteins from each strain were separated by SDS- PAGE and proteins were transferred to nitrocellulose filters. The filters were blocked overnight at 4°C by agitation in blocking solution (10% skimmed milk and 0.05% Tween in PBS), followed by incubation for 90 minutes at 37°C with a 1 : 1 ,000 dilution of rabbit anti-vIII polyclonal antibodies in 3% skimmed milk and 0.05% Tween in PBS. After 3 washing steps in PBS-Tween, the filters were incubated in a 1 :2,000 dilution of peroxidase- conjugated anti-rabbit immunoglobulin (Dako) in 3% skimmed milk and 0.05% Tween in PBS for 1 hour, and after 3 washing steps, bound conjugated IgGs were detected using the Super Signal West Pico chemo- luminescent substrate (Pierce). As shown in Figure 5A and Figure 5B, all four fusion proteins were found expressed

in the cell lysates. No immune reactive bands were detected in total lysates from E. coli cells carrying either pET-fHbp expressing full length fHbp or empty pET21 plasmid.

Analysis of fHbpvIII and fHbpDomAvIII expression in OMVs

Having demonstrated that the four fusion proteins were expressed in E. coli

BL21(DE3 JompA strain, the presence of the antigens in the OMV fraction was analyzed. The four recombinant strains BL21(DE3)/ziom/?^(pETfHbpFL-vIII), BL21(DE3)/zfom/¾4(pETfHbpFLvIIIx3), BL2\(DE3)/AompA (pETfHbpDomAvIII) and BL21(DE3)/z/om/? ^ 4(pETfHbpDomAvIIIx3) were grown in LB medium and when the cultures reached an OD 6 oo value of 0.6, IPTG was added at 1 mM final concentration. After two additional hours of growth at 37°C, vesicles were purified from culture supematants by using ultrafiltration coupled to ultracentrifugation. More specifically, OMVs were collected from culture supematants by filtration through a 0.22 μιη pore size filter (Millipore) and by high-speed centrifugation (200.000 x g for 2 hours). Pellets containing OMVs were finally suspended in PBS. The presence of the antigens in OMVs preparations was verified by Western Blot analysis as described in the previous section (Figure 6). Data indicate that both fHbp and fHbpDomA carrying either one or three copies of vIII peptide were incorporated into OMVs. The presence of high molecular weight bands seem to suggest that when fused to vIII peptide fHbp and fHbpDomA can also form stable homo- oligomers which did not easily dissociate even if OMV preparations were treated at 100°C in the presence of 1% SDS and reducing agent (Figure 6B, C, D).

(continued) petrev CATATGTATATCTCCTTCTTAAAGTTAAAC nohisflag TAACATCACCATCACCATCACGATTACAAAGA

ggaattccatatgAAAATAAAAACAGGTGCACGCATC

Ndl-MalEf

Rl-MalEr ccttttttttcttccagCTTGGTGATACGAGTCTGCG

taacaacgtagttaCCTTTTTTTTCTTCCAGCTTGGTGA

R2-MalEr

Xhl-R3- ccccgctcgagttagtggtcggTAACAACGTAGTTACCTTTTTTTTCTTCC MalEr

pET21- ggagatatacatatgAAAATAAAAACAGGTGCACGCATC

MalEf

pET21-R3- gtgatggtgatgttagtggtcggTAACAACGTAGTTACCTTTTTTTTCTTCC MalEr

ggaattccatatgATGAAGCGCAATATTCTGGC

Ndl-OmpFf

Xhl-OmpFr ccccgctcgagTTAGAACTGGTAAACGATACCCAC

pET210mpF

ggagatatacatatgATGAAGCGCAATATTCTGGC

f

pET210mpF

gtgatggtgatgttaGAACTGGTAAACGATACCCAC

r

LooplVf aaaggtaactacgttgttaccgaccacGGCGACATGACCTATGCCC

aacgtagttaccttttttttcttccagAAAATAATGCAGACCAACAGCTTTACC

LooplVr

G

aaaggtaactacgttgttaccgaccacGGTAACAAAACGCGTCTGGC

Loop2Vf

Loop2Vr aacgtagttaccttttttttcttccagGTTACCCTGGAAGTTATATTCCCAC

Loop4Vf aaaggtaactacgttgttaccgaccacAACGGCGACGGTGTTGGC

Loop4Vr aacgtagttaccttttttttcttccagGTTTTTACCCAGGTACTGAACAGC

Loop6Vf aaaggtaactacgttgttaccgaccacGCCAACAAAACGCAAGACGTTCTG

Loop6Vr aacgtagttaccttttttttcttccagCGTAGCGTTACGGGTTTCACC

LoopWf aaaggtaactacgttgttaccgaccacGTGAACTACTTTGAAGTGGGCG

LoopWr aacgtagttaccttttttttcttccagTTTCGCTTTAGATTTGGTGTAAGCGAT

(continued) Loop8Vf aaaggtaactacgttgttaccgaccacACCGTTGCTGTGGGTATCGTT

Loo 8Vr aacgtagttaccttttttttcttccagCTGGTTGATGATGTAGTCAACATAGG

Ll_3x_If GGTCTGCATTATTTTGGTTCCCTGGAAGAAAAGAAGGG

Ll_3x_Ir GGTCATGTCGCCGCCGGAATGGTCGGTAACCAC

L4_3x_If GGGTAAAAACGGTTCCCTGGAAGAAAAGAAGGG

L4_3x_Ir CCGTCGCCGTTGCCGGAATGGTCGGTAACCAC

Ll_3x_Vf GGCGACATGACCTATGCCCG

Ll_3x_Vr AAAATAATGCAGACCAACAGCTTTACCG

L4_3x_Vf AACGGCGACGGTGTTGGCGG

L4_3x_Vr GTTTTTACCCAGGTACTGAACAGC

NHBA-F GGAGATATACATATG 1 1 1 AAACGCAGCGTAATC

NHBA-R GTGATGGTGATGTTATCAATCCTGCTCTTTTTTG

NHBA VIII

aac gta gtt acc ttt ttt ttc ttc cag ATC CTG CTC TTT TTT GCC GG IX R

NHBA VIII Aaaggtaactacgttgttaccgaccac TAA CAT CAC CAT CAC CAT CAC IX F GAT TAC AAA GA

NHBA-VIII- CCGGCAAAAAAGAGCAGGAT ggttccctggaagaaaagaaggg

3X-i-F

NHBA-VIII- GTGATGGTGATGTTA gccggaatggtcggtaaccac

3X-i-R

NHBA-VIII-

ATCCTGCTCTTTTTTGCCGG

3x v-R

NHBA-VIII-

TAACATCACCATCACCATCACGATTACAAAGA

3x-v-f

Ag_fHbp-F Gaaggagatatacat ATG GTT TAC CCT GTT ATA ACG

Ag fHbp-R ATGGTGATGGTGATGTTCTTT TTTACCTGCCAA ACC

pET 2-R CATATGTATATCTCCTTCTTAAAGTTAAACaaaattatttc

pET HIS-F catcaccatcaccatcacTAAGATTACAAAGACGATGATGACAAGtga

Analysis of cellular localization of fllbpvIII and fHbpvIIIx3

The localization of recombinant fHbpvIII and fHbpvIIIx3 fusion proteins was evaluated by flow cytometry. To this aim, recombinant E. coli strains BL21(DE3)/Jom/¾4(pETfHbpFL-vIII), BL21(DE3)/Jom/¾4(pETfHbpFLvIIIx3) and E. coli BL21(DE3)/ziom/¾4(pET21), as negative control, were grown at 37°C under agitation. When cultures reached an OD 6 oo value of 0.6, IPTG was added at a final concentration of 1 mM and bacteria were grown for 2 additional hours. Subsequently, bacteria cells corresponding to those contained in 1 ml culture at OD 6 oo = 1 were collected by centrifugation at 13,000 x g for 5 minutes and pellets were re-suspended in 50 ml of PBS containing 1% BSA. 50 μΐ of cell suspensions were incubated with 50 μΐ of an appropriate dilution of anti-vIII rabbit polyclonal antibodies raised against the vIII peptide conjugated to Keyhole Limpet Hemocyanin (KLH) or with 50 μΐ of PBS containing 1% BSA as negative control. After 1 hour, 100 μΐ of PBS containing 1% BSA were added and the suspensions were centrifuged at 3,000 x g for 10 minutes and supernatants discarded. Pellets were washed with 200 μΐ of PBS containing 1% BSA and bacteria subsequently incubated for 30 minutes on ice with goat anti-rabbit antibodies (Alexa flour488, Life Technology) added at a final dilution of 1 :2,000. Finally, After 2 wash steps, pellets were re-suspended in 200 μΐ of PBS and analyzed with FACS CANTOII evaluating collected data with Flow Jo software. As shown in Figure 7A, in the presence of anti-vIII antibodies, a clear shift in fluorescence intensity was observed in a substantial fraction of bacterial cells expressing both fHbpvIII and fHbpvIIIx3. No difference in fluorescence intensity was observed when E. coli BL2l(OE3)/ ompA were incubated with anti-vIII antibodies. These data indicate not only that fHbvIII and fHbpvIIIx3 were expressed in E. coli BL2l(OE3)/ AompA and localized in OMVs, but also that the fused proteins associated to the outer membrane, with their C-terminal portion carrying the vIII peptide exposed to the extracellular compartment.

Engineered OMVs carrying recombinant fHbpvIII and fHbpDomAvIII fusion proteins induce high antibodies titers in immunized mice

To test whether OMVs purified from fHbpvIIIx3 and fHbpDomAvIIIx3 recombinant strains were capable of inducing vlll-specific antibody responses, CD l mice were i.p. immunized three times at two-week intervals with 10 μg of OMVs formulated in Alum. Blood samples were collected nine days after second dose (post2) and seven days after third dose (post3) administration and anti-vIII IgGs were detected by using plates coated in each well with 0.5 μg of synthetic vIII peptide conjugated to Keyhole limpet hemocyanin (vIII-KHL). Serum deriving from mice immunized with empty OMVs was used as negative control. More specifically, coating was carried out by incubating plates overnight at 4°C with 100 μΐ of conjugated peptide (5 μg/ml). Subsequently, wells were washed three times with PBST (0.05% Tween 20 in PBS, pH 7.4), incubated with 100 μΐ of 1% BSA in PBS for 1 h at room temperature and washed again three times with PBST. Serial dilutions of serum samples in PBST containing 1% BSA were added to the plates, incubated 2 h at 37°C, and washed three times with PBST. Then 100 μΐ/well of 1 :2.000 diluted, alkaline phosphatase-conjugated goat anti-mouse IgGs were added and left for 2 h at 37°C. After triple PBST wash, bound alkaline phosphatase-conjugated antibodies were detected by adding ΙΟΟμΙ/well of 3mg/ml para-nitrophenyl-phosphate disodium hexahydrate (Sigma- Aldrich) in 1M diethanolamine buffer (pH 9.8). After 10 minute incubation at room temperature, the reaction was stopped with 100 μΐ 7% EDTA and substrate hydrolysis was analyzed at 405 nm in a microplate spectrophotometer.

As shown in Figure 8, OMVs carrying fHbpvIIIx3 and fHbpDomAvIIIx3 fusion proteins were able to induce high anti-vIII IgG titers in mice. In particular, OMV carrying fHbpDomA3xvIII induced maximum antibody responses even after only two immunizations.

OMV engineering with NHBAvIII

Construction of pET-NHBAvIII plasmids

In order to express and deliver the EGFRvIII peptide to the membrane compartment of E. coli OMVs the Neisseria meningitidis NHBA lipoprotein was used as a carrier. To this purpose one or three copies of the EGFRvIII peptide were fused to the 3' end of the full length NHBA. The first step to achieve this was to amplify the sequence coding for full length NHBA and to clone the amplified sequences into pET21 plasmid. NHBA coding sequences (SEQ ID NO: 113) were amplified by PCR from N. meningitidis MC58 genomic DNA using primers NHBA-F/NHBA-R (Table 1 and Figure 3C), to generate extremities complementary to pET21 expression vector linear DNA, amplified with primers petrev/nohisflag (Table 1), using the polymerase incomplete primer extension (PIPE) cloning method (Klock H.E. and Lesley S.A (2009) Methods Mol. Biol. 498, 91-103). Primer NHBA-F was designed to include at the 5' end of the amplified products the sequence coding for the leader peptide for secretion and the lipobox. PCR products were then mixed together and used to transform E. coli HK-100 strain, generating plasmids pET- NHBA. The correctness of the cloning of NHBA was verified by sequence analysis (SEQ ID NO: 113). To clone the EGFRvIII peptide as translational fusion to the C-terminus of NHBA the polymerase incomplete primer extension (PIPE) cloning method was used. In particular, to fuse one copy of the EGFRvIII peptide, pET-NHBA plasmid was PCR amplified using primers NHBA VIII IXF/ NHBA VIII IX R (Table 1). These primers carries partially complementary 5' tails which when annealed reconstitute the nucleotide sequence coding for the EGFRvIII peptide. PCR-amplification followed by E. coli HK-100 transformation generated pET-NHBAvIII plasmid encoding chimeric proteins carrying one copy of EGFRvIII peptide fused to the C-terminus of NHBA (Figure 3C). The correctness of the cloning of NHBA-vIII fusion was verified by sequence analysis (SEQ ID NO: 114). To fuse three copies of the EGFRvIII peptide to the C-termini of full length NHBA, was used the strategy schematized in Figure 4. In brief, a DNA fragment, named vlllx3, coding for three copies of vIII separated by the GlySer dipeptide and carrying single stranded 3 ' protruding ends complementary to the protruding single stranded 3' ends generated by EcoRI and BamHI restriction sites was chemically synthesized and cloned in pUC plasmid cut with EcoRI and BamHI. The synthetic DNA and the linear pUC were in vitro ligated and the ligation mixture was used to transform E. coli HK-100, thus generating plasmid pUC-vIIIx3 (Figure 4A). Subsequently, pET-NHBA plasmid was PCR amplified using primers NHBA-vIII-3x-v-f/ NHBA-vIII-3x-v-r (Table 1), while the vlllx3 insert was PCR- amplified from pUC-vIIIx3 using primers NHBA- vIII-3x-i-f/ NHBA- vIII-3x-i-r (Table 1). Finally, the PCR products were mixed together and used to transform HK-100 competent cells, obtaining plasmids pET-NHBAvIIIx3 Figure 4D). The correctness of the cloning of NHBA-vIIIx3 fusion was verified by sequence analysis (SEQ ID NO: l 15).

Expression of the NHBAvIII heterologous protein in E. coli BL21 (DE3)/ ompA strain

The two recombinant plasmids encoding NHBA fused to one copy and three copies of the EGFRvIII peptide were used to transform E. coli strain BL2l(OE3)/ AompA. Two recombinant strains were obtained: BL21(DE3)/ziom/¾4(pET-NHBA-vIII) and BL21(DE3)/ziom/¾4(pET-NHBAvIIIx3). Each strain was grown in LB medium and when cultures reached an OD 6 oo value of 0.6, IPTG was added at 1 mM final concentration. After two additional hours of growth at 37°C, cells were collected and total protein extracts were analyzed by Western Blot. To this aim, 25 μg of total proteins from each strain were separated by SDS-PAGE and proteins were transferred to nitrocellulose filters. The filters were blocked overnight at 4°C by agitation in blocking solution (10% skimmed milk and 0.05% Tween in PBS), followed by incubation for 90 minutes at 37°C with a 1 : 1 ,000 dilution of rabbit anti-vIII polyclonal antibodies in 3% skimmed milk and 0.05% Tween in PBS. After 3 washing steps in PBS-Tween, the filters were incubated in a 1 :2,000 dilution of peroxidase-conjugated anti-rabbit immunoglobulin (Dako) in 3% skimmed milk and 0.05%) Tween in PBS for 1 hour, and after 3 washing steps, bound conjugated IgGs were detected using the Super Signal West Pico chemo-luminescent substrate (Pierce). As shown in Figure 5A and Figure 5B, both fusion proteins were found expressed in the cell lysates. No immune reactive bands were detected in total lysates from E. coli cells carrying either pET-NHBA expressing full length NHBA or empty pET21 plasmid.

Analysis of cellular localization of NHBAvIII and NHBAvIIIx3

The localization of recombinant NHBAvIII and NHBAvIIIx3 fusion proteins was evaluated by flow cytometry. To this aim, recombinant E. coli strains BL21(DE3)/zfom/¾4(pET-NHBAvIII), BL21(DE3)/zfom/¾4(pET-NHBAvIIIx3) and E. coli BL21(DE3)/ziom/¾4(pET21), as negative control, were grown at 37°C under agitation. When cultures reached an OD 6 oo value of 0.6, IPTG was added at a final concentration of 1 mM and bacteria were grown for 2 additional hours. Subsequently, bacteria cells corresponding to those contained in 1 ml culture at OD 6 oo = 1 were collected by centrifugation at 13,000 x g for 5 minutes and pellets were re-suspended in 50 ml of PBS containing 1% BSA. 50 μΐ of cell suspensions were incubated with 50 μΐ of an appropriate dilution of anti-vIII rabbit polyclonal antibodies raised against the vIII peptide conjugated to Keyhole Limpet Hemocyanin (KLH) or with 50 μΐ of PBS containing 1% BSA as negative control. After 1 hour, 100 μΐ of PBS containing 1% BSA were added and the suspensions were centrifuged at 3,000 x g for 10 minutes and supernatants discarded. Pellets were washed with 200 μΐ of PBS containing 1% BSA and bacteria subsequently incubated for 30 minutes on ice with goat anti-rabbit antibodies (Alexa flour488, Life Technology) added at a final dilution of 1 :2,000. Finally, After 2 wash steps, pellets were re-suspended in 200 μΐ of PBS and analyzed with FACS CANTOII evaluating collected data with Flow Jo software. As shown in Figure 7B, in the presence of anti-vIII antibodies, a clear shift in fluorescence intensity was observed in a substantial fraction of bacterial cells expressing both NHBAvIII and NHBAvIIIx3. No difference in fluorescence intensity was observed when E. coli BL2l(OE3)/ ompA were incubated with anti-vIII antibodies. These data indicate not only that NHBAvIII and NHBAvIIIx3 were expressed in E. coli BL2l(OE3)/ ompA and that the fused proteins are associated to the outer membrane with their C-terminal portion carrying the vIII peptide exposed to the extracellular compartment.

OMV engineering with OmpFvIII

Construction of pET-OmpFvIII plasmids

With the aim of expressing EGFRvIII peptide on the OMVs surface, the fusion of vIII peptide to Outer Membrane Protein F (OmpF) was attempted. OmpF is a protein embedded in the outer membrane with eight loops (L) exposed to the extracellular compartment (Fig. 1C). Therefore, if extracellular loops are removed and replaced with vIII peptide, the peptide should theoretically be delivered to the membrane surface. To test this hypothesis, the coding sequence of the extracellular loops LI, L2, L4, L6, L7, L8 was substituted with the nucleotide sequence coding for the 13 amino acid EGFRvIII peptide (Fig. 2A), thus generating six constructs, each construct having one loop substituted. The first step to obtain the six engineered OmpF constructs was to clone the OmpF gene into pET21 plasmid. To this aim, the entire OmpF coding sequence was amplified from E. coli K12-MG1655 genomic DNA using primers Ndl-OmpFf/XhI-OmpFr (Table 1). Then, the resulting fragment was re-amplified with primers pET21 OmpFf/pET21 OmpFr (Table 1 ) to make the extremities complementary to pET21 expression vector which was amplified with primers petrev/nohisfiag (Table 1), using the polymerase incomplete primer extension (PIPE) method. The two linear DNAs were mixed together and used to transform the highly competent E. coli strain strain HK-100, to obtain the pET-OmpF plasmid. The correctness of the cloning of the ompF gene in pET21 was verified by sequence analysis (SEQ ID NO:91).

To substitute loops LI, L2, L4, L6, L7 and L8 with EGFRvIII, the polymerase incomplete primer extension (PIPE) cloning method was used. pET-OmpF was amplified by PCR using the primer couples LooplVf/LooplVr, Loop2Vf/Loop2Vr, Loop4Vf/Loop4Vr, Loop6Vf/Loop6Vr, Loop7Vf/Loop7Vr and Loop8Vf/Loop8Vr, respectively (Table 1). Each couple was designed to anneal to the flanking regions of a selected loop and to carry complementary tails that when annealed reconstituted the sequence coding for vIII peptide. PCR-amplification followed by E. coli HK100 transformation resulted in replacement of the loop with the DNA sequence coding for EGFRvIII. Plasmids pET-OmpFvIII Ll, pET-OmpFvIII_L2, pET-OmpFvIII_L4, pET- OmpFvIII_L6, pET-OmpFvIII_L7 and pET-OmpFvIII_L8 were obtained. The correctness of the replacement of each loop coding sequence with the small fragment coding for the vIII peptide was verified by sequence analysis (SEQ ID NOs:92, 94, 95, 97, 98 and 99).

In an attempt to maximize the exposure of the peptide to the extracellular milieu two other constructs were generated, in which LI or L4 were substituted with three copies of EGFRvIII (Fig. 2B). To obtain pET-OmpFvIIIx3_Ll, the vlllx3 coding sequence was amplified frompUC-vIIIx3 (see previous section) using primers Ll_3x_If/Ll_3x_Ir (Table 1). The primers carried 5' ends complementary to the sequences preceding and following the region coding for LI . In parallel, pET-OmpF plasmid was amplified using primers Ll_3x_Vf/Ll_3x_Vr. The two DNA fragments were mixed together and used to transform E. coli HK-100 competent cells, thus obtaining a clone carrying plasmid pET- OmpFvIIIx3_Ll . A similar strategy was used to obtain pET-OmpFvIIIx3_L4 plasmid. The vlllx3 coding sequence was amplified from pUC-vIIIx3 (see previous section) using primers L4_3x_If/L4_3x_Ir (Table 1). The primers carried 5' ends complementary to the sequences preceding and following the region coding for L4. In parallel, pET-OmpF plasmid was amplified using primers L4_3x_Vf/L4_3x_Vr (Table 1). The two DNA fragments were mixed together and used to transform E. coli HK-100 competent cells, thus obtaining a clone carrying plasmid pET-OmpFvIIIx3_L4. The correctness of the replacement of Loop 1 and Loop4 with the small fragment coding for three copies of the vIII peptide was verified by sequence analysis (SEQ ID NOs: 93 and 96).

Expression of OmpFvIII and OmpFvIIIx3 heterologous proteins in E. coli BL21(DE3)/AompA

The eight plasmids: pET-OmpFvIII Ll, pET-OmpFvIII_L2, pET-OmpFvIII_L4, pET-OmpFvIII_L6, pET-OmpFvIII_L7, pET-OmpFvIII_L8, pET-OmpFvIIIx3_Ll and pET-OmpFvIIIx3_L4 were used to transform E. coli strain BL21(DE3)/AompA. Expression of engineered OmpF carrying the vIII peptide in each loop was analyzed in bacterial lysates by SDS-PAGE and Western Blot. E. coli BL21(DE3)/AompA strains carrying each plasmid encoding engineered OmpF were grown in LB medium and when cultures reached an OD 6 oo value of 0.6, IPTG was added at 1 mM final concentration. After two additional hours of growth at 37°C, cells were collected and 25 μg of total protein extracts were analyzed by SDS-PAGE. As shown in Figure 9 and Figure 10A, a protein band with the electrophoretic mobility similar to OmpF accumulated in all protein extracts prepared after IPTG induction. In the case of the two strains carrying plasmid pET- OmpFvIIIx3_Ll and plasmid pET-OmpFvIIIx3_L4, respectively, the bands had an electrophoretic mobility slightly higher than OmpF, in line with the fact that three copies of vIII were used to replace Loop 1 and Loop 4, respectively. Western Blot of total cell extracts was also carried out to confirm the expression of the engineered OmpF proteins in strains carrying plasmids pET-OmpFvIII_L2, pET-OmpFvIII_L4, pET-OmpFvIII_L6, pET-OmpFvIIIx3_Ll and pET-OmpFvIIIx3_L4. To this aim, 13 μg of total proteins from each strain were separated by SDS-PAGE and proteins were transferred to nitrocellulose filters. The filters were blocked overnight at 4°C by agitation in blocking solution (10% skimmed milk and 0.05% Tween in PBS), followed by incubation for 90 minutes at 37 C with a 1 : 1.000 dilution of rabbit anti-vIII polyclonal antibodies in 1% skimmed milk and 0.05%) Tween in PBS. After 3 washing steps in PBS-Tween, the filters were incubated in a 1 :5.000 dilution of peroxidase-conjugated anti-mouse immunoglobulin (Dako) in 1% skimmed milk and 0.05% Tween in PBS for 1 hour at room temperature, and after 3 washing steps, bound conjugated IgGs were detected using the Super Signal West Pico chemo-luminescent substrate (Pierce). As shown in Figure 10B intense immune reactive bands with electrophoretic mobility identical to the corresponding engineered OmpF proteins visible in SDS-polyacrylamide gel stained with Coomassie Blue were detected.

Expression of OmpFvIII and OmpFvIIIx3 heterologous proteins into OMVs

Having demonstrated that engineered OmpF carrying one or three copies of vIII peptide in correspondence of one of the OmpF external loops were expressed in E. coli BL21(DE3)/AompA, the presence of engineered OmpF in the OMV fraction was analyzed. To this aim, the recombinant strains BL21(DE3)/ziom/?^(pET-OmpFvIII_Ll), BL2\(DE3)/AompA (pET-OmpFvIII_L2), BL2\(DE3)/AompA (pET-OmpFvIII_L4), BL2\(DE3)/AompA (pET-OmpFvIII_L6), BL2\(DE3)/AompA (pET-OmpFvIII_L7), and BL21(DE3 )/AompA (pET-OmpFvIII_L8) were grown in LB medium and when the cultures reached an OD 6 oo value of 0.6, IPTG was added at 1 mM final concentration. After two additional hours of growth at 37°C, vesicles were purified from culture supernatants by using ultrafiltration coupled to ultracentrifugation. More specifically, OMVs were collected from culture supernatants by filtration through a 0.22 μιη pore size filter (Millipore) and by high-speed centrifugation (200,000 x g for 2 hours). Pellets containing OMVs were finally suspended in PBS. The presence of the engineered OmpF proteins in each OMV preparation was verified by Western Blot analysis as already described using anti-vIII polyclonal antibodies. As shown in Figure 11, all OMVs purified from the supernatants of the six recombinant strains carried the respective engineered OmpF proteins. In particular, the OmpF proteins carrying the vIII peptide in place of Loop 1, Loop 2 and Loop 4 appear to accumulate in OMVs with remarkably high efficiency. As far as the two engineered OmpF proteins carrying three copies of vIII peptide in loop 1 and 4 are concerned, their presence in OMVs was verified by SDS-PAGE. B 2l(OE3)/AompA (pET- OmpFvIIIx3_Ll), and BL2l(OE3)/ ompA (pET-OmpFvIIIx3_L4) strains were grown in LB medium and when the cultures reached an OD 6 oo value of 0.6, IPTG was added at 1 mM final concentration. After two additional hours of growth at 37°C, vesicles were purified from culture supernatants by using ultrafiltration coupled to ultracentrifugation as described above. 25 μg of OMVs were separated on an SDS-polyacrylamide gel and after protein separation the gel was stained with Coomassie Blue. As shown in Figure 12, a protein band with an apparent molecular weight slightly higher than OmpF was clearly visible in both OMV preparations. The simultaneous presence of both wild type and engineered OmpF in OMVs was in line with the fact that both protein species were expressed in the strains, the wild type OmpF being encoded by the chromosomal DNA while the engineered OmpF being encoded by the recombinant plasmid.

Analysis of vIII expression on the surface of OmpFvIII and OmpFvIIIx3 recombinant strains

Finally, the surface expression of the vIII peptide was analyzed by Flow Cytometry in BL21(DE3)/AompA(pET-OmpFvIII_L2), BL21(DE3)/AompA (pET-OmpFvIII_L4), BL21(DE3)/AompA (pET-OmpFvIII_L6), BL21(DE3)/AompA (pET-OmpFvIIIx3_Ll), and BL21(DE3)/ AompA (pET-OmpFvIIIx3_L4), strains. Bacterial cultures were grown at 37°C under agitation. When cultures reached an OD600 value of 0.6, IPTG was added at a final concentration of 1 mM and bacteria were grown for 3 additional hours. Subsequently, lml bacteria cells were collected by centrifugation at 5,000 x g for 5 minutes. After a wash in 1% BSA/PBS bacteria were resuspended in 5 ml 1% BSA/PBS. 50 μΐ of cell suspensions were incubated with 5 μΐ of an appropriate dilution of rabbit anti-vIII polyclonal antibodies raised against the vIII peptide conjugated to Keyhole Limpet Hemocyanin (KLH) or, as negative control, with 5 μΐ of PBS containing 1% BSA. After 1 hour, the suspensions were centrifuged at 5.000 x g for 5 minutes and supematants discarded. Pellets were washed with 100 μΐ of PBS containing 1% BSA and bacteria subsequently incubated for 1 hour at 4°C with goat anti-rabbit antibodies (Alexa flour488, Life Technology) added at a final dilution of 1 :200. Finally, after a wash step, pellets were fixed with 100 μΐ 4% formaldehyde/PBS, re-suspended in 100 μΐ of PBS and analyzed with FACS CANTOII evaluating collected data with Flow Jo software. As shown in Figure 13, in the presence of anti-vIII antibodies, a clear shift in fluorescence intensity was observed in a substantial fraction of bacterial cells expressing the engineered OmpF proteins. No difference in fluorescence intensity was observed when E. coli BL21(DE3)/AompA expressing OmpF wt was incubated with anti- vIII antibodies. These data indicate that when used to replace the OmpF external loops, vIII peptide appeared on the surface of bacteria expressing the engineered OmpF proteins. Surface exposition was particularly pronounced when three copies of vIII peptide were used to replace the LI and L4 loops of OmpF.

OMV engineering with MBPvIII

pET-MBPvIII plasmid construction

In an attempt to deliver EGFRvIII peptide to the lumen of OMVs, the Maltose binding protein (MBP), which is naturally delivered to the periplasm, was used as a carrier (Fig. ID). In essence, a fusion protein was designed constituted by the whole MBP carrying the 13 amino acid vIII peptide at its carboxyl terminus (Figure 2A). For this purpose, the DNA sequence coding for MBP was amplified by PCR from E. coli K12-MG1655 genomic DNA, using primers Ndl-MalEf and Rl-MalEr. The forward primer (Ndl-MalEf) anneals to the 5' end of malE, the gene coding for MBP; the reverse primer (Rl-MalEr) anneals to the 3' end of malE (excluding the stop codon) and its 5' tail contains nucleotides 1-17 of the vIII sequence. Then, in order to complete the vIII coding sequence at the 3' of malE, a second and a third polymerase chain reactions were performed using the same forward primer (Ndl-MalEf) but different reverse primers. The reverse primer used in the second amplification step (R2-MalEr) anneals to a region containing the 3' end of malE and nucleotides 1-17 of the vIII sequence; its 5' tail contains nucleotides 18-31 of the vIII sequence. The reverse primer used in the third amplification step (XhI-R3-MalEr) anneals to nucleotides 2-31 of the vIII sequence; its 5' tail contains nucleotides 32-39 of the vIII sequence and the TAA stop codon (Table 1 and Fig. 14). With the aim of cloning the fusion gene coding for MBPvIII in pET21 expression vector the PIPE cloning method was used (Klock and Lesley, 2009). Briefly, pET21 was amplified using primers petrev/nohisflag and the fusion gene coding for MBPvIII was reamp lifted using primers pET21- MalEf/pET21-R3-MalEr, which have 5' tails that generate extremities complementary to pET21 expression vector linear DNA (Table 1 and Figure 14). E. coli HK100 transformation leads to the generation of pET-MBPvIII plasmid. The correctness of the cloning of the MBPvIII fusion in pET21 was verified by DNA sequencing (SEQ ID NO: 101).

Expression of MBPvIII heterologous protein in E. coli BL21 (DE3)/AompA

The recombinant plasmid encoding MBP fused to one copy of the EGFRvIII peptide was used to transform E. coli strain BL21(DE3)/AompA.

One recombinant strain was obtained: BL21(DE3)/AompA(pET-MBPvIII). The strain was grown in LB medium and when culture reached an OD600 value of 0.6, IPTG was added at 1 mM final concentration.

After two additional hours of growth at 37°C, cells were collected and 25 μg of total protein extracts were analyzed by SDS-PAGE. As shown in Figure 15 A, a protein band with an electrophoretic mobility corresponding to MBPvIII accumulated in the protein extract prepared after IPTG induction. Western Blot of total cell extracts was also carried out to confirm the expression of MBPvIII. To this aim, 13 μg of total proteins from each strain were separated by SDS-PAGE and proteins were transferred to nitrocellulose filters. The filters were blocked for 2h at room temperature by agitation in blocking solution (10% skimmed milk and 0.05% Tween in PBS), followed by incubation for 90 minutes at 37°C with a 1 : 1.000 dilution of rabbit anti-vIII polyclonal antibodies or mouse anti-MBP monoclonal antibody in 1% skimmed milk and 0.05% Tween in PBS. After 3 washing steps in PBS-Tween, the filters were incubated in a 1 :5,000 dilution of peroxidase-conjugated anti-rabbit or anti-mouse immunoglobulin (Dako) in 1% skimmed milk and 0.05% Tween in PBS for 1 hour at room temperature, and after 3 washing steps, bound conjugated IgGs were detected using the Super Signal West Pico chemo- luminescent substrate (Pierce). As shown in Figure 15B, EGFRvIII fused to MBP was found expressed in the cell lysates. No immune reactive bands were detected in total lysates from E. coli cells carrying empty pET21 plasmid.

Expression of MBPvIII into OMVs

Having demonstrated that MBPvIII was expressed in E. coli BL21(DE3)/AompA strain, the presence of the antigen in the OMV fraction was analyzed. Strain BL21(DE3)/AompA(pET-MBPvIII) was grown in LB medium and when the culture reached an OD600 value of 0.6, IPTG was added at 1 mM final concentration. After two additional hours of growth at 37°C, vesicles were purified from culture supernatants by using ultrafiltration coupled to ultracentrifugation. More specifically, OMVs were collected from culture supernatants by filtration through a 0.22 um pore size filter (Millipore) and by highspeed centrifugation (200,000 x g for 2 hours). Pellets containing OMVs were finally suspended in PBS. The presence of the antigens in OMVs preparations was verified by SDS- PAGE and Western Blot analyses as described in the previous section. Data indicate that MBP carrying one copy of vIII peptide was incorporated into OMVs (Figure 16A and B).

Antibody titers elicited in mice immunized with MBPvIII in engineered OMVs To test whether OMVs purified from BL21(DE3)/AompA strain expressing MBPvIII were able to induce EGFRvIII-specific antibody responses, CD1 mice were i.p. immunized three times at two-week interval with 10 μg OMVs in Alum. Blood samples were collected nine days after third dose administration and anti-vIII IgGs were detected by using plates coated in each well with 0.5 μg of synthetic vIII peptide conjugated to Keyhole limpet hemocyanin (vIII-KHL). PBS containing 1% BSA was used as negative control. More specifically, coating was carried out by incubating plates overnight at 4°C with 100 μΐ of conjugated peptide (5 μg/ml). Subsequently, wells were washed three times with PBST (0.05% Tween 20 in PBS, pH 7.4), incubated with 100 μΐ of 1% BSA in PBS for 1 h at room temperature and washed again three times with PBST. Serial dilutions of serum samples in PBST containing 1 % BSA were added to the plates, incubated 2 h at 37°C, and washed three times with PBST. Then 100 μΐ/well of 1 :2,000 diluted, alkaline phosphatase-conjugated goat anti-mouse IgGs were added and left for 2 h at 37°C. After triple PBST wash, bound alkaline phosphatase-conjugated antibodies were detected by adding ΙΟΟμΙ/well of 3mg/ml para-nitrophenyl-phosphate disodium hexahydrate (Sigma- Aldrich) in 1M diethanolamine buffer (pH 9.8). After 10 minute incubation at room temperature, the reaction was stopped with 100 μΐ 7% EDTA and substrate hydrolysis was analyzed at 405 nm in a microplate spectrophotometer. As shown in Figure 17, OMVs carrying MBPvIII fusion protein were able to induce high anti-vIII IgG titers in mice.

Engineered OMVs expressing FAT1 peptide

Two strategies were designed to deliver the FAT1 peptide to the E. coli OMVs. The first strategy envisaged the fusion of FAT 1 peptide to the carboxyl terminus of either full length fHbp or fHbpDomA. For this purpose, a synthetic DNA fragment encoding three copies of IQVEATDKDLGPNGHVTYSIVTDTD peptide was ligated to the 3' end of the full length fHbp gene and to the DNA sequence coding for the Domain A of fhbp (fHbpDomA), thus generating chimeric proteins carrying the FAT1 peptide at their C-terminus.

The second strategy was designed to deliver the FAT1 peptide into the lumen of OMVs. To this aim, the synthetic DNA coding for three copies of the FAT1 peptide was fused to the 3 ' end of the MBP gene to create an in frame C-terminal fusion.

Construction of pET21 FL-fHbp-FATl and pET21 JHbp-DomA-FATl plasmids The fusion of three copies of the FAT1 peptide to full length fHbp (FL-fHbp) and to fHbp-DomA was carried out in two main steps. First, the DNA fragments encoding FL- fHbp and fHbp-DomA were cloned into plasmid pET21, thus generating plasmids pET21- fHbp and pET21 -fHbp-DomA. Subsequently, the two plasmids were linearized to clone the synthetic DNA encoding three copies of FATl peptide (FATl Minigene) at the end of fHbp and fHbp-DomA coding sequences. The sequence of FATl Minigene was designed taking into consideration BL21 E. coli codon usage (SEQ ID NO: 102).

pET21_FL-fHbp and pET21 fHbp-DomA were generated as described in section 5.1.1. To generate pET21_FL-fHbp-FATl and pET21 fHbp-DomA- FATl plasmids, pET21-fHbp and pET21 -fHbp-DomA were linearized by PCR amplification using the two couples of primers FHBP-F/FHBPFL-R and FHBP -F/FHBPD A-R (Table 2) and the linear fragments were combined with the synthetic DNA coding for FATl Minigene (Table 2, Sequence 18). FATl Minigene was constructed by assembling six complementary oligonucleotides the sequence of which is reported in Table 2 and the assembled DNA fragment was amplified with primers F-FATFH/R-FATFH F-FATDomA/R-FATFH (Table 2) to make its extremities complementary to linearized pET21_FL-fHbp and pET21_FL-fHbpDomA, respectively. The DNA mixtures were then used to transform E. coli HK100 competent cells and clones carrying pET21_FL-fHbp-FATl and pET21_fHbp- DomA-FATl plasmids were selected on LB agar plates supplemented with 100 μg/ml Amplicillin. From one clone of each transformation the plasmid was purified and the correctness of the fHbp-FATl and fHbpDomA-FATl gene fusions was verified by DNA sequencing (SEQ ID NOs: 104 and 103).

pET-MBP-FATl plasmid construction

To express FATl peptide in the lumen of OMVs, the Maltose binding protein (MBP) which is naturally delivered to the periplasm was used as a carrier and the FATl Minigene was cloned as an in frame fusion to the 3' end of the MBP gene. For this purpose, plasmid pET- MBPvIII (see Section 5.1.3) has been used as template for a PCR reaction carried out according to the PIPE method (Klock H.E. and Lesley S.A (2009) Methods Mol. Biol. 498, 91-103), using primers pET21-MBPF and pET21-MBPR (see Table 2) to generate a linear fragment missing the vIII coding sequence. Then, the linear fragment was ligated to FATl Minigene, which was assembled in vitro using the six synthetic oligonucleotides reported in Table 2 and subsequently amplified with primers MBPFA-F and MBPFA-R. The DNA mixture was used to transform MK-100 competent cells and clones carrying pET-MBP-FATl plasmid were selected on LB agar plates supplemented with 100 μg/ml of Ampicillin. The correctness of the MBP and FAT1 Minigene fusion in pET-MBP-FATl plasmid purified from one of the Ampicillin resistant clones was verified by DNA sequencing (SEQ ID NO: 105).

Expression offHBP- FAT1 and MBP-FAT1 peptide in E.coli BL21(DE3)AompA

Plasmids pET21_FL-fHbp-FATl, pET21 _fHbp-DomA-F AT 1 and pET-MBP- FAT1 were used to transform BL21(DE3) AompA strain. Recombinant clones were grown in 200 ml LB medium at 37 C and when the cultures reached OD600 = 0.6, the expression of the fusion proteins was induced by addition of 1 mM IPTG. After 2 hour growth, the expression of protein fusions was assessed by SDS-PAGE and Western Blot. To this aim, the equivalent in volume of 1 OD 6 oo of each bacterial culture was collected, centrifuged at 13,000 X g for 5 minutes and pellets were lysed in 200 μΐ of BPer Reagent, Lysozime lmg/ml, DNAase 10 U/ml and 0.1 mM MgCh for 30 minutes. Then the samples were centrifuged at 13.000 x g for 20 minutes to separate the supernatants (soluble fraction) from the pellets (insoluble fraction). The soluble fraction was collected (200 μΐ) and diluted with 100 μΐ of 4x SDS-PAGE loading buffer while the pellets were re-suspended in 300 μΐ of 2X loading buffer. 20 μΐ of each sample were loaded onto an SDS-polyacrylamide gel and the proteins separated by electrophoresis (SDS-PAGE).

TABLE 2

Then the gel was stained with Coomassie blue overnight at room temperature. For

Western blot analysis 5 μΐ of the same samples were loaded onto a 4-12% polyacrilamide gel (Invitrogen). After electrophoretic separation proteins were transferred onto nitrocellulose filter by standard methods. The filters were blocked overnight at 4°C by agitation in blocking solution (10% skimmed milk and 0.1% Tween in PBS), followed by incubation for 90 minutes at 37°C with anti-FATl mAbl98.3 at 3 μg/ml in 1% skimmed milk and 0.1% Tween in PBS. After three washing steps in PBS-Tween, the filters were incubated in a 1 :5.000 dilution of peroxidase-conjugated anti-mouse immunoglobulin (PerkinElmer) in 1% skimmed milk and 0,1% Tween in PBS for an hour, and after three washing steps, bound conjugated IgGs were detected using the Super Signal West Pico chemo-luminescent substrate (Pierce) and the resulting signal was detected by using the Western lighting plus ECL (PerkinElmer).

As shown in Figure 18A and Figure 18B, protein species with apparent molecular mass corresponding fHbp-FATl, fHbpDomA-FATl and MBP-FATl were clearly visible after Coomassie Blue staining of SDS-polyacrylamide gels. The proteins largely compartmentalized in the insoluble fraction and were expressed after IPTG induction. The Western Blot analysis using anti-FATl mAbl98.3 confirmed that the proteins expressed after IPTG induction carried the FAT1 peptide (Figure 18C).

Expression of fHbp-FATl, fHbpDomA-F ATI and MBP-FATl into OMVs

Having demonstrated that fHbp-F AT 1 , fHbpDomA-FAT 1 and MBP-F AT 1 peptide were expressed in E. coli BL2l(OE3)/AompA, the presence of the fusions in the OMV fraction was analyzed. 200 ml of the bacterial cultures BL21(DE3 )/AompA (pET-fHbp- FAT1), BL2\(DE3)/AompA (pET-fHbpDomA-F AT 1 ) and BL2\(DE3)/AompA (pET- MBP-FAT1) were grown in LB and when the cultures reached an OD 6 oo=0.6 were induced with ImM IPTG. Vesicles were purified from the culture supernatants by using ultrafiltration coupled to ultracentrifugation. Briefly, cultures were centrifuged at 4000 x g for 15min, and the supernatants were concentrated using a membrane with a cut off of 100 kDA (Amicon) until a final volume of 30 ml was reached. Then an ultracentrifugation step was performed at 160.000 x g for 2 hours. The pellets were re-suspended in 200 μΐ of PBS. Protein quantization was performed by DC Protein Assay (BioRad). The presence of the specific antigen in OMVs preparation was verified by Western Blot and SDS-PAGE analysis as already described by loading 10 μg of total proteins of each OMV preparation. As shown in Figure 19A the three fusion constructs compartmentalized in OMVs with bands corresponding to the expected molecular mass visible by Coomoassie Blue staining of the gels.

The presence of FAT 1 peptide in the three fusion proteins was also confirmed by Western Blot, where the anti FAT1 mAbl98.3 recognized the three protein species visible in the OMVs preparations by Coomassie Blue staining (Figure 19B). Analysis of fHBP-FATl expression on the surface of E.coli BL21(DE3) AompA by

FACS

In order to confirm the ability of fHbp to deliver the FATl peptide on the outer membrane of E.coli AompA, FATl expression on the bacterial surface was analyzed by FACS.

BL2\(DE3)/AompA (pET-fHbp-F AT 1 ) and BL2\(DE3)/AompA (pET-fHbpDomA- FAT1) E.coli strains were grown in 10 ml LB medium at 37 C and when the cultures reached OD 6 oo = 0.6, the expression of the fusion proteins was induced by addition of 1 mM IPTG. After 2 hour growth bacteria cells corresponding to those contained in 1 ml culture at OD 6 oo = 1 were collected by centrifugation at 13.000 x g for 5 minutes and pellets were re-suspended in 50 ml of PBS containing 1% BSA. 50 μΐ of cell suspensions were incubated with 50 μΐ of an appropriate dilution of anti-FATl mAbl98.3 or, as negative control, with 50 μΐ of PBS containing 1% BSA. After 1 hour, 100 μΐ of PBS containing 1% BSA were added and the suspensions were centrifuged at 3.000 x g for 10 minutes and supernatants discarded. Pellets were washed with 200 μΐ of PBS containing 1% BSA and bacteria subsequently incubated for 30 minutes on ice with goat anti-mouse antibodies (Alexa flour488, Life Technology) added at a final dilution of 1 :200. Finally, After 2 wash steps, pellets were re-suspended in 200 μΐ of PBS and analyzed with FACS CANTOII evaluating collected data with Flow Jo software. As shown in Figure 20, the FATl peptide, fused to fHbpDomA and fHbpFL, was clearly expressed on the surface of E.coli as indicated by the shift in fluorescence intensity observed after incubation of bacterial cells with anti-FATl specific monoclonal antibody.

Antibody titers elicited in mice immunized with fHbp-FATl, fHbpDomA-FATl and MBP-FAT1 in engineered OMVs

To test whether OMVs purified from BL21(DE3)/ AompA strain expressing fHbpFLFATl, fHbpDomA-F AT 1 and MBP-FAT1 were capable of inducing FATl peptide-specific antibody responses, groups of CD1 mice (5 mice per group) were i.p. immunized three times at two-week interval with 20 ug OMVs in Alum. After two weeks from the third immunization, sera were collected and pooled to analyze anti-FATl antibody titers by ELISA. ELISA was performed using plates coated with synthetic FAT1 peptide conjugated to the carrier protein Keyhole limpet hemocyanin (KLH). Coating was carried out at room temperature for 14 hours by adding to each well 100 μΐ of conjugated peptide at a concentration of 5 μg/ml. After three washes with 200 μΐ/well of PBS supplemented with 0.05% Tween 20 (PBST) the plates were incubated one hour at 37 C with 100 μΐ/well of PBS containing 1% BSA and subsequently washed three times with PBST. Different dilutions of sera in PBST containing 0.1% BSA were added in duplicate in a final volume of 100 μΐ/well and plates were stored at 37 C for 2 hours. After three washes in PBST, 100 μΐ of goat anti-mouse antibodies conjugated to alkaline phosphatase (SouthernBiotech, Cat. 1030-04, 1 :2.000 dilution) were added to each well and incubated at 37 C for 1 hour. Finally, after three washes, the phosphatase substrate (4-Nitrophenyl phosphate disodium salt) was added to each well at a concentration of 1 mg/ml (100 ml/well) and after 30 minutes incubation at room temperature in the dark, substrate hydrolysis was measured spectrophotometrically at 405 nm. As shown in Figure 23B, OMV engineered with FAT1 induced high concentrations of anti-FATl antibodies detectable even at serum dilutions higher them 1 :24.000. In particular, OMV expressing fHbpFL-FATl and MBP-FAT1 induced antibodies titers which at the highest dilution (1 :24.000) still gave saturating OD values similar to what observed with the anti-FATl mAb 198.3 which at 1 :24.000 dilution corresponded to a concentration of 100 ng/ml. No significant titer against conjugated FAT1 was detected using sera from mice immunized with OMV engineered with the uncorrelated peptide MUC1.

Engineered OMVs expressing MUC1 peptide

Two strategies were designed to deliver the MUC1 peptide to E. coli OMVs. The first strategy was designed to deliver MUC1 peptide to the membrane of the vesicles by fusing MUC1 to the C-terminus of fHbpDomA. To achieve this, a synthetic DNA fragment encoding five copies of the MUC1 peptide GVTSAPDTRPAPGSTAPPAH was ligated to the 3' end of the gene coding for Domain A of fHbp (fHbpDomA). The second strategy was designed to deliver the MUC1 peptide into the lumen of OMVs. To this aim, the synthetic DNA coding for five copies of the peptide was fused to the 3' end of the MBP gene to create an in frame C-terminal fusion.

OMV engineering with flibpDomA-MUCl peptide

The fusion of five copies of the MUC1 peptide to fHbp-DomA was carried out in two main steps. First, the DNA fragment encoding fHbp-DomA were cloned into plasmid pET21 thus generating plasmid pET21 fHbp-DomA as already described. Subsequently, the plasmid was linearized by PCR and ligated to the synthetic DNA fragment encoding five copies of MUC1 peptide (MUC1 Minigene). The sequence coding for MUC1 Minigene was designed taking into consideration the E. coli codon usage. More specifically, pET21 fHbp-DomA was generated as follows. The fHbpDomA gene was amplified by PCR from Neisseria meningitidis serogroup B strain MC58 genome. The primers were designed to amplify the gene with its natural leader sequence for secretion and the lipobox. The polymerase incomplete primer extension (PIPE) cloning method (Klock H. et al, 2009) was used to insert the PCR product into plasmid pET21b which was amplified using primers pet-rev/nohisflag (Table 2). In so doing plasmid pET21 fHbp-DomA was generated (see also above for additional details).

To generate pET21_fHbp-DomA-FATl plasmid the PIPE cloning method was used. pET21 fHbp-DomA was linearized by PCR amplification using primers FHBP-F and FHBPDA-R (Table 2) and the linear fragment was combined with the synthetic DNA coding for MUC1 Minigene (Table 2 and Sequence 22). MUC1 Minigene was constructed by assembling ten complementary oligonucleotides the sequence of which is reported in Table 3 and the assembled DNA fragment was amplified with primers RMUCFH and FMUCDomA (Table 3) to make its extremities complementary to the amplified vector. The DNA mixture was then used to transform E. coli HK100 competent cells and clones carrying pET21_fHbp-DomA-MUCl plasmid were selected on LB agar plates supplemented with 100 μg/ml Amplicillin. From one clone the plasmid was purified and the correctness of the fHbpDomA- MUC1 gene fusion was verified by DNA sequencing (SEQ ID NO: 107).

pET-MBP-MUCl plasmid construction

To express MUC1 peptide in the lumen of OMVs, the Maltose binding protein (MBP) which is naturally delivered to the periplasm was used as a carrier and the MUC1 Minigene was cloned as an in frame fusion to the 3 ' end of the MBP gene. For this purpose, plasmid pET-MBPvlll (see Section 5.1.3) has been used as template for a PCR reaction carried out according to the PIPE method (Klock H.E. and Lesley S.A (2009) Methods Mol. Biol. 498, 91-103), using primers pET21-MBPF and pET21-MBPR (see Table 2) to generate a linear fragment missing the vIII coding sequence. Then, the linear fragment was ligated to MUC1 Minigene constructed as described above and amplified with primers MBPMU-F and MBPMU-R (Table 3) to make its extremities complementary to the amplified vector. The DNA mixture was then used to transform E. coli HK100 competent cells and clones carrying pET21 MBP-MUCl plasmid were selected on LB agar plates supplemented with 100 μg/ml Amplicillin. From one clone the plasmid was purified and the correctness of the MBP-MUCl gene fusion was verified by DNA sequencing (SEQ ID NO: 108).

TABLE 3

Expression offlibpDomA- MUC1 and MBP-MUC1 fusions in E.coli BL21(DE3)

AompA

Plasmids pET21_fHbp-DomA-MUCl and pET-MBP-MUCl were used to transform BL21(DE3) AompA strain. Recombinant clones were grown in 200 ml LB medium at 37 C and when the cultures reached OD600 = 0.6, the expression of the fusion proteins was induced by addition of 1 mM IPTG. After 2 hour growth, the expression of protein fusions was assessed by SDS-PAGE. To this aim, the equivalent in volume of 1 OD 6 oo of bacteria culture was collected, centrifuged at 13.000 X g for 5 minutes and pellets were lysed in 200 μΐ of BPer Reagent, Lysozime lmg/ml, DNAase lOU/ml and 0.1 mM MgCk for 30 minutes. Then the samples were centrifuged at 13.000 x g for 20 minutes to separate the supematants (soluble fraction) from the pellets (insoluble fraction). The soluble fractions were collected (200ul) and diluted with 100 μΐ of 4x SDS-PAGE loading buffer while the pellets were re-suspended in 300 μΐ of 2x loading buffer. 20 μΐ of each sample were loaded onto an SDS-polyacrylamide gel and proteins separated by electrophoresis (SDS-PAGE). Then the gel was stained with Coomassie blue overnight at room temperature. As shown in Figure 21, protein species with apparent molecular mass corresponding fHbpDomA-MUCl and MBP-MUC1 were visible after Coomassie Blue staining of SDS-polyacrylamide gels. The proteins largely compartmentalized in the insoluble fractions and were expressed after IPTG induction.

Compartmentalization of fHbpDomA-MUCl and MBP-MUC1 fusions into OMVs Having demonstrated that fHbpDomA-MUCl and MBP-MUC1 were expressed in

E. coli BL21(DE3) 'ΔοτηρΑ, the presence of the fusions in the OMV compartment was analyzed. To this aim, 200 ml of the bacterial cultures BL21(DE3 )IAompA (pET- fHbpDomA-MUCl) and B 2l(OE3)/AompA (pET-MBP-MUCl) were grown in LB and when the cultures reached an OD 6 oo=0.6 were induced with ImM IPTG. Vesicles were purified from the culture supernatants by using ultrafiltration coupled to ultracentrifugation. Briefly, cultures were centrifuged at 4.000 x g for 15min, and the supernatants were concentrated using a membrane with a cut off of 100 kDA (Amicon) until a final volume of 30 ml was reached. Then an ultracentrifugation step was performed at 160.000 x g for 2 hours. The pellets were re-suspended in 200 μΐ of PBS. Protein quantization was performed by DC Protein Assay (BioRad). The presence of the fusion proteins in the OMVs preparation was verified by SDS-PAGE analysis as already described by loading 10 μg of total proteins of each OMV preparation. As shown in Figure 22, the two fusion constructs compartmentalized in OMVs as indicated by the appearance of protein bands corresponding to the expected molecular mass.

Antibody titers elicited in mice immunized with engineered OMVs carrying fHbpDomA-MUCl and MBP-MUC1

To test whether OMVs purified from BL21 (DE3 'AompA (pET-fHbpDomA-MUC 1 ) and BE2\{OE3)IAompA (pET-MBP-MUCl) strains were capable of inducing MUC1- specific antibody responses, CD1 mice (5 mice per group) were i.p. immunized three times at two-week intervals with 20 μg of engineered OMVs in Alum. After two weeks from the third immunization, sera were collected and pooled to analyze anti-MUCl antibody titers by ELISA. ELISA was performed using plates coated with the synthetic MUCl peptide GVTSAPDTRPAPGSTAPPAH. Coating was carried out at room temperature for 14 hours by adding to each well 100 μΐ of a solution of synthetic MUCl peptide at a concentration of 5 μg/ml. After three washes with 200 μΐ/well of PBS supplemented with 0.05% Tween 20 (PBST) the plates were incubated one hour at 37 C with 100 μΐ/well of PBS containing 1% BSA and subsequently washed three times with PBST. Different dilutions of sera in PBST containing 0.1% BSA were added in duplicate in a final volume of 100 μΐ/well and plates were stored at 37 C for 2 hours. After three washes in PBST, 100 μΐ of goat anti- mouse antibodies conjugated to alkaline phosphatase (SouthernBiotech, Cat. 1030-04, 1 :2.000 dilution) were added to each well and incubated at 37 C for 1 hour. Finally, after three washes, the phosphatase substrate (4-Nitrophenyl phosphate disodium salt) was added to each well at a concentration of 1 mg/ml (100 μΐ/well) and after 30 minute incubation at room temperature in the dark, substrate hydrolysis was measured spectrophotometrically at 405 nm. As shown in Figure 23 A, OMV engineered with MUCl induced high concentrations of anti-MUCl antibodies detectable even at serum dilutions higher them 1 :24.000. In particular, both engineered OMV expressing fHbpDomA-MUCl and MBP- MUCl induced antibodies titers which at the highest dilution (1 :24.000) still gave saturating OD values. No relevant titers against MUCl peptide was detected using sera from mice immunized with OMV engineered with the uncorrected peptide FAT1.

Engineered OMVs expressing Aa-fHbp

pET-Aa-flibp-HIS8 plasmid construction

The gene encoding the 828bp lipoprotein gnal870 like-protein was chemically synthetized (Gene Art™ Gene Synthesis, Thermo Fisher Scientific) using the reported gene sequence (EnsemblBacteria gene ID HMPREF9996 00541) from Aggregatibacter actinomycetemcomitans Y4 with the exception that the natural GTG start codon was replaced by an ATG start codon. The synthetic gene (SEQ ID NO: 1 16) was cloned into pET21b + fused to a 8-HIS tag at the C-term for subsequent detection of the protein using an anti-HIS polyclonal antibody. For cloning: the synthetic gene was amplified by PCR using the AgfHbp F and AgfHbp R primers, and annealed to pET21b + plasmid backbone amplified with pET HIS-F and pET 2-R primers (Table 1).

Expression of the Aa-fHbp heterologous protein in E. coli BL21(DE3)/AompA strain

To investigate whether Aa-fHbp was surface-associated when expressed in E. coli, the pET Aa-fHbp-Hiss recombinant plasmid was used to transform E. coli BL21 AompA and the expression and localization of Aa-fHbp was analyzed as described previously. Each strain was grown in LB medium and when cultures reached an OD 6 oo value of 0.6, IPTG was added at 1 mM final concentration. After two additional hours of growth at 37°C, cells were collected and total protein extracts were analyzed by SDS-PAGE followed by Coomassie staining. No bands were visible in total lysates fromE. coli cells carrying empty pET21 plasmid. A band corresponding to the Aa-fHbp protein is detected in the total lysate of the strain carrying the pET Aa-fHbp-Hiss recombinant plasmid as detected by Coomassie.

Analysis of Aa-fHbp expression in OMVs

Having demonstrated that the exogenous Aa-fHbp protein was well expressed in E. coli BL21(DE3 )/ AompA strain, we then analysed its cellular localization and compartimentalization to the OMV fraction. The recombinant strain BL21(DE3)/ziom/¾4(pETAa-fHbp-_HIS8) was grown in LB medium and when the cultures reached an OD 6 oo value of 0.6, IPTG was added at 1 mM final concentration. After two additional hours of growth at 37°C, vesicles were purified from culture supernatants by using ultrafiltration coupled to ultracentrifugation. More specifically, OMVs were collected from culture supernatants by filtration through a 0.22 μιη pore size filter (Millipore) and by high-speed centrifugation (200.000 x g for 2 hours). Pellets containing OMVs were finally suspended in PBS. The presence of the Aa-fHbp_HIS8 fusion protein in OMVs preparations was verified by Coomassie and Western Blot analysis as described in the previous section (Figure 24). Data indicate that the recombinant protein was incorporated into OMVs as shown by the presence of the corresponding correct molecular weight band in the Coomassie stained SDS-PAGE and a specular specific band in the western blot analysis probed by anti-His tag antibody.

Analysis of cellular localization of Aa-fHbp

The localization of recombinant Aa-fHbp protein was evaluated by flow cytometry. To this aim, recombinant E. coli strains BL21(DE3)/ziom/¾4(pET- Aa-fHbp-HIS8) and E. coli BL21(DE3)/ziom/¾4(pET21), as negative control, were grown at 37°C under agitation. When cultures reached an OD 6 oo value of 0.6, IPTG was added at a final concentration of 1 mM and bacteria were grown for 2 additional hours. Subsequently, bacteria cells corresponding to those contained in 1 ml culture at OD 6 oo = 1 were collected by centrifugation at 13,000 x g for 5 minutes and pellets were re-suspended in 50 ml of PBS containing 1% BSA. 50 μΐ of cell suspensions were incubated with 50 μΐ of an appropriate dilution of anti-His-tag antibody with 50 μΐ of PBS containing 1% BSA as negative control. After 1 hour, 100 μΐ of PBS containing 1% BSA were added and the suspensions were centrifuged at 3,000 x g for 10 minutes and supernatants discarded. Pellets were washed with 200 μΐ of PBS containing 1% BSA and bacteria subsequently incubated for 30 minutes on ice with goat anti-mouse antibodies (Alexa flour488, Life Technology) added at a final dilution of 1 :2,000. Finally, After 2 wash steps, pellets were re-suspended in 200 μΐ of PBS and analyzed with FACS CANTOII evaluating collected data with FlowJo software. As shown in Figure 24, in the presence of anti-his-tag antibody, a clear shift in fluorescence intensity was observed in a substantial fraction of bacterial cells expressing Aa-fHbp-HIS8 fusion protein. No difference in fluorescence intensity was observed when E. coli BL21(DE3 )IAompA were incubated with anti-his-tag antibody. Taken together these data indicate not only that AafHbp is expressed in E. coli BL21(DE3 )IAompA and is associated to the outer membrane but is also capable of exposing a foreign tag fused to its C-terminal portion to the extracellular compartment.