The invention relates to the field of immunity against coronaviruses. In this respect, the invention provides a lentiviral-based immunogenic agent that is suitable for use in boost or target immunization treatment in a subject, in particular a human subject, who had previously developed an immunity against Severe Acute Respiratory Syndrome coronavirus 2 (SARS- CoV-2) based on: (i) vaccination with a first generation of vaccines against SARS-CoV-2 infection or disease such as a protein, an mRNA, an adenovirus, an inactivated virus or a protein subunit vaccine composition against SARS-CoV-2 infection or disease, in particular a protein- or an mRNA-based vaccine, or (ii) SARS-CoV-2-induced or correlated disease. The invention accordingly concerns a lentiviral-based immunogenic agent that in particular may help overcome the deficiencies of available vaccines against SARS-CoV-2, especially may be efficient in overcoming the waning immune response or insufficient cellular memory response observed after immunization with available first generation of vaccines such as a protein, an mRNA, an adenovirus, an inactivated virus or a protein subunit vaccine, in particular protein or mRNA vaccine, by triggering a mucosal humoral and cellular immune response against coronaviruses, including a long-lasting immune response.

Considering: (i) the sustained pandemicity of coronavirus disease 2019 (COVID-19),

(ii) weakening protection potential of the first-generation vaccines against SARS-CoV-2, and

(iii) ceaseless emergence of new viral Variants of Concerns (VOCs), new effective vaccine platforms can be critical for the future primary or booster vaccines (Global COVID-19 Vaccination - Strategic Vision for 2022, World Health Organization, SAGE meeting October 2021). The inventors recently demonstrated the strong performance of a lentiviral vaccination vector (LV) encoding the full-length sequence of Spike glycoprotein (S) from the ancestral SARS-CoV-2 (LV::S), when used in systemic prime followed by intranasal (i.n.) boost in multiple preclinical models (Ku MW, et al. Cell Host Microbe, 29(2), 236-249 e236, 2021). LV::S ensures complete (cross) protection of the respiratory tract against ancestral SARS- CoV-2 and VOCs (Ku MW, etal. EMBO Mol Med, e14459, 2021). In addition, in new transgenic mice, expressing human Angiotensin Converting Enzyme 2 (hACE2) and displaying unprecedented permissiveness of the brain to SARS-CoV-2 replication, an i.n. boost with LV::S is required for full protection of the central nervous system (Ku MW, et al. EMBO Mol Med, e14459, 2021). LV::S is intended to be used as a primary vaccine or a booster to reinforce and broaden protection against emerging VOCs with immune evasion potential (Juno JA, Wheatley AK. Nat Med, 27(11), 1874-1875, 2021). The duration of the protection conferred by the first generation COVID-19 vaccines is not yet well established, hardly predictable with serological laboratory tests and variable in diverse individuals and against distinct VOCs. Despite high vaccination rates, the current exacerbation of the world-wide pandemic indicates that repeated booster immunizations will be needed to ensure individual and collective immunity against COVID-19. In this context, the safety and potential adverse effects of multiple additional homologous doses of the first generation COVID-19 vaccines, for instance related to allergic reaction to polyethylene glycol (PEG) contained in mRNA vaccines, have to be taken into account ( Castells MC, Phillips EJ. N Engl J Med, 384(7), 643-649, 2021). Importantly, unmatched vaccine delivery methods, i.e., heterologous prime-boost format, has been proven to be a more successful strategy than homologous prime-boost approach in numerous preclinical models of various infectious diseases (He Q, et al. Emerg Microbes Infect, 10(1), 629-637, 2021; Lu S. Curr Opin Immunol, 21(3), 346-351, 2009; Nordstrom P, et al. Lancet Reg Health Eur, 100249, 2021). Therefore, new efficient vaccination platforms are of particular interest to develop heterologous boosters against COVID-19. The LV::S vaccine candidate has a serious potential for prophylactic use against COVID-19, mainly based on its strong capacity to induce, not only strong neutralizing humoral responses, but also robust protective T-cell responses which are not impacted by the escape mutations accumulated in the SARS-CoV-2 VOCs (Ku MW, et al. EMBO Mol Med, e14459, 2021).

Besides, it has been observed that with the decreasing prophylactic potential of the immunity initially induced by the first-generation vaccines, especially against new VOCs, administration of additional vaccine doses becomes essential (Global COVID-19 Vaccination - Strategic Vision for 2022, World Health Organization, SAGE meeting October 2021). As an alternative to additional doses of the same vaccines, combining vaccine platforms in a heterologous prime-boost regimen may hold promise for gaining protective efficacy (Barros- Martins J, et al. Nat Med, 27(9), 1525-1529, 2021). Compared to homologous vaccine dose administration, heterologous prime-boost strategies may reinforce better the specific adaptive immune responses and long-term protection, without triggering/reinforcing vector-specific immunity or the risk of aggravation of possible reactogenicity to the vaccines themselves or excipients. Furthermore, the sequence of the Spike antigen has to be adapted according to the dynamics of SARS-CoV-2 VOC emergence in order to induce the greatest neutralization breadth. In addition, whereas protection against symptomatic SARS-CoV-2 infection is mainly related to sero-neutralizing activity, protection against severe COVID-19 involves CD8 ⁺ T-cell immunity. These cells with their ability to cytolyze virus-infected cells, especially controls the virus replication and result in resolution of SARS-CoV-2 infection (Sette A, Crotty S. Cell, 184(4), 861-880, 2021). Therefore, an appropriate B- and T-cell vaccine platform, including an adapted Spike sequence, is of utmost interest at the current step of the pandemic.

The inventors have reasoned that LV::S could be remarkably suitable to be used as a heterologous i.n. booster vaccine, to reinforce and broaden protection against the SARS-CoV- 2 in particular against its known and emerging VOCs (including but not limited to Alpha, Beta, Gamma, Delta and Omicron variants of SARS-CoV-2), while collective immunity in early vaccinated nations is waning only a few months after completion of the initial immunization, and while new waves of infections are on the rise (Juno JA, Wheatley AK. Nat Med, 27(11), 1874- 1875, 2021).

LVs for use in the present invention are in particular non-integrating, non-replicative, non-cytopathic and negligibly inflammatory (Hu B, Tai A, Wang P. Immunol Rev, 239(1), 45- 61, 2011; Ku MW, Charneau P, Majlessi L. Expert Rev Vaccines, 1-16, 2021). These vectors are pseudotyped with the heterologous glycoprotein from Vesicular Stomatitis Virus (VSV-G) which confers them a broad tropism for diverse cell types, notably including dendritic cells. The latter are mainly non-dividing cells and thus barely permissive to gene transfer. Hence, LVs possess the central property to efficiently transfer genes to the nuclei of non-dividing cells, which therefore renders possible efficient transduction of dendritic cells. The resulting endogenous antigen expression in these cells with unique ability to activate naive T cells (Guermonprez P, et al. J. Int Rev Cell Mol Biol, 349, 1-54, 2019) correlates with outstanding ability of LV at inducing high-quality effector and memory T cells (Ku MW, et al. Commun Biol, 4(1), 713, 2021). Importantly, VSV-G pseudo-typing also avoids LVs to be targets of preexisting vector-specific immunity in humans which is key in vaccine development (Hu B, Tai A, Wang P. Immunol Rev, 239(1), 45-61, 2011; Ku MW, Charneau P, Majlessi L. Expert Rev Vaccines, 1-16, 2021). The safety of LV has been established in humans in a phase l/lla Human Immunodeficiency Virus-1 therapeutic vaccine trial (EU Clinical Trials Register, Clinical Trials for 2011 -006260-52). Because of their non-cytopathic and non-inflammatory properties (Cousin C, etal. Cell Rep, 26(5), 1242-1257 e1247, 2019; Lopez J, etal. An optimized lentiviral vector induces CD4+ T-cell immunity and predicts a booster vaccine against tuberculosis, in revision), LVs are well suitable for mucosal vaccination. The i.n. administration route presents well-recognized advantages of triggering mucosal IgA responses, as well as resident memory B and T lymphocytes in the respiratory tract (Lund FE, Randall TD. Science, 373(6553), 397- 399, 2021). This route has also been shown to be the most effective at reducing SARS-CoV- 2 transmission in both hamster and macaque preclinical models (van Doremalen N, et al. Sci Transl Med, 13(607), 2021). Induction of mucosal immunity by i.n. immunization allows SARS- CoV-2 neutralization, directly at the gateway to the host organism, before it gains access to major infectable anatomical sites (Ku MW, et al. Cell Host Microbe, 29(2), 236-249 e236, 2021).

In the present invention, the inventors generated an LV encoding the down-selected Scov-2 of the Beta variant, stabilized by K ⁹⁸⁶ P and V ⁹⁸⁷ P substitutions in the S2 domain of Scov-2 (LV::Sseta-2p). In mice, primed and boosted intramuscularly (i.m.) with mRNA-1273 (Moderna) vaccine, and in which the (cross) sero-neutralization potential was progressively turning down, the inventors compared the systemic and mucosal immune responses and the protective potential of an i.n. LV::Sseta-2P heterologous boost vs an i.m. mRNA-1273 (Moderna) (Jackson LA, etal. Preliminary Report. N Engl J Med, 383(20), 1920-1931, 2020; Wang F, etal. Med Sci Monit, 26, e924700, 2020) homologous boost. The inventors observed multiple advantages with respect to the former regimen with an improved antigen design, new vaccine delivery LV platform and the alternative i.n. administration route.

According to a first aspect, the invention hence relates to a pseudotyped lentiviral vector particle encoding a Spike (S) protein of a Severe Acute Respiratory Syndrome coronavirus 2 (SARS-CoV-2) or a derivative thereof for use as a heterologous boost or target immunization agent in a vaccine regimen for administration to the upper respiratory tract of a subject, in particular a human subject, who received a prime immunization with a vaccine composition against SARS-CoV-2 infection or disease selected from the group consisting of a protein, an mRNA, an adenovirus, an inactivated virus and a protein subunit vaccine composition against SARS-CoV-2 infection or disease, in particular a protein or an mRNA vaccine composition against SARS-CoV-2 infection or disease. Non-limited examples of protein subunit vaccine compositions against SARS-CoV-2 infection or disease according to the invention may include vaccines based on adjuvanted recombinant Spike protein or vaccines based on recombinant Spike protein packaged in nanoparticles.

Spike (S) protein of SARS-CoV-2 virus is well identified in the art as an envelop- anchored glycoprotein (Walls et al, 2020, Structure, Function, and Antigenicity of the SARS- CoV-2 Spike Glycoprotein. Cell 181:281-292 e286). More precisely, the SARS-CoV-2 S (Scov- ₂ ) is a (180 kDa) ₃ homotrimeric class I viral fusion protein, which engages the carboxypeptidase Angiotensin-Converting Enzyme 2 (ACE2), expressed on host cells. The monomer of Scov-2 protein possesses an ecto-domain, a transmembrane anchor domain, and a short internal tail. Scov-2 is activated by a two-step sequential proteolytic cleavage to initiate fusion with the host cell membrane. Subsequent to Scov-2- ACE2 interaction, which leads to a conformational reorganization, the extracellular domain of Scov-2 is first cleaved at the highly specific furin 682 ^RRAR 685 (SEQ ID NO: 21 ) site (Guo et al., 2020, The origin, transmission and clinical therapies on coronavirus disease 2019 (COVID-19) outbreak - an update on the status. Mil Med Res 7, 17; Walls et al., 2020), a key factor determining the pathological features of the virus, linked to the ubiquitous furin expression (Wang eta!., 2020 A Unique Protease Cleavage Site Predicted in the Spike Protein of the Novel Pneumonia Coronavirus (2019-nCoV) Potentially Related to Viral Transmissibility. Virol Sin 2020 Jun;35(3):337-339. doi: 10.1007/s12250-020-00212-7. Epub 2020 Mar 20). The resulted subunits are constituted of: (i) S1 , which harbors the ACE2 Receptor Binding Domain (RBD), with the atomic contacts restricted to the ACE2 protease domain and also harbors main B-cell epitopes, targeted of neutralizing antibodies (NAbs) (Walls et al., 2020), and (ii) S2, which bears the membranefusion elements. Like for Scov-i, the shedding of S1 renders accessible on S2 the second proteolytic cleavage site 797 ^R , namely S2' (Belouzard et al., 2009, Activation of the SARS coronavirus spike protein via sequential proteolytic cleavage at two distinct sites. Proc Natl Acad Sci U S A 106:5871-5876). According to the cell or tissue types, one or several host proteases, including furin, trypsin, cathepsins or TransMembrane Protease Serine Protease (TMPRSS)-2 or -4, can be involved in this second cleavage step (Coutard et al., 2020, The spike glycoprotein of the new coronavirus 2019-nCoV contains a furin-like cleavage site absent in CoV of the same clade. Antiviral Res 176:104742). The consequent “fusogenic” conformational changes of S result in a highly stable post-fusion form of Scov-2 that initiates the fusion reaction with the host cell membrane (Sternberg and Naujokat, 2020 Structural features of coronavirus SARS-CoV-2 spike protein: Targets for vaccination. Life Sci 257, 118056) and lead to the exposure of a Fusion Peptide (FP), adjacent to S2'. Insertion of FP to the host cell/vesicle membrane primes the fusion reaction, whereby the viral RNA release into the host cytosol (Lai et al., 2017 The SARS-CoV Fusion Peptide Forms an Extended Bipartite Fusion Platform that Perturbs Membrane Order in a Calcium-Dependent Manner. J Mol Biol 429:3875- 3892). The facts that the Scov-2-ACE2 interaction is the only mechanism, thus far identified for the host cell infection by SARS-CoV-2, and that the RBD contains numerous conformational B-cell epitopes (Walls etal., 2020), designate this viral envelop glycoprotein as the main target for neutralization antibodies (NAbs).

The S protein for expression by the lentiviral particles of the invention may originate from a SARS-CoV-2 strain and accordingly maybe characterized by an amino acid sequence that is the native sequence of the viral protein. In a particular embodiment the invention is performed using the S protein of known SARS-CoV-2 strains such as the S protein of the Ancestral strain (wherein the amino acid sequence is SEQ ID NO: 1 ), or of variant strains discovered later such as the Alpha, Beta, Gamma, Delta or Omicron strain (all regarded as variant strains with respect to one another). The invention may alternatively be performed with a derivative of the S protein, i.e., a derivative of a native S protein obtained by mutation in the amino acid sequence of the S protein, as will be disclosed herein. In order to express LV::S recombinant particles, the nucleic acid encoding the S protein may have the sequence of the gene present in the viral strain of origin or may be a codon-optimized acid nucleic suitable for expression in mammalian cells, in particular in human cells. In order to express LV recombinant particles expressing a derivative of the S protein, the nucleic acid encoding the derivative of the S protein may have the sequence deduced from the sequence of the gene of the S protein present in a viral strain and may be a codon-optimized acid nucleic suitable for expression in mammalian cells.

In a particular embodiment, the recombinant lentiviral particles (LV) used in the invention are HIV-1 -based lentiviral particles. Accordingly, where the expressions “lentiviral particle” of "LV” are used herein it is in particular directed to the HIV-1 based lentiviral particles especially LV particles pseudotyped with VSV-G protein, in particular LV as illustrated in the examples.

The expressions “boost” or “boost immunization” or “boost administration” or “target immunization” refer according to the invention to an administration of the immunogenic agent that comes after a first administration of a heterologous immunization agent, in particular a heterologous vaccine, or after a second or later administration of such heterologous immunization agent or vaccine. Otherwise stated the immunization agent used according to the invention is administered to a subject who previously received a prime administration, or a prime and further one or multiple administration doses, of a heterologous immunization agent or vaccine against the same SARS-CoV-2 or against a variant strain thereof. The boost or target immunization is achieved through administration to the upper respiratory tract, in particular as an intranasal administration, that accordingly distinguish over administration route of a first generation of vaccines against SARS-CoV-2 infection or disease such as a protein, an mRNA, an adenovirus, an inactivated virus or a protein subunit vaccine composition against SARS-CoV-2 infection or disease, in particular protein or mRNA vaccines that most often make use of systemic, including intramuscular, intradermal or subcutaneous administration route. The boost or target immunization is intended to enhance, improve or lengthen the immune response previously raised and possibly to broaden such response to elicit cross-neutralization against multiple SARS-CoV-2 viruses. Improvement of the response may arise from the capability of the immunization agent used in the invention to elicit a mucosal response and to accordingly protect, not only the systemic sites, but also the upper and lower respiratory tracts and the central nervous system that may not have been successfully targeted or protected with heterologous vaccines against SARS-CoV-2 infection or disease such as a protein, an mRNA, an adenovirus, an inactivated virus or a protein subunit vaccine composition against SARS- CoV-2 infection or disease, in particular protein or mRNA vaccines injected via systemic routes. In a particular embodiment the boost administration is intended to raise crossneutralizing immune response in the subject against emerging stains of the virus. In a particular embodiment, the boost or target immunization may be administered to a subject who received a heterologous immunization agent as disclosed herein and who had and recovered from infection by SARS-CoV-2 or disease related to such infection such as COVID-19. Additional features relating to the use of the immunization agent and to the treatment course of the subject will be disclosed in the following description.

Administration “to the upper respiratory tract” includes any type of administration that results in delivery to the mucosa lining of the upper respiratory tract and includes in particular nasal administration. Administration to the upper respiratory tract includes without limitation, aerosol inhalation, nasal instillation, nasal insufflation, and all combinations thereof. In some embodiments the administration is by aerosol inhalation. In some embodiments the administration is by nasal instillation. In some embodiments the administration is by nasal insufflation.

According to a particular embodiment, the pseudotyped lentiviral vector particle encoding a SARS-CoV-2 S protein or a derivative thereof is for administration as intranasal mucosal boost or target immunization in a subject who received a prime administration with a vaccine composition against SARS-CoV-2 infection or disease selected from the group consisting of a protein, an mRNA, an adenovirus, an inactivated virus and a protein subunit vaccine composition against SARS-CoV-2 infection or disease, in particular a protein or an mRNA vaccine composition against SARS-CoV-2 infection or disease.

According to a particular embodiment, the pseudotyped lentiviral vector particle encoding a S protein of a SARS-CoV-2 or a derivative thereof for use in accordance with the embodiments disclosed herein is further characterized by the following features:

- the S protein is from a SARS-CoV-2 virus which is pathogenic for a human host, in particular (i) a S protein from a SARS-CoV-2 virus selected from the group consisting of SARS-CoV-2 Ancestral strain, D614G strain, Alpha strain, Beta strain, Gamma strain, Delta strain and Omicron strain, preferably from Beta strain or Omicron strain, more preferably from Beta strain, or (ii) a S protein from a variant of said Ancestral, D614G strain, Alpha, Beta, Gamma, Delta or Omicron strains wherein such variants encode a S protein with an amino acid sequence at least 90% identical to SEQ ID NO: 1 or, - the S protein is a derivative of the native S protein of one of the Ancestral, D614G, Alpha, Beta, Gamma, Delta or Omicron strains by mutation of 1 to 12, especially 1 to 6 amino acid residues, in particular by substitution and/or deletion of 1 to 12, especially 1 to 6 amino acid residues, in particular (a) is a stabilized form of the S protein characterized by substitution of two consecutive amino acid residues for Proline residues in the S2 domain of the S protein at positions provided as residues 986 and 987 by reference to SEQ ID NO: 1 in the amino acid sequence of the S protein or (b) is a prefusion form of the S protein by deletion of the furin site located from residue 675 to residue 685 by reference to SEQ ID NO: 1 in the amino acid sequence of the S protein or (c) is a stabilized prefusion form of the S protein by deletion of the furin site and substitution of two consecutive amino acid residues for Proline residues in the S2 domain at positions 986 and 987 by reference to SEQ ID NO: 1 .

In a particular embodiment, the S protein of the Ancestral strain of SARS-CoV-2 has an amino acid sequence of SEQ ID NO: 1 and the native sequence of the polynucleotide encoding the S protein of the Ancestral strain of SARS-CoV-2 is defined in SEQ ID NO: 2.

In another particular embodiment, the native sequence of the polynucleotide encoding the S protein of the D614G strain of SARS-CoV-2 comprising a mutation of the aspartic acid residue to the glycine residue at position 614 of the amino acid sequence of SEQ ID NO: 1 (D614G), a mutation of the lysine residue to the proline residue at position 986 of the amino acid sequence of SEQ ID NO: 1 (K986P) and a mutation of the valine residue to the proline residue at position 987 of the amino acid sequence of SEQ ID NO: 1 (V987P), i.e. mutation 2P, is defined in SEQ ID NO: 3 (SD614G-2P). The S protein of the D614G strain of SARS-CoV-2 comprising said mutation 2P (SD614G-2P) has an amino acid sequence of SEQ ID NO: 4.

In another particular embodiment, the native sequence of the polynucleotide encoding the S protein of the Alpha strain of SARS-CoV-2 comprising a mutation of the lysine residue to the proline residue at position 986 of the amino acid sequence of SEQ ID NO: 1 (K986P) and a mutation of the valine residue to the proline residue at position 987 of the amino acid sequence of SEQ ID NO: 1 (V987P), i.e. mutation 2P, is defined in SEQ ID NO: 5 (SAi _P ha-2p). The S protein of the Alpha strain of SARS-CoV-2 comprising said mutation 2P (SAi _P ha-2p) has an amino acid sequence of SEQ ID NO: 6.

In another particular embodiment, the native sequence of the polynucleotide encoding the S protein of the Beta strain of SARS-CoV-2 is defined in SEQ ID NO: 7 (Sseta). The S protein of the Beta strain of SARS-CoV-2 (Sseta) has an amino acid sequence of SEQ ID NO: 8. In another particular embodiment, the native sequence of the polynucleotide encoding the S protein of the Beta strain of SARS-CoV-2 comprising a mutation of the lysine residue to the proline residue at position 986 of the amino acid sequence of SEQ ID NO: 1 (K986P) and a mutation of the valine residue to the proline residue at position 987 of the amino acid sequence of SEQ ID NO: 1 (V987P), i.e. mutation 2P, is defined in SEQ ID NO: 9 (Ss _e ta-2 ). The S protein of the Beta strain of SARS-CoV-2 comprising said mutation 2P (Ss _e ta-2 ) has an amino acid sequence of SEQ ID NO: 10.

In another particular embodiment, the native sequence of the polynucleotide encoding the S protein of the Gamma strain of SARS-CoV-2 comprising a mutation of the lysine residue to the proline residue at position 986 of the amino acid sequence of SEQ ID NO: 1 (K986P) and a mutation of the valine residue to the proline residue at position 987 of the amino acid sequence of SEQ ID NO: 1 (V987P), i.e. mutation 2P, is defined in SEQ ID NO: 1 1 (SGamma-2p). The S protein of the Gamma strain of SARS-CoV-2 comprising said mutation 2P (SGamma-2p) has an amino acid sequence of SEQ ID NO: 12.

In another particular embodiment, the native sequence of the polynucleotide encoding the S protein of the Delta strain of SARS-CoV-2 comprising a mutation of the lysine residue to the proline residue at position 986 of the amino acid sequence of SEQ ID NO: 1 (K986P) and a mutation of the valine residue to the proline residue at position 987 of the amino acid sequence of SEQ ID NO: 1 (V987P), i.e. mutation 2P, is defined in SEQ ID NO: 13 (Soeita-2p). The S protein of the Delta strain of SARS-CoV-2 comprising said mutation 2P (Soeita-2p) has an amino acid sequence of SEQ ID NO: 14.

In another particular embodiment, the native sequence of the polynucleotide encoding the S protein of the Omicron strain of SARS-CoV-2 is defined in SEQ ID NO: 15 (Somicron). The S protein of the Omicron strain of SARS-CoV-2 (Somicron) has an amino acid sequence of SEQ ID NO: 16.

In another particular embodiment, the native sequence of the polynucleotide encoding the S protein of the Omicron strain of SARS-CoV-2 comprising a mutation of the lysine residue to the proline residue at position 986 of the amino acid sequence of SEQ ID NO: 1 (K986P) and a mutation of the valine residue to the proline residue at position 987 of the amino acid sequence of SEQ ID NO: 1 (V987P), i.e. mutation 2P, is defined in SEQ ID NO: 17 (Somicron- ₂ p). The S protein of the Omicron strain of SARS-CoV-2 comprising said mutation 2P (Somicron- ₂ p) has an amino acid sequence of SEQ ID NO: 18.

In another particular embodiment, the native sequence of the polynucleotide encoding the S protein of the Omicron BA.1 strain of SARS-CoV-2 is defined in SEQ ID NO: 23 (Somicron- BA.I). The S protein of the Omicron strain of SARS-CoV-2 (Somicron-BA.i) has an amino acid sequence of SEQ ID NO: 24.

In another particular embodiment, the native sequence of the polynucleotide encoding the S protein of the Omicron BA.4 or BA.5 strain of SARS-CoV-2 is defined in SEQ ID NO: 25 (Somicron-BA.4/5). The S protein of the Omicron BA.4 or BA.5 strain of SARS-CoV-2 (Somicron-BA.4/5) has an amino acid sequence of SEQ ID NO: 26.

In another particular embodiment, the native sequence of the polynucleotide encoding the S protein of the Ancestral strain of SARS-CoV-2 comprising a mutation of the lysine residue to the proline residue at position 986 of the amino acid sequence of SEQ ID NO: 1 (K986P) and a mutation of the valine residue to the proline residue at position 987 of the amino acid sequence of SEQ ID NO: 1 (V987P), i.e. mutation 2P, is defined in SEQ ID NO: 19 (S2P). The S protein of the Ancestral strain of SARS-CoV-2 comprising said mutation 2P (S2P) has an amino acid sequence of SEQ ID NO: 20.

In a particular embodiment, the pseudotyped lentiviral vector particle encodes the S protein of the Beta strain of SARS-CoV-2 comprising the mutation 2P (Sseta-2 ) that is encoded by the vector pFlap-ieCMV-S-B351 -2P-WPREm that has been deposited at the COLLECTION NATIONALE DE CULTURES DE MICROORGANISMES (CNCM) located at Institut Pasteur, 25-28 rue du Docteur Roux, 75724 Paris Cedex 15 FRANCE, on July 6, 2021 under N°CNCM 1-5710.

Also provided is the vector pFlap-ieCMV-S-B351 -2P-WPREm (CNCM 1-5710). The nucleotide sequence of pFlap-ieCMV-S-B351 -2P-WPREm is defined in SEQ ID NO: 22.

Also provided is a host cell comprising the vector pFlap-ieCMV-S-B351 -2P-WPREm (CNCM 1-5710 or SEQ ID NO: 22).

Also provided is a pseudotyped lentiviral vector particle encoding the S protein of the Beta strain of SARS-CoV-2 comprising the mutation 2P (Ss _e ta-2 ), wherein the pseudotyped lentiviral vector particle is made by a method comprising co-transfection of a host cell with the vector pFlap-ieCMV-S-B351 -2P-WPREm (CNCM 1-5710 or SEQ ID NO: 22).

In a particular embodiment, the pseudotyped lentiviral vector particle encoding a S protein of a SARS-CoV-2 or a derivative thereof for use in accordance with the embodiments disclosed herein, is further characterized by the following features: the amino acid sequence of the S protein is SEQ ID NO: 1 or is a derivative thereof having an amino acid sequence at least 90% identical to SEQ ID NO: 1 , and the derivative of the S protein of SARS-CoV-2 comprises at least five amino acid mutations including (i) a mutation of the lysine residue to the asparagine residue at position 417 of the amino acid sequence of SEQ ID NO: 1 (K417N), (ii) a mutation of the glutamic acid residue to the lysine residue at position 484 of the amino acid sequence of SEQ ID NO: 1 (E484K) or a mutation of the glutamic acid residue to the alanine residue at position 484 of the amino acid sequence of SEQ ID NO: 1 (E484A), (iii) a mutation of the asparagine residue to the tyrosine residue at position 501 of the amino acid sequence of SEQ ID NO: 1 (N501 Y), (iv) a mutation of the lysine residue to the proline residue at position 986 of the amino acid sequence of SEQ ID NO: 1 (K986P) and (v) a mutation of the valine residue to the proline residue at position 987 of the amino acid sequence of SEQ ID NO: 1 (V987P).

In a particular embodiment, the pseudotyped lentiviral vector particle encoding a SS protein of a SARS-CoV-2 or a derivative thereof for use according to the invention, is such that the S protein of SARS-CoV-2 further comprises amino acid mutations selected from the group consisting of (vi) a mutation of the glycine residue to the serine residue at position 446 of the amino acid sequence of SEQ ID NO: 1 (G446S), (vii) a mutation of the threonine residue to the lysine residue at position 478 of the amino acid sequence of SEQ ID NO: 1 (T478K), (viii) a mutation of the glutamine residue to the arginine residue at position 493 of the amino acid sequence of SEQ ID NO: 1 (Q493R) and (ix) a mutation of the glutamine residue to the arginine residue at position 498 of the amino acid sequence of SEQ ID NO: 1 (Q498R).

In a particular embodiment, the pseudotyped lentiviral vector particle encoding a S protein of a SARS-CoV-2 or a derivative thereof for use according to the invention is such that the encoded mutated S protein of SARS-CoV-2 has the amino acid sequence of SEQ ID NO: 10 or SEQ ID NO: 18, preferably of SEQ ID NO: 10.

In another particular embodiment, the pseudotyped lentiviral vector particle encoding a S protein of a SARS-CoV-2 or a derivative thereof for use according to the invention is such that the encoded mutated S protein of SARS-CoV-2 has the amino acid sequence of SEQ ID NO: 24 or SEQ ID NO: 26.

In a particular embodiment, the pseudotyped lentiviral vector particle encoding a S protein of a SARS-CoV-2 or a derivative thereof for use according to the invention is pseudotyped with a vesicular stomatitis virus glycoprotein G (VSV-G) protein.

In particular the VSV-G protein is advantageously provided by a VS virus of the Indiana strain or the New-Jersey strain.

In a particular embodiment, the pseudotyped lentiviral vector particle encoding a S protein of a SARS-CoV-2 or a derivative thereof for use according to the invention is such that the pseudotyped lentiviral vector particle is non-integrative, non-cytopathic and non-replicative.

In accordance with the herein disclosed use of the immunogenic agent according to the invention, in some embodiments, the immunogenic agent or composition comprising the agent is for use in a method of prevention of infection of a human subject by SARS-CoV-2. In some embodiments, the immunogenic agent or composition is for use in a method of protection against SARS-CoV-2 replication in a human subject at risk of being exposed to SARS-CoV-2 or infected by SARS-CoV-2. In some embodiments, the immunogenic composition is for use in a method of preventing development of symptoms or development of a disease associated with infection by SARS-CoV-2, such as COVID-19 in a human subject at risk of being exposed to SARS-CoV-2 or infected by SARS-CoV-2. In some embodiments, the immunogenic composition is for use in a method of preventing the onset of neurological outcome associated with infection by SARS-CoV-2 in a human subject at risk of being exposed to SARS-CoV-2 or infected by SARS-CoV-2. In some embodiments, the immunogenic composition is for use in a method of protecting the Central Nervous System (CNS) of a human subject at risk of being exposed to SARS-CoV-2 or infected by SARS-CoV-2. In some embodiments the vaccine provides protection against the infection by SARS-CoV-2, especially sterilizing protection.

In any of these applications for use in a method disclosed, the immunogenic agent or composition is to be administered to the subject as a prophylactic agent in a boost or target administration step in an effective amount for administration to the upper respiratory tract in order to elicit an immune response against SARS-CoV-2.

In some embodiment the immunogenic composition is for use in a method of protection of a human subject against SARS-CoV-2 infection or against development of the symptoms or the COVID-19 disease associated with SARS-CoV-2 infection, wherein the subject is at risk of developing lung and/or CNS pathology. In particular the human subject is in need of immune protection of CNS from SARS-CoV-2 replication because he/she is affected with comorbid conditions, in particular comorbid conditions affecting the CNS.

In a particular embodiment, the pseudotyped lentiviral vector particle encoding a S protein of a SARS-CoV-2 or a derivative thereof for use according to any one of the embodiments disclosed herein is administered in a subject selected from the group consisting of (a) a subject that has previously received a vaccine composition against SARS-CoV-2 infection or disease selected from the group consisting of a protein, an mRNA, an adenovirus, an inactivated virus and a protein subunit vaccine composition against SARS-CoV-2 infection or disease, in particular a protein- or an mRNA-based vaccine against SARS-CoV-2 infection or disease as a systemic prime and/or boost administration(s) such as intramuscular, intradermal or sub-cutaneous administration(s), in particular an intramuscular prime and/or boost administration(s), (b) a subject that has received a systemic prime administration such as intramuscular, intradermal or sub-cutaneous administration(s), in particular an intramuscular prime administration, of a vaccine composition against SARS-CoV-2 infection or disease selected from the group consisting of a protein, an mRNA, an adenovirus, an inactivated virus and a protein subunit vaccine composition against SARS-CoV-2 infection or disease, in particular a protein- or an mRNA-based vaccine against SARS-CoV-2 infection or disease and has then recovered from coronavirus disease such as coronavirus disease 2019 (COVID-19), (c) a subject that has first recovered from coronavirus disease such as COVID- 19 and has then received a systemic prime administration such as intramuscular, intradermal or sub-cutaneous administration(s), in particular an intramuscular administration, of a protein- or an mRNA-based vaccine against SARS-CoV-2 infection or disease, and (d) a subject that has received more than two, in particular more than three, systemic administrations such as intramuscular, intradermal or sub-cutaneous administration(s), in particular an intramuscular administration, of a protein- or an mRNA-based vaccine against SARS-CoV-2 infection or disease.

In a particular embodiment, the pseudotyped lentiviral vector particle encoding a S protein of a SARS-CoV-2 or a derivative thereof is for use in a prime/boost or a target immunization regimen for elicitation of a long-lasting protective mucosal humoral immune response and/or a long-lasting mucosal cellular immune response against SARS-CoV-2 infection or disease, wherein said response protects the respiratory system and/or the CNS of the subject.

In a particular embodiment, the pseudotyped lentiviral vector particle encoding a S protein of a SARS-CoV-2 or a derivative thereof is for use in an immunization regimen wherein the pseudotyped lentiviral vector particle elicits a CD8 ⁺ T-cell response against SARS-CoV-2.

In a particular embodiment, the pseudotyped lentiviral vector particle encoding a S protein of a SARS-CoV-2 or a derivative thereof is for use in an immunization regimen wherein the pseudotyped lentiviral vector particle elicits lung-resident memory CD8 ⁺ T cells (Trm) and/or effector CD8 ⁺ T cells (Tc1 ) specific to Spike and able to produce Interferon-gamma (IFN-y)/Tumor Necrosis Factor (TNF)/lnterleukin-2 (IL-2) cytokines. In a particular embodiment, the pseudotyped lentiviral vector particle encoding a S protein of a ARS-CoV-2 or a derivative thereof for use according to the invention is used in an immunization regimen wherein the subject has a waning immunity from week 12 after the first injection of the initial vaccination with a vaccine composition against SARS-CoV-2 infection or disease selected from the group consisting of a protein, an mRNA, an adenovirus, an inactivated virus and a protein subunit vaccine composition against SARS-CoV-2 infection or disease, in particular a protein- or an mRNA-based vaccine against SARS-CoV-2 infection or disease or post SARS- CoV-2 disease recovery, in particular post-COVID-19 recovery.

In pre-clinical results in mice, the inventors showed that 4 to 5 months after the last administration of an mRNA-based vaccine expressing the Spike antigen, there was no longer protective immunity against SARS-CoV-2 ( Vesin et al., 2022, Mol Ther 30, 2984-2997). Furthermore, it is now well established that the level of neutralizing anti-Spike antibodies is significantly reduced in the serum of vaccinated individuals 3 to 10 months after the last administration of an mRNA-based vaccine expressing the Spike antigen (Decru et al., 2022, Front Immunol. 13, 909910).

It is also widely established that after pre-exposure of the immune system to an antigen, including SARS-CoV-2 Spike, in the context of vaccination - regardless of vaccination strategy - or in the context of infection, a memory immunity is generally induced in the B and T cell compartments ( Valyl -Nagy et al., 2022, I nt J Mol Sci, 23.10.3390). As far as can be currentlyassessed, in the case of anti-Spike immunity, induced by vaccination or infection, such memory immunity is expected to last on average at least one year (Gallais et al., 2021, EBioMedicine, 71, 103561). This memory immunity will be boostable, at least until one year after the last administration of an mRNA vaccine.

In a particular embodiment, the pseudotyped lentiviral vector particle encoding a S protein of a SARS-CoV-2 or a derivative thereof for use according to the invention is administered in a subject according to the invention as an intranasal mucosal boost or target immunization at least 3 months, in particular from 3 to 24 months, preferably from 3 to 12 months, after the last contact with SARS-CoV-2 or administration of a vaccine composition against SARS-CoV-2 infection or disease selected from the group consisting of a protein, an mRNA, an adenovirus, an inactivated virus and a protein subunit vaccine composition against SARS-CoV-2 infection or disease, in particular a protein or an mRNA vaccine composition against SARS-CoV-2 infection or disease.

In a particular embodiment, the pseudotyped lentiviral vector particle encoding a S protein of a SARS-CoV-2 or a derivative thereof for use according to the invention is formulated as a liquid composition or a dry powder for an administration as intranasal aerosols, intranasal drops or intranasal insufflations. In a particular embodiment, the pseudotyped lentiviral vector particle encoding a S protein of a SARS-CoV-2 or a derivative thereof for use according to the invention, is used in an immunization regimen wherein the administration regimen comprises administration of one or more dosage form(s) of the pseudotyped lentiviral vector particle wherein the dose of each dosage form is from 10 ⁷ to 10 ⁹ Transduction Unit (TU).

According to another aspect the invention also concerns an immunogenic composition comprising a pseudotyped lentiviral vector particle encoding a S protein of a SARS-CoV-2 or a derivative thereof and a pharmaceutically acceptable carrier, wherein the pseudotyped derivative of the S protein of SARS-CoV-2 comprises at least nine amino acid mutations including (i) a mutation of the lysine residue to the asparagine residue at position 417 of the amino acid sequence of SEQ ID NO: 1 (K417N), (ii) a mutation of the glutamic acid residue to the alanine residue at position 484 of the amino acid sequence of SEQ ID NO: 1 (E484A), (iii) a mutation of the asparagine residue to the tyrosine residue at position 501 of the amino acid sequence of SEQ ID NO: 1 (N501 Y), (iv) a mutation of the lysine residue to the proline residue at position 986 of the amino acid sequence of SEQ ID NO: 1 (K986P), (v) a mutation of the valine residue to the proline residue at position 987 of the amino acid sequence of SEQ ID NO: 1 (V987P), (vi) a mutation of the glycine residue to the serine residue at position 446 of the amino acid sequence of SEQ ID NO: 1 (G446S), (vii) a mutation of the threonine residue to the lysine residue at position 478 of the amino acid sequence of SEQ ID NO: 1 (T478K), (viii) a mutation of the glutamine residue to the arginine residue at position 493 of the amino acid sequence of SEQ ID NO: 1 (Q493R) and (ix) a mutation of the glutamine residue to the arginine residue at position 498 of the amino acid sequence of SEQ ID NO: 1 (Q498R).

This immunogenic composition may be such that the pseudotyped lentiviral vector particle encodes mutated S protein of SARS-CoV-2 the amino acid sequence of which is SEQ ID NO: 18.

According to another embodiment, this immunogenic composition may be such that the pseudotyped lentiviral vector particle encodes mutated S protein of SARS-CoV-2 the amino acid sequence of which is SEQ ID NO: 24 or SEQ ID NO: 26.

According to a further embodiment, the immunogenic composition is formulated for intranasal administration as disclosed in the embodiments herein.

The invention also concerns kits suitable for use in practicing a use or a method disclosed herein. In some embodiments the kit comprises a dosage form for administration to the upper respiratory tract of a subject of the pseudotyped lentiviral vector particle encoding a SARS-CoV-2 S protein or a derivative thereof according to this disclosure, and an applicator. In some embodiments the applicator is an applicator for aerosol inhalation. In some embodiments the applicator is an applicator for nasal instillation. In some embodiments the applicator is an applicator for nasal insufflation. Suitable examples of each are known in the art and may be used.

Preparation of recombinant LV particles is known in the art, including to obtain non- integrative, non-replicative recombinant LV particles. In particular reference is made to the disclosure in Ku MW, etal. Cell Host Microbe, 29(2), 236-249 e236, 2021; Ku MWetal, EMBO Mol Med, e14459, 2021, Ku MW, Charneau P, Majlessi L. Expert Rev Vaccines, 1-16, 2021). Polynucleotide constructs may be adapted with the sequence encoding the selected Spike protein or derivative thereof. In some embodiments, the lentiviral vector particle comprises HIV-1 Gag and Pol proteins. In some embodiments, the lentiviral vector particle comprises subtype D, especially HIV-1 NDK, Gag and Pol proteins.

According to some embodiments, the lentivector particles are obtained in a host cell transformed with a DNA plasmid.

Such a DNA plasmid can comprise:

- bacterial origin of replication (ex: pUC ori);

- antibiotic resistance gene (ex: AmpiR or KanR) for selection; and more particularly:

- a lentiviral vector comprising at least one nucleic acid encoding a SARS-CoV-2 S protein or a derivative thereof, transcriptionally linked to a promoter, for example a CMV promoter.

Such a method allows producing a recombinant vector particle for use according to the invention, comprising the following steps of: i) transfecting a suitable host cell with a lentiviral vector; ii) transfecting said host cell with a packaging plasmid vector, containing viral DNA sequences encoding at least structural and polymerase (+ integrase) activities of a retrovirus (preferably lentivirus); Such packaging plasmids are described in the art (Dull et al., 1998, J Virol, 72(1 1 ):8463-71 ; Zufferey et a!., 1998, J V /'o/ 72(12):9873-80); iii) culturing said transfected host cell in order to obtain expression and packaging of said lentiviral vector into lentiviral vector particles; and iv) harvesting the lentiviral vector particles resulting from the expression and packaging of step iii) in said cultured host cells.

An appropriate host cell is preferably a human cultured cell line as, for example, a HEK cell line, such as a HEK293T line.

Alternatively, the method for producing the vector particle is carried out in a host cell, which genome has been stably transformed with one or more of the following components: a lentiviral vector DNA sequence, the packaging genes, and the envelope gene. Such a DNA sequence may be regarded as being similar to a proviral vector according to the invention, comprising an additional promoter to allow the transcription of the vector sequence and improve the particle production rate.

In a preferred embodiment, the host cell is further modified to be able to produce viral particle in a culture medium in a continuous manner, without the entire cells swelling or dying. One may refer to Strang et al., 2005, J Virol 79(3):1165-71 ; Relander et al., 2005, Mol Ther 1 1 (3):452-9; Stewart et al., 2009, Gene Ther, 16(6) :805-14; and Stuart et al., 2011 , Hum gene Ther, with respect to such techniques for producing viral particles. The lentiviral particle vectors can comprise the following elements, as previously defined:

- cPPT/CTS polynucleotide sequence; and

- a nucleic acid encoding a CAR under control of a p2m or Major Histocompatibility Complex classe I (MHC-I) promoter, and optionally one of the additional elements described above.

Other features and advantages of the invention will be apparent from the examples which follow and will also be illustrated in the figures.

BRIEF DESCRIPTION OF THE FIGURES

Figure 1 : Down-selection of a Scov-2 variant with the highest potential to induce cross sero-neutralizing antibodies. (A) Timeline of prime-boost vaccination with LV::SAi _P ha, LV::Sseta or LV::SGamma and (cross) sero-neutralization assays in C57BL/6 mice (n = 4-5/group). (B) Half maximal Effective Concentration (EC50) of neutralizing activity of sera from vaccinated mice was evaluated before and after the boost, against pseudo-viruses carrying Scov-2 from D614G, Alpha, Beta or Gamma variants. (C) EC50 of sera from C57BL/6 mice, vaccinated following the regimen detailed in (A) with LV encoding for SDGUG, either WT or carrying the K ⁹⁸⁶ p - v ⁹⁸⁷ P substitutions in the S2 domain. EC50 was evaluated before and after the boost, as indicated in (B).

Figure 2: Anti-Scov-2 humoral responses in mRNA-1273-vaccinated mice which were further intranasally boosted with LV::SBeta-2P- (A) Timeline of mRNA-1273 intramuscular (i.m.)-i.m. prime-boost vaccination in C57BL/6 mice which were later immunized i.n. by escalating doses of LV::Sseta-2P (n = 4-5/group) and the (cross) sero-neutralization followup. (B) Serum EC50 determined at the indicated time points against pseudo-viruses carrying Scov-2 from D614G, Alpha, Beta, Gamma, Delta or Delta-i- variants. (C) Anti-Scov-2 IgG or IgA in the sera two weeks after late LV::Sseta-2P i.n. or mRNA-1273 (Moderna) i.m. boost. Statistical significance was determined by Mann-Whitney test (*= p < 0.05, **= p < 0.01 , ***= p < 0.001 ).

Figure 3: Lung resident memory B-cell subsets in mRNA-1273-vaccinated mice which were further intranasally boosted with LV::SBeta-2P. The mice are those detailed in Figure 2. Mucosal immune cells were studied two weeks after LV::Sseta-2P i.n. boost. (A) Cytometric gating strategy to detect lung B resident memory cells, and (B) percentages of Brm among surface IgM/lgD B cells in various groups. Figure 4: Systemic CD8 ⁺ T-cell responses to Scov-2 in mRNA-1273-vaccinated mice which were further intranasally boosted with LV::SBeta-2P- The mice are those detailed in Figure 2. T-splenocyte responses were evaluated two weeks after LV::Sseta-2P i.n. boost by IFN-y ELISPOT after stimulation with S:256-275, S:536-550 or S:576-590 synthetic 15-mer peptides encompassing Scov-2 MHC-l-restricted epitopes, pertinent in C57BL/6 (H-2 ^b ) mice. Statistical significance was evaluated by Mann-Whitney test (*= p < 0.05).

Figure 5: Mucosal CD8 ⁺ T-cell responses to Scov-2 in mRNA-1273-vaccinated mice which were further intranasally boosted with LV::SBeta-2P- The mice are those detailed in Figure 2. (A) Representative IFN-y response by lung CD8 ⁺ T cells detected by IntraCellular Cytokine Staining (ICS) after in vitro stimulation with a pool of S:256-275, S:536-550 and S:576-590 peptides. Cells are gated on alive CD45 ⁺ CD8 ⁺ T cells.

Figure 6. Lung resident memory T-cell subset in mRNA-1273-vaccinated mice which were further intranasally boosted with LV::SBeta-2P. The mice are those detailed in Figures 2 and 3. Mucosal immune cells were studied two weeks after LV::Sseta-2P i.n. boost. (A) Cytometric gating strategy to detect lung CD8 ⁺ T resident memory cells (Trm, CD44 ⁺ CD62L ^_ CD69 ⁺ CD103 ⁺ ), and (B) percentages of Trm among CD8 ⁺ CD44 ⁺ T-cells in various groups.

Figure 7. Full protective potential of a late LV::SBeta-2P i.n. boost in mRNA-1273- primed and-boosted mice. (A) Timeline of mRNA-1273 i.m.-i.m. prime-boost vaccination in C57BL/6 mice which were later immunized i.n. with the sub-optimal dose of 1 x 10 ⁸ TU/mouse of LV::Sseta-2P (n = 4-5/group), pre-treated i.n. with Ad5::hACE-2 4 days before i.n. challenge with 0.3 x 10 ⁵ TCID50 (Tissue Culture Infectious Dose of 50 %) of SARS-CoV-2 Delta variant. (B) Comparative quantitation of hACE-2 mRNA in the lungs of different groups of mice pretreated i.n. with Ad5::hACE-2. (C) Viral RNA contents, evaluated by conventional E _Co v-2- specific (top) or sub-genomic Esg-specific (bottom) qRT-PCR at 3 dpi. The bottom line indicates the detection limits. Statistical significance was evaluated by Mann-Whitney test (*= p < 0.05, **= p < 0.01 ).

Figure 8. Anti-Scov-2 humoral responses in mRNA-1273-vaccinated mice which were further intranasally boosted with LV::SBeta-2P. (A) Follow-up of anti-Scov-2 (left) and anti-RBD (right) IgG in the sera of mice initially primed and boosted i.m. with mRNA-1273. (B) Anti-Scov-2 IgG (top), anti-RBD IgG (middle), and anti-Scov-2 I A (bottom) in the total lung extracts of mice initially primed and boosted i.m. with mRNA-1273 and then boosted later with a third i.m. dose of mRNA-1273 or an i.n. boost of LV::Sseta-2P.

Figure 9. Absence of mucosal CD8 ⁺ Tc2 responses to Scov-2 in mRNA-1273- vaccinated mice which were further intranasally boosted with LV: :SBeta-2P- The mice are those detailed in Figure 2. Absence of IL-4, IL-5, IL-10 and IL-13 production by lung CD8 ⁺ T cells after in vitro stimulation with a pool of S:256-275, S:536-550 and S:576-590 peptides, studied by ICS in parallel to the assay performed to detect IFN-y/TNF/IL-2 (see Figure 5). Cells are gated on alive CD45 ⁺ CD8 ⁺ T cells.

Figure 10. Maps of plasmids used for production of LV encoding (A) S _D6 I4G-2P, (B) SAIpha-2P; (C) Scamma-2P, (D) Soelta-2P, (E) S[3eta-2P and (F) Somicron-2P antigens.

Figure 11. Amino acid sequence of Spike from (A) Omicron BA.1 (top) or (B) Omicron BA 4/5 (bottom) sub-variants. Sequences indicated in bold and underlined are murine MHC-I- restricted T-cell epitopes in H-2 ^b mice (Ku MW, et al. EMBO Mol Med, e14459, 2021). Sequences highlighted in gray are MHC-I or -Il-restricted human T-cell epitopes identified in HHD-DR1 MHC-humanized mice.

Figure 12. Humoral immunity in hamsters immunized i.m. with various LV::S. (A) Schematic representation of LV encoding Scov-2 proteins from either ancestral WA1 or Beta SARS-CoV-2 strain. Codon-optimized sequences encoding Scov-2 were cloned into the pFLAP lentiviral vector plasmid, under the control of human Pcwvie promoter; RRE, rev response element; cPPT, central polypurine tract. The LV::SWAI includes the entire sequence of Scov-2. RBD, S1/S2, S2’ cleavage sites, 675 ^QTQTNSPRRAR 685 sequence encompassing RRAR furin cleavage site of SEQ ID NO: 27, and K ⁹⁸⁶ P, V ⁹⁸⁷ P, K ⁴¹⁷ N, E ⁴⁸⁴ K and N ⁵⁰¹ Y substitutions are indicated. (B) Western blot analysis to detect expression of SWAI , S WAI-2P, S WAI-AF-2P, and Ss _e ta2- P in LV-transduced 293T cells. Total cell lysates were analyzed under non-reduced conditions using an anti-S2 rabbit polyclonal antibody. LV::GFP was included as negative control. Full length Spike (S) and S2 subunit are indicated. (C) Syrian golden hamsters (n = 4/group) were immunized i.m. with 1 x 10 ⁸ TU of LV::SWAI , LV: :SWAI-2P, LV: :SWAIAF-2P or LV::Sseta-2P. Five weeks later, serum anti-SwAi responses expressed as mean endpoint dilution titers, were determined by ELISA. Errors bars represent the standard error of the mean (SEM). The statistical significance of differences was determined by Kruskal-Wallis test followed by Dunn’s multiple comparisons test and were found not significant. Dotted lines indicate the limit of detection (LOD).

Figure 13. Humoral immunity in hamsters following LV::S administration. (A) Timeline of the LV::S prime-boost vaccination regimen and WA1 SARS-CoV-2 challenge in hamsters (n = 6/group). (B) Serum anti-SwAi or -RBDWAI IgG responses expressed as mean endpoint dilution titers, determined by ELISA. (C) Neutralizing activity (EC50) of sera, taken prior the WA1 SARS-CoV-2 challenge, or of lung homogenates, taken at 4 dpi as determined by use of pseudoviruses harboring Scov-2 from D614G SARS-CoV-2 variant. Data are presented as mean ± SEM. Asterisks indicate the significance of differences between the groups, p-values were determined by using Kruskal-Wallis tests followed by Dunn’s multiple comparisons tests; *p < 0.05, **p < 0.01 . Only significant differences are shown. Dotted lines indicate the LOD.

Figure 14. Single i.n. LV::S injection fully protects hamsters against WA1 SARS- CoV-2. Hamsters are those described in the legend to the Figure 13. (A) Lung viral loads quantitated by total E (left) or Esg qRT-PCR (right) at 4 dpi. Bars represent geometric means. (B) Percentages of weight loss in LV::S- or LV ctrl-vaccinated hamsters at 4 dpi. (C) Expression of inflammatory cytokines in lung tissues after challenge. The heatmap recapitulates relative Iog2 fold changes in the expression of inflammation-related mediators in LV::S vaccinated or LV ctrl-administered individuals, as analyzed at 4 dpi by use of RNA extracted from total lung homogenates and normalized versus samples from untreated controls. Six individual hamsters per group are shown in the heatmap. Statistical differences between LV:: S and LV Ctrl groups were determined by Kruskal-Wallis test followed by Dunn’s multiple comparisons test and are indicated by asterisks; *p < 0.05; **p < 0.01 ; ***p < 0.001 . Comparisons were made between vaccinated groups and LV Ctrl.

Figure 15. Single i.n. LV:: S injection largely reduced lung histopathogy. (A) Lung histological H&E analysis, as studied at 4 dpi. Heatmap recapitulating the histological scores, for: 1 ) inflammation score and 2) interstitial syndrome. (B) Representative alveolo-interstitial syndrome and (C) severe inflammation in an LV ctrl-injected and infected hamster. Here the structure of the organ is largely obliterated, while remnants of alveolar spaces and bronchiolar lumens can be seen. (D-F) bronchiolar lesions in LV ctrl-immunized animals. Shown are epithelial cells and cell debris in the bronchiolar lumen (black arrows) (D), papillary projections of the bronchiolar epithelium into the lumen (star) (E) and degenerative lesions with effacement of the epithelium (green arrow) (F). (G) Mild alveolar infiltration in a vaccinated hamster. Some of the alveoli (arrow) are partially or completely filled with cells and an eosinophilic exudate. (H) Representative Ncov- ₂ -specific IHC image performed on lungs of SARS-CoV-2-infected hamsters. Lower panels show enlarged views from upper panels. Scale bars are 1 mm for upper panels and 25 pm for lower panels.

Figure 16. Decreased SARSCoV-2 omicron infectious virus in lungs and nasal turbinates by i.n administration of a single or booster dose of LV::SBeta-2P. (A) Timeline of single or prime-boost vaccination and Omicron SARS-CoV-2 challenge. Hamsters (n= 4- 5/group) were primed i.n. or i.m. with 1 x 10 ⁸ TU of LV::Sseta-2P. Three weeks later, some of them were boosted i.n. with the same amount of LV::Sseta-2P or LV Ctrl. Serum samples were collected at wks 3 and 7 for serological analyses. (B) Serum or anti-Somicron (upper panels) or -RBDomicron (lower panels) IgG responses, expressed as mean endpoint dilution titers, determined by ELISA. Data are presented as mean ± SEM. Percentages of weight loss in LV::Sseta -2P-, or LV ctrl-vaccinated hamsters following challenge (C) and at 4 dpi (D). (E) Lung and (F) NT viral loads were quantitated by Esg qRT-PCR at 4 dpi. Bars represent geometric means. Statistical differences were determined by Kruskal-Wallis tests followed by Dunn’s multiple comparisons test and are indicated by asterisks. *p < 0.05, **p < 0.01 , ***p < 0.001 . Dotted lines indicate the LOD.

Figure 17. Immunodetection of the NCoV-2 antigen performed on lungs of Omicron SARS-CoV-2-infected hamsters. Hamsters are those described in Figure 16. (A) One example of each vaccination regimen is shown at low magnification. Solid arrows denote foci of inflammatory infiltrates, and dotted arrows areas where the immunodetection signal is discernable even at this low magnification. (B) The close-up views depict the concentration of viral antigen (brown) within the inflammatory foci (bottom), while areas harboring no or little inflammation (top) display only scarce staining.

Figure 18. Robust humoral responses in hamsters vaccinated by LV: :SWAI-2P or LV::SBeta-2P prime (i.m.) - LV::SBeta-2P boost (i.n.). (A) Timeline of prime-boost vaccination. Hamsters (n= 4/group) were primed i.m. with 1 x 10 ⁸ TU of LV::SWAI-2P or LV::Sseta-2P. Five weeks later, hamsters were boosted i.n. with the same amount of LV::Sseta-2P. (B) Serum anti- SWAI or -RBDWAI (left panels) or anti-Somicron or -RBDomicron (right panels) IgG responses, expressed as mean endpoint dilution titers ± SEM, determined by ELISA. Statistical differences are indicated by asterisks. *p < 0.05. Figure 19. Anti-Scov-2 antibody imprinting in hamsters vaccinated by LV::S prime (i.m.) - LV::S _Be ta-2P boost (i.n.). Hamsters are those described in the legend to the Figure 18. (A-C) EC50 determined by use of pseudo-viruses carrying Scov-2 from D614G, Alpha, Beta, Gamma, Delta or Omicron variants. EC50 (A) 5 wks post prime in sera or 2 wks post boost in sera (B) and in lung homogenates (C). Data are expressed as the geometric mean EC50. Statistical significances were analyzed using two-way ANOVA followed by Sidak’s multiple comparisons test; **p < 0.01 ; ****p < 0.0001 . Dotted lines indicate the lower limit of detection (LOD).

Figure 20. Full protective capacity of LV::SBeta-2P used in a prime (i.m.) boost (i.n.) regimen against Omicron variant. (A) Timeline of prime-boost vaccination and i.n. challenge with 0.3 x 10 ⁵ TCID50 of SARS-CoV-2 Omicron variant in B6.K18-hACE2 ^{IP THV} transgenic mice (n = 5/group). (B) Lung viral RNA contents, evaluated by sub-genomic Esg-specific qRT-PCR at 5 dpi. Bottom lines indicate the detection limits. Statistical significance was evaluated by Mann-Whitney test (*= p < 0.05, **= p < 0.01 ).

The LV-based strategy, which is highly productive, not only in inducing humoral responses but also and particularly in establishing high quality and memory T-cell responses (Ku MW, etal. Commun Biol, 4(1), 713, 2021), is a favorable platform for a heterologous boost, even if it is also largely efficacious by its own as a primary COVID-19 vaccine candidate (Ku MW, et al. Cell Host Microbe, 29(2), 236-249 e236, 2021; Ku MW, et al. EMBO Mol Med, e14459, 2021). Furthermore, and importantly, LV is non-cytopathic, non-replicative and scarcely inflammatory and thus can be used to perform non-invasive i.n. boost, to efficaciously induce sterilizing mucosal immunity which protects the respiratory system, as well as the CNS (Ku MW, et al. Cell Host Microbe, 29(2), 236-249 e236, 2021; Ku MW, et al. EMBO Mol Med, e14459, 2021). The i.n. route of vaccination has been shown by several teams to be the most efficacious route at reducing viral contents in nasal swabs and nasal olfactory neuroepithelium (Bricker TL, Cell Rep, 36(3), 109400, 2021; Hassan AO, et al. Cell Rep Med, 2(4), 100230, 2021). The inventors thus hypothesized that i.n. vaccination could contribute efficaciously in blocking/reducing the respiratory chain of SARS-CoV-2 transmission.

Another major advantage of LV-based immunization is the induction of strong T-cell immune responses with high cross-reactivity of T-cell epitopes from Spike of diverse VOCs. Therefore, when the neutralizing antibody fails or wanes, the T-cell arm remains largely protective, as the inventors recently described in antibody-deficient, B-cell compromised pMT KO mice (Ku MW, et al. EMBO Mol Med, e14459, 2021). This property is relative to a high- quality and long-lasting T-cell immunity induced against multiple preserved T-cell epitopes, despite the mutation accumulated in the Spike of the emerging VOCs (Ku MW, et al. EMBO Mol Med, e14459, 2021).

In the present invention, the inventors first down-selected Sseta antigen which induced the greatest neutralization breadth against the VOCs and designed a non-integrating LV encoding a stabilized version of this antigen. In mRNA-1273-primed and -boosted mice with waning (cross) sero-neutralization capacity, the inventors used escalating doses of LV::Sseta- ₂ p in i.n. boost. The inventors demonstrated a dose-dependent increase in anti-Scov-2 I G and IgA titers, and a broadened sero-neutralization potential both in the sera and lung homogenates against VOCs. No anti-Scov-2 IgA was detected in the lungs of mice injected i.m. with the third dose of mRNA-1273. Increasing proportions of lung surface IgM/D- B cells with CD38 ⁺ CD73 ⁺ CD62L ⁺ CD69 ⁺ CD80 ⁺ phenotype, which have been associated with resident memory characteristic, were detected in a dose-dependent manner, in these mice. No such increases in this B-cell subset were detected in their counterparts which received an additional mRNA-1273 dose via i.m.

Spike-specific, effector lung CD8 ⁺ Tc1 cells were largely detected in the initially mRNA- 1273-primed and boosted mice which received a late i.n. LV::Sseta-2P boost. These lung CD8 ⁺ T cells did not display Tc2 phenotype. Increasing proportions of lung CD8 ⁺ CD44 ⁺ CD69 ⁺ CD103 ⁺ T rm were also detected, in a dose-dependent manner, only in LV::Sseta-2P i.n. boosted mice, but not in their counterparts boosted i.m. with mRNA-1273. The systemic CD8 ⁺ T-cell responses against various immunogenic regions of Scov-2were also increased with 1 x 10 ⁸ or 1 x 10 ⁹ TU of LV::Sseta-2P i.n. boost in initially mRNA-1273-primed and boosted mice. The highest i.n. dose of LV::Sseta-2P was comparable to the additional i.m. dose of mRNA-1273. The fact that the i.n. administration of LV::Sseta-2P has a boost effect on the systemic T-cell immunity represents another advantage of this vaccination regimen.

Evaluation of the protective potential of the lungs of mice primed and boosted with mRNA-1273, without additional boosting, showed that 20 wks after the first injection of mRNA- 1273, there was no detectable protective capacity left against the Delta variant of SARS-CoV- 2. In these mice, an i.n. booster injection of suboptimal dose, i.e., 1 x 10 ⁸ TU of LV::Sseta-2P completely inhibited SARS-CoV-2 replication in the lungs. A third late i.m. booster injection of mRNA-1273 reduced SARS-CoV-2 RNA content in the lungs in a similar manner, but did not completely inhibit viral replication in all mice.

The lack of protective potential against the Delta variant 4 months after prime-boost to mRNA-1273 could be explained by the hypothesis that adaptive immune memory is likely to be located in secondary lymphoid organs, l.e., distant anatomical sites far from the respiratory tract. In this context, the extraordinarily rapid replication of new VOCs, such as Delta and Omicron variants, at the site of the infection would not allow sufficient time for reactivation of immune memory and early enough to prevent local mucosal infection, replication and viral transmission.

In the present invention, the inventors provided numerous evidences that LV:: Sseta-2P i.n. boost can be used to induce robust systemic and mucosal adaptive immunity, to broaden the specificity of the protective response. The LV::Sseta-2P i.n. boost strengthen the intensity, broaden the VOC cross-recognition, and targets B-and T-cell immune responses to the principal entry point of SARS-CoV-2 to the mucosal respiratory of the host organism and avoid the infection of main anatomical sites. A phase l/lla clinical trial is currently in preparation for the use of i.n. boost by LV::Sseta-2P in previously vaccinated persons or in COVID- convalescents.

The inventors established that B-cell independent and antigen-specific T-cell immunity played a major role in LV-mediated protection against SARS-CoV-2 infection (Ku MW, et al. EMBO Mol Med, e14459, 2021). This was consistent with the strong T-cell responses induced by LV::SBeta-2P at the systemic level and in the lungs, detectable in vaccinated mice ( Vesin, B., et al. Mol Ther 30, 2984-2997, 2022). Importantly, all or the vast majority of murine and human T-cell epitopes, identified on the ancestral Scov-2 sequence and included in SBeta-2P, were preserved in the mutated Spike from the emerging variants, including Omicron BA.1 and BA.4/5 sub-variants (Figure 11).

These observations indicated the strong capability of LV expressing the Spike protein of SARS-CoV-2 especially LV::Sseta-2P at inducing full protection against not only the ancestral but also against the recently emerging SARS-CoV-2 variants by eliciting robust T-cell responses.

In contrast to the B-cell epitopes which were targets of neutralizing antibodies, the so far identified T-cell epitopes had not been impacted or were barely impacted by mutations accumulated in the Scov-2 of the emerging variants (Figure 11).

This observation indicates why the LV platform, which induces strong T-cell immunity, has an invariable and complete protective capacity against the newly emerging variants.

Materials and Method

Mice immunization and SARS-CoV-2 infection

Female C57BL/6JRj mice were purchased from Janvier (Le Genest Saint Isle, France), housed in individually-ventilated cages under specific pathogen-free conditions at the Institut Pasteur animal facilities and used at the age of 7 wks. Mice were immunized i.m. with 1 pg/mouse of mRNA-1273 (Moderna) vaccine. For i.n. injections with LV, mice were anesthetized by i.p. injection of Ketamine (Imalgene, 80255mg/kg) and Xylazine (Rompun, 5 mg/kg). For protection experiments against SARS-CoV-2, mice were transferred into filtered cages in isolator. Four days before SARS-CoV-2 inoculation, mice were pretreated with 3 x 10 ⁸ IGU of Ad5::hACE2 as previously described (Ku MW, et al. Cell Host Microbe, 29(2), 236- 249 e236, 2021). Mice were then transferred into a level 3 biosafety cabinet and inoculated i.n. with 0.3 x 10 ⁵ TCID50 of the Delta SARS-CoV-2 clinical isolate (Lescure FX, et al. Lancet Infect Dis, 20(6), 697-706, 2020) contained in 20 pl. Mice were then housed in filtered cages in an isolator in BioSafety Level 3 animal facilities. The organs recovered from the infected animals were manipulated according to the approved standard procedures of these facilities.

Ethical approval of animal experimentation.

Experimentation on animals was performed in accordance with the European and French guidelines (Directive 86/609/CEE and Decree 87-848 of 19 October 1987) subsequent to approval by the Institut Pasteur Safety, Animal Care and Use Committee, protocol agreement delivered by local ethical committee (CETEA #DAP20007, CETEA #DAP200058, and Ministry of High Education and Research APAFIS#24627- 2020031 117362508 v1 , APAFIS#28755-20201221 10238379 v1 ).

Construction and production of vaccinal LVs

First, a codon-optimized sequence of Spike from the Ancestral, D614G, Alpha, Beta or Gamma VOCs were synthetized and inserted into the pMK-RQ S-2019-nCoVJS501 YV2 plasmid. The S sequence was then extracted by BamHI/Xhol digestion to be ligated into the pFlap lentiviral plasmid between the BamHI and Xhol restriction sites, located between the native human ieCMV promoter and the mutated atg starting codon of Woodchuck Posttranscriptional Regulatory Element (WPRE) sequence (See the plasmid maps, Figure 10). To introduce the K ⁹⁸⁶ P-V ⁹⁸⁷ P “2P” double mutation in SDGUG or Seeta, a directed mutagenesis was performed by use of Takara In-Fusion kit on the corresponding pFlap plasmids. Various pFlap-ieCMV-S-WPREm or pFlap-ieCMV-Sap-WPREm plasmids were amplified and used to produce non-integrative vaccinal LV, as described elsewhere, (Ku MW, et al. Cell Host Microbe, 29(2), 236-249 e236, 2021; Ku MW, Charneau P, Majlessi L. Expert Rev Vaccines, 1-16, 2021). Analysis of humoral and systemic T-cell immunity

Anti-Scov-21 g G and IgA antibody titers were determined by ELISA by use of recombinant stabilized Scov-2 or RBD fragment for coating. Neutralization potential of clarified and decomplemented sera or lung homogenates was quantitated by use of lentiviral particles pseudo-typed with Scov-2 from diverse variants, as previously described (Ku MW, et al. Cell Host Microbe, 29(2), 236-249 e236, 2021; Sterlin D, et al. Sci Trans! Med, 13(577), 2021).

T-splenocyte responses were quantitated by IFN-y ELISPOT after in vitro stimulation with S:256-275, S:536-550 or S:576-590 synthetic 15-mer peptides which contain Scov-2 MHC- l-restricted epitopes in H-2 ^d mice (Ku MW, et al. Cell Host Microbe, 29(2), 236-249 e236, 2021). Spots were quantified in a CTL Immunospot S6 ultimate-V Analyser by use of CTL Immunocapture 7.0.8.1 program.

Phenotypic and Functional cytometric analysis of lung immune cells

Enrichment and staining of lung immune cells were performed as detailed elsewhere (Ku MW, etal. Cell Host Microbe, 29(2), 236-249 e236, 2021; Ku MW, etal. EMBO Mol Med, e14459, 2021) after treatment with 400 U/ml type IV collagenase and DNase I (Roche) for a 30- minute incubation at 37°C and homogenization by use of GentleMacs (Miltenyi Biotech). Cell suspensions were then filtered through 100pm-pore filters, centrifuged at 1200 rpm and enriched on Ficoll gradient after 20 min centrifugation at 3000 rpm at RT, without brakes. The recovered cells were co-cultured with syngeneic bone-marrow derived dendritic cells loaded with a pool of A, B, C peptides, each at 1 pg/ml or negative control peptide at x290pg/ml. The following mixture was used to detect lung Tc1 cells: PerCP-Cy5.5-anti-CD3 (45-0031 - 82, eBioScience), eF450-anti-CD4 (48-0042-82, eBioScience) and APC-anti-CD8 (17-0081 -82, eBioScience) for surface staining and BV650-anti- IFN-g (563854, BD), FITC-anti-TNF (554418, BD) and PE-anti-IL- 2 (561061 , BD) for intracellular staining. The following mixture was used to detect lung Tc2 cells: PerCP-Cy5.5-anti-CD3 (45-0031 -82, eBioScience), eF450-anti-CD4 (48-0042-82, eBioScience), BV71 1 -anti-CD8 (563046, BD Biosciences), for surface staining and BV605-anti-IL-4 (504125, BioLegend Europe BV), APC-anti-IL-5 (504306, BioLegend Europe BV), FITC-anti-IL-10 (505006, BioLegend Europe BV), PE-anti- IL-13 (12-7133-81 , eBioScience) for intracellular staining. The intracellular staining was performed by use of the Fix Perm kit (BD), following the manufacturer's protocol. Dead cells were excluded by use of Near IR Live/Dead (Invitrogen). Staining was performed in the presence of Fcyll/lll receptor blocking anti-CD16/CD32 (BD).

To identify lung resident memory CD8 ⁺ T-cell subsets, a mixture of PerCP-Vio700-anti- CD3 (130-119-656, Miltenyi Biotec), PECy7-CD4 (552775, BD Biosciences), BV510-anti-CD8 (100752, BioLegend), PE-anti-CD62L (553151 , BD Biosciences), APC-anti-CD69 (560689, BD Biosciences), APC-Cy7-anti-CD44 (560568, BD Biosciences), FITC-anti-CD103 (1 1 - 1031 -82, eBiosciences) and yellow Live/Dead (Invitrogen) was used. Lung B cells were studied by surface staining with a mixture of PerCP Vio700-anti-lgM (130-106-012, Miltenyi), and PerCP Vio700-anti-lgD (130-103-797, Miltenyi), APC-H7-anti-CD19 (560143, BD Biosciences), PE-anti-CD38 (102708, BioLegend Europe BV), PE-Cy7-anti-CD62L (ab25569, AbCam), BV711 -anti-CD69 (740664, BD Biosciences), BV421 -anti-CD73 (127217, BioLegend Europe BV), FITC-anti-CD80 (104705, BioLegend Europe BV) and yellow Live/Dead (Invitrogen).

Cells were incubated with appropriate mixtures for 25 minutes at 4°C, washed in PBS containing 3% FCS and fixed with Paraformaldehyde 4% after an overnight incubation at 4°C. Samples were acquired in an Attune NxT cytometer (Invitrogen) and data analyzed by FlowJo software (Treestar, OR, USA).

Determination of viral RNA content in the organs

Organs from mice were removed and immediately frozen at -80°C on dry ice. RNA from circulating SARS-CoV-2 was prepared from lungs as described elsewhere (Ku MW, et al. Cell Host Microbe, 29(2), 236-249 e236, 2021). Lung homogenates were prepared by thawing and homogenizing in lysing matrix M (MP Biomedical) with 500 pl of PBS using a MP Biomedical Fastprep 24 Tissue Homogenizer. RNA was extracted from the supernatants of organ homogenates centrifuged during 10 min at 2000g, using the Qiagen Rneasy kit, except that the neutralization step with AVL buffer/carrier RNA was omitted. The RNA samples were then used to determine viral RNA content by E-specific qRT-PCR. To determine viral RNA content by Esg-specific qRT-PCR, total RNA was prepared using lysing matrix D (MP Biomedical) containing 1 mL of TRIzol reagent (ThermoFisher) and homogenization at 30 s at 6.0 m/s twice using MP Biomedical Fastprep 24 Tissue Homogenizer. The quality of RNA samples was assessed by use of a Bioanalyzer 2100 (Agilent Technologies). Viral RNA contents were quantitated using a NanoDrop Spectrophotometer (Thermo Scientific NanoDrop). The RNA Integrity Number (RIN) was 7.5-10.0. SARS-CoV-2 E or E sub-genomic mRNA were quantitated following reverse transcription and real-time quantitative TaqMan® PCR, using SuperScriptTM III Platinum One-Step qRT- PCR System (Invitrogen) and specific primers and probe (Eurofins), as recently described (Ku MW, et al. EMBO Mol Med, e14459, 2021).

Examples

Antigen design and down-selection of a lead candidate To select the most suitable Scov-2 variant, which can consider the dynamics of virus spread of the known variants and able to induce the greatest neutralization breadth, the inventors generated LVs encoding the full length Scov-2 from the Alpha, Beta or Gamma SARS- CoV-2 VOCs. C57BL/6 mice (n = 5/group) were primed i.m. (wk 0) and boosted i.m. (wk 3) with 1 x 10 ⁸ TU/mouse of each individual LV and the (cross) neutralization potential of their sera was assessed before boost (wk 3) and after boost (wk 5) against pseudoviruses carrying various Scov-2 (Figure 1A). Immunization with LV: :SAi _P ha generated appropriate neutralization capacity against SDGUG and SAipha but not against Sseta and Scamma (Figure 1 B). Between LV::Sseta and LV::SGamma, the former generated the highest cross sero-neutralization potential against SDGUG, S^pha and Scamma variants. In accordance with previous observations using other vaccination strategies in the context of immunization with LV, the K ⁹⁸⁶ P - V ⁹⁸⁷ P substitutions in the S2 domain of SCoV-2 improved the (cross) sero-neutralization potential (Figure 1 C), which the most probably results from an extended half-life of Sc _o v-2-2P ( Walls AC, et al. Cell, 181(2), 281-292 e286, 2020).

Taken together these data allowed to down select Sseta-2p as the best cross-reactive antigen candidate to be used in the context of LV (LV::Sseta-2p) to strengthen the waning immunity previously induced by the first generation COVID-19 vaccines, like mRNA-1273.

Follow-up of humoral immunity in mRNA-1273-primed and -boosted mice and effect of LV::SBeta-2p i.n. boost

The inventors analyzed the potential of LV::Sseta-2P i.n. boost vaccination to strengthen and broaden the immune responses in mice which have been initially primed and boosted with mRNA-1273 and in which the (cross) sero-neutralization potential is decreasing. C57BL/6 mice were primed i.m. at wk 0 and boosted i.m. at wk 3 with 1 pg/mouse of mRNA-1273, defined as the optimal dose of this vaccine in mice (Nature, 2020, Vol. 586, 567- 571) (Figure 2A). Longitudinal serological follow-up demonstrated that at 3 wks post prime, cross-neutralization activities against both SDGUG and SAipha were readily detectable (Figure 2B). Cross sero-neutralization was also detectable, although to a lesser degree, against Scamma, but not against Sseta, Soeita or SDeita+. At wk 6, i.e., 3 wks post boost, cross sero- neutralization activities against all Scov-2 variants were detectable, although at significantly lesser extents against Sseta, Soeita and SDeita+. From wk 6 to wk 10, cross sero-neutralization against Sseta, Soeita, or Soeita+ gradually and significantly decreased.

At wk 15, groups of mRNA-1273-primed and -boosted mice received i.n. escalating doses of 1 x 10 ⁶ ,1 x 10 ⁷ , 1 x 10 ⁸ , or 1 x 10 ⁹ TU/mouse of LV: :Sseta-2P (Figure 2A). Control mRNA-1273-primed and -boosted mice received i.n. 1 x 10 ⁹ TU of an empty LV (LV Ctrl). In parallel, at this time point, mRNA-1273-primed and -boosted mice were injected i.m. with 1 pg of mRNA-1273 or PBS. Unprimed, age-matched mice received i.n. 1 x 10 ⁹ TU of LV::Sseta-2P or PBS.

In mRNA-1273-primed mice, serum anti-Scov-2 and anti-RBD IgG were detected at wk 3, increased after mRNA-1273 boost as studied at wk 6 and 10, and then decreased at wk 17 in the absence of additional boost (Figure 8A). In the mRNA-1273-primed and -boosted mice, boosted at wk 15, a significant increase was observed in the titer of anti-Scov-2 IgG in the mice injected with 1 x 10 ⁸ or 1 x 10 ⁹ TU of LV::Sseta-2P or a third dose of mRNA-1273 (Figure 2C). The titers of anti-Scov-2 IgA were higher in the mice injected with 1 x 10 ⁹ TU of LV::Sseta-2pthan those injected with a third dose of mRNA-1273 (Figure 2C). At the mucosal level, at this time point, titers of anti-Scov-2 and anti-RBD IgG in the total lung extracts increased in a dosedependent manner in LV::SBeta-2P-injected mice, with the highest dose of LV::Sseta-2P was comparable to the third i.m. dose of mRNA-1273 (Figure 8B). Importantly, significant titers of lung anti-Scov-2 IgA were only detected in LV::SBeta-2P-boosted mice but were barely detectable in their counterparts boosted late via i.m. with mRNA-1273 (Figure 8B).

At the mucosal cellular level, inside the lung IgM /IgD CD19 ⁺ (Ig switched) B-cell population, increased percentages of CD38 ⁺ CD62L ⁺ CD69 ⁺ CD73 ⁺ CD80 ⁺ cells, which may constitute resident memory B cells, were detected in a dose-dependent manner in mice boosted i.n. with LV::Sseta-2P, but not in their mRNA-1273 i.m. -boosted counterparts (Figure 3A, B).

Features of systemic and mucosal T-cell immunity after i.n. l_V::SBeta-2P boost in previously mRNA-1273-primed and -boosted mice

Systemic anti-Scov-2 T-cell immunity was assessed by I FN-y-specific ELISPOT in the spleen of individual mice immunized following the above-mentioned regimen (Figure 2A) after in vitro stimulation with individual S:256-275, S:536-550 or S:576-590 peptide, encompassing immunodominant Scov-2 regions for CD8 ⁺ T cells in H-2 ^b mice (Ku MW, et al. Cell Host Microbe, 29(2), 236-249 e236, 2021). Importantly, the weak anti-S CD8 ⁺ T-cell immunity, detectable in the spleen of mRNA-1273 primed-boosted mice at wk 17, largely increased following i.n. boosting with 1 x 10 ⁸ and 1 x 10 ⁹ TU/mouse of LV::Sseta-2P (Figure 4). The booster effect of the 1 x 10 ⁹ TU dose of LV::Sseta-2P i.n. was comparable, or had a tendency to be higher than, the additional dose of mRNA-1273 i.m. on the systemic T-cell immunity.

In parallel, in the same animals, the mucosal anti-Scov-2 T-cell immunity was assessed by intracellular Tc1 and Tc2 cytokine staining in T-cell enriched fraction from individual mice, after in vitro stimulation with autologous bone-marrow-dendritic cells loaded with a pool of S:256-275, S:536-550 and S:576-590 peptides (Figure 5A). In previously mRNA-1273-primed and -boosted mice, only a few Scov-2-specific IFN-y/TNF/IL-2 CD8 ⁺ T-cell responses were detected in the lungs (Figure 5A). The i.n. administration of LV::Sseta-2P boosted, in a dose dependent manner, these Tc1 responses with sizable percentages of such Tc1 cells induced with the doses of 1 x 10 ⁸ or 1 x 10 ⁹ TU. The mRNA-1273 i.m. administration had a substantially lower boost effect on mucosal T cells (Figure 5A, B). Tc2 responses (IL-4, IL-5, IL-10 and IL- 13) were not detected in any experimental group (Figure 9).

Importantly, the proportions of lung resident memory CD8 ⁺ T cells (Trm) were significantly increased in mice boosted i.n. with escalating doses of LV::Sseta-2P, in net contrast to their counterparts which received i.m. a third dose of mRNA-1273. In fact, the latter did not have significant proportions of this Trm cell population which is very frequently correlated with protective potential in numerous infectious diseases (Figure 6A, B).

Full protective potential of a late LV::SBeta-2P i.n. boost in mRNA-1273-primed and -boosted mice

The inventors then evaluated the protective vaccine efficacy of LV::Sseta-2P i.n. boost in mRNA-1273-primed and -boosted mice, following a vaccination regimen comparable to the above-mentioned one. At wk 15, mRNA-1273-primed and -boosted mice received i.n. the suboptimal dose of 1 x 10 ⁸ TU of LV::Sseta-2p or control empty LV (Figure 7A). The choice of such suboptimal dose was based on numerous previous observations from the inventors with this dose which was effective in protection in a homologous LV prime-boost experiment (Ku MW, et al. Cell Host Microbe, 29(2), 236-249 e236, 2021; Ku MW, et al. EMBO Mol Med, e14459, 2021). Control mRNA-1273-immunized mice received mRNA-1273 i.m. or PBS. Unvaccinated, age-and sex-matched controls were left unimmunized. Four weeks after the late boost, i.e. wk 20, all mice were pre-treated with 3 x 10 ⁸ Infectious Genome Units (IGU) of an adenoviral vector serotype 5 encoding hACE2164(Ad5::hACE2) (Ku MW, et al. Cell Host Microbe, 29(2), 236-249 e236, 2021) to render their lungs permissive to SARS-CoV-2 replication (Figure 7B). Four days later, mice were challenged with SARS-CoV-2 Delta variant, which, at the time of the present invention, i.e., November 2021 , was the most expanded SARS- CoV-2 variant worldwide.

At day 3 post infection (dpi), analysis of the total lung RNA first showed that hACE-2 was expressed homogeneously in all mice after Ad5::hACE2 in vivo transduction (Figure 7B). Lung viral loads were then determined at 3 dpi by assessing total E RNA and sub-genomic (Esg) ECOV-2 RNA qRT-PCR, the latter being an indicator of active viral replication (Chandrashekar A, et al. Science, 369(6505), 812-817, 2020; Tostanoski LH, et al. Nat Med, 26(11), 1694- 1700, 2020; Wolfel R, Corman VM, Guggemos W et al. Virological assessment of hospitalized patients with COVID-2019. Nature, 581(7809), 465-469, 2020). In mice initially primed and boosted with mRNA-1273, and then injected i.n. with the control LV or i.m. with PBS alone, no significant protection potential was detectable against the challenge with SARS-CoV-2 Delta variant. In net contrast, the LV::Sseta-2P i.n. boost drastically reduced the total E _Co v-2 RNA content of SARS-CoV-2 and no copies of the replication-related Esg E _Co v-2 RNA were detected in this group (Figure 7B). The content of total E _Co v- ₂ RNA was also significantly reduced in the group which received a late mRNA-1273 i.m. boost. The content of Esg Ecov-2 RNA was undetected in 3 out of 5 in this group.

• Full Lung Prophylaxis against SARS-CoV-2 by One Shot or Booster Intranasal Lentiviral Vaccination in Syrian Golden Hamsters

The inventors demonstrated that a single intranasal administration of a vaccinal lentiviral vector encoding a stabilized form of the original SARS-CoV-2 Spike glycoprotein induced full lung protection of respiratory tracts and strongly reduced pulmonary inflammation in the susceptible Syrian golden hamster model against the prototype SARS-CoV-2. In addition, the inventors showed that a lentiviral vector encoding stabilized Spike of SARS-CoV- 2 Beta variant (LV::Sseta-2p) prevented pathology and reduced infectious viral loads in lungs and nasal turbinates following inoculation with the SARS-CoV-2 Omicron variant. Importantly, an intranasal boost with LV::Sseta-2P improved cross-seroneutralization much better in LV::Sseta- ₂ p-primed hamsters than in their counterparts primed with an LV encoding Spike from the ancestral SARS-CoV-2. These results strongly suggested that an immune imprint with the original Spike sequence had a negative impact on cross-protection against new variants. The results of the inventors tackled the issue of vaccine effectiveness in people who have already been vaccinated and have vanished immunity and indicated the efficiency of LV-based intranasal vaccination, either as a single dose or as booster.

Construction and production of LV expressing S protein

Construction of LV::SWAI and LV::SWAI-AF2P were described previously (Ku MW, et al. Cell Host Microbe, 29(2), 236-249 e236, 2021; Ku MW, et al. EMBO Mol Med, e14459, 2021).

Production and titration of LV

Lentiviral particles were produced by transient calcium phosphate co-transfection of HEK293T cells with the vector plasmids pFlap/Sc _ov-2 , a vesicular stomatitis virus G Indiana envelope plasmid and an encapsidation plasmid pD64V for the production of integrationdeficient vectors. Supernatants were harvested at 48 h post-transfection, clarified by 6-min centrifugation at 2500 rpm at 4°C. LV were aliquoted and stored at -80°C. Vector titers were determined by transducing HEK293T cells treated with aphidicolin. The titer, proportional to the efficacy of nuclear gene transfer, was determined as Transduction Unit (TU)/mL by qPCR on total lysates at day 3 post-transduction, by use forward and reverse primers specific to pFLAP plasmid, and forward and reverse primers specific to the host housekeeping gadph gene, as previously described (Iglesias et a!., J.Gene Med., 2006, 8, 265-274).

SDS-PAGE and western blotting

Six-well plates were seeded with HEK293T cells (2 10 ⁶ cells/well), and after overnight growth transduced with LV encoding SARS-CoV-2 S transgenes at a multiplicity of infection of 10. Cell lysates were harvested 48 h post transduction and quantified. After heating for 5 min at 95°C bolt with Bolt sample buffer, samples were loaded on a precast Bolt 4-12% Bis-Tris gel (Invitrogen). Proteins were transferred to a nitrocellulose membrane using an iBlot2 dry blotting system (Invitrogen), and membrane was blocked with TBST blocker (Tris-buffered saline (TBS) containing 0.2% Tween 20 and 5% milk). Following 1 h blocking, the membrane was incubated overnight with an anti- SARS-CoV-2 S2 rabbit polyclonal antibody (SinoBiological 40590-T62) in TBST blocker. The membrane was then washed three times with TBST for 10 min and subsequently incubated for 1 h with 1 :2,500 Dy Light 800-conjugated goat anti-rabbit IgG (H+L) secondary antibody (Invitrogen, Cat # SA5-35571 ) in TBST Blocker. Finally, the membrane was washed three times with TBST for 10 min and developed using an ODYSSEY CLx Infrared Imaging System (Li-COR). E-PAGE SeeBlue Pre-stained Standard (Invitrogen) was used as ladder.

Hamsters

Male Mesocricetus auratus Syrian golden hamsters (Le Genest Saint Isle, France) were purchased mature and weighed between 80 to 100 gr at the beginning of the experiments. Hamsters were housed in individually ventilated cages under specific pathogen- free conditions during the immunization period. For SARS-CoV-2 infection these hamsters were transferred into individually filtered cages placed inside isolators in the animal facility of Institut Pasteur. Prior to i.m. or i.n. injections, hamsters were sedated with isoflurane inhalation or i.p. injection of Ketamine (Imalgene, 100 mg/kg) and Xylazine (Rompun, 5 mg/kg).

Ethical Approval of Animal Studies

Experimentation on hamsters was realized in accordance with the European and French guidelines (Directive 86/609/CEE and Decree 87-848 of 19 October 1987) subsequent to approval by the Institut Pasteur Safety, Animal Care and Use Committee, protocol agreement delivered by local ethical committee (CETEA #DAP200007) and Ministry of High Education and Research (APAFIS#24627-2020031 117362508 v1 ).

Production of SARS-CoV-2 Spike proteins

Codon-optimized nucleotide fragments encoding a stabilized version of SARS-CoV-2 WA1 or Omicron BA.1 spike (HexaPro) ectodomain (followed by a foldon trimerization motif) and WA1 or Omicron BA.1 RBD proteins containing C-terminal tags (Hisx8-tag, Strep-tag, and AviTag) were synthesized and cloned into pcDNA3.1/Zeo ⁽⁺⁾ expression vector (Thermo Fisher Scientific). Recombinant proteins were produced by transient transfection of exponentially growing Freestyle 293-F suspension cells (Thermo Fisher Scientific, Waltham, MA) using polyethylenimine (PEI) precipitation method as previously described (PMID: 25910833). Proteins were purified from culture supernatants by high-performance chromatography using the Ni Sepharose® Excel Resin according to manufacturer’s instructions (GE Healthcare), dialyzed against PBS using Slide-A-Lyzer® dialysis cassettes (Thermo Fisher Scientific), quantified using NanoDrop 2000 instrument (Thermo Fisher Scientific), and controlled for purity by SDS-PAGE using NuPAGE 3-8% Tris-acetate gels (Life Technologies), as previously described (PMID: 25910833).

Humoral response

Immunoglobulin G (IgG) Abs were detected by an enzyme-linked immunosorbent assay (ELISA) by use of recombinant stabilized Scov-2 and RBD proteins from SARS-CoV-2 WA1 or Omicron strains. Nunc Polysorp ELISA plates (ThermoFisher, 475094) were coated at 1 pg/mL in 50mM Na2CO3 pH 9.6 at 4°C overnight. After incubation, plates were washed with 1X PBS + 0.05% Tween-20 (PBST) and blocked with PBST + 1% BSA for 2 to 3 h at 37°C. Plates were incubated with serial dilutions of sera in PBS-T + 1 % BSA for 1 .5 hr at 37°C. Following washes, rabbit anti-hamster IgG-horseradish peroxidase conjugate (Jackson Immuno Research, M37470) was used as secondary Ab, and 3,5,3'5'-tetramethylbenzidine (Eurobio Scientific, 5120-0047) was used as the substrate to detect Ab responses. Reactions were stopped with 50 pL of 2 M sulfuric acid. Endpoint titers were calculated as the highest serum dilution that resulted in an absorbance value greater than that mean +3SD of pre- immune sera.

SARS-CoV-2 inoculation

Hamsters were anesthetized by i.p. injection of Ketamine and Xylazine mixture, transferred into a biosafety cabinet 3 and inoculated i.n. with 50 pl of viral inoculum containing 0.3 x 10 ⁵ TCID50 of the WA1 (Lescure et al., Lancet Infect. Dis., 2020, 20, 697-706} or the Omicron BA.1 variant (Pango lineage BA.1 , GISAID: EPI ISL 6794907 and EPI ISL 7413964) of SARS-CoV-2 clinical isolate (Planas eta!., Nature, 2022, 602, 671-675). Animals were housed in an isolator in BioSafety Level 3 animal facilities of Institut Pasteur. The organs recovered from the infected animals were manipulated according to the approved standard procedures of these facilities.

Pseudovirus neutralization assay

Nab quantification was assessed via an inhibition assay which uses HEK293T cells stably expressing human ACE2 (HEK 293T-ACE2) and non-replicative Scov-2 pseudo-typed LV particles which harbor the reporter luciferase firefly gene, allowing quantitation of the host cell invasion by mimicking fusion step of native SARS-CoV-2 virus, as previously described (Sterlin et al., Sci. Transl. Med., 2021, 13, eabd2223). Serum samples or clarified lung homogenates were heat inactivated at 56°C for 30 minutes. Serial four-fold dilutions of samples diluted in 25 pl DMEM-glutamax (Gibco, 21063-029) containing 10% heat-inactivated FCS, 100 U/mL penicillin and 100 mg/mL streptomycin and 1 mM sodium pyruvate (Gibco, 1 1360-070) were mixed with 1 ng of Scov-2 pseudo-typed LV p24 equivalent in 25 pl for 30 min at room temperature, in U-bottom plates. Samples were then transferred into clear-flat-bottom 96-well-black-plates (Corning, CLS3603) containing 2 10 ⁴ HEK 293T-ACE2 cells. The plates were incubated for 72 hours at 37°C and then assayed for luciferase expression using the ONE-Glo™ Luciferase Assay System (Promega, E6120) on an EnSpire plate reader (PerkinElmer). EC50 was reported as the reciprocal of the serum dilution conferring 50% of infection of HEK 293T-ACE2 cells by lentiviral vectors bearing the indicated Scov-2 variants.

Determination of viral loads in the organs

Lungs and nasal turbinates (NT) were removed aseptically and immediately frozen at - 80°C. RNA from circulating SARS-CoV-2 was prepared from lungs as recently described (Ku MW, et al. Cell Host Microbe, 29(2), 236-249 e236, 2021). Briefly, lung homogenates were prepared by thawing and homogenizing of the organs using lysing matrix A (MP Biomedicals, 116913050-CF) in 500 pl of ice-cold PBS in an MP Biomedical Fastprep 24 Tissue Homogenizer and were used to determine viral loads by E-specific qRT-PCR. Alternatively, total RNA was prepared from lungs or NT by addition of lysing matrix D (MP Biomedical, 1 16910050-CF) containing 1 mL of TRIzol reagent (ThermoFisher, 15596026) and homogenization at 30 s at 6.0 m/s twice using MP Biomedical Fastprep 24 Tissue Homogenizer. These RNA preparations were used to determine viral loads by Esg-specific qRT-PCR or inflammatory mediators. SARS-CoV-2 E gene or E sub-genomic mRNA (Esg RNA), was quantitated following reverse transcription and real-time quantitative TaqMan® PCR, using SuperScript™ III Platinum™ One-Step qRT-PCR Kit (Invitrogen, 11732020) and specific primers and probe (Eurofins) as previously described (Corman et al. Euro Surveill. 2020, 25(3); Wolfel et al., Nature 2020, 581 (7809) :465-9). The standard curve of Esg mRNA assay was performed using in vitro transcribed RNA derived from PCR fragment of “T7 SARS-CoV-2 Esg mRNA”. The in vitro transcribed RNA was synthesized using T7 RiboMAX Express Large Scale RNA production system (Promega, P1320) and purified by phenol/chloroform extraction and two successive precipitations with isopropanol and ethanol. Concentration of RNA was determined by optical density measurement, diluted to 10 ⁹ genome equivalents/pL in RNAse-free water containing 100pg/mL tRNA carrier, and stored at -80°C. Serial dilutions of this in vitro transcribed RNA were prepared in RNAse-free water containing 10pg/ml tRNA carrier to build a standard curve for each assay. PCR conditions were: (i) reverse transcription at 55°C for 10 min, (ii) enzyme inactivation at 95°C for 3 min, and (iii) 45 cycles of denaturation/amplification at 95°C for 15 s, 58°C for 30 s. PCR products were analyzed on an ABI 7500 Fast real-time PCR system (Applied Biosystems). RNA copy values were extrapolated from the standard curve and multiplied by the volume to obtain RNA copies per organ. The limit of detection was based on the standard curve and defined as the quantity of RNA that would give a Ct value of 40.

The qRT-PCR quantification of inflammatory mediators in the lungs and brain of hamsters was performed in total RNA extracted by TRIzol reagent, as recently detailed (Ku MW, et al. Cell Host Microbe, 29(2), 236-249 e236, 2021).

Histopathology

Samples from the lungs of hamsters were fixed in formalin for 7 days and embedded in paraffin. Paraffin sections (5-pm thick) were stained with Hematoxylin and Eosin (H&E). Histopathological lesions were qualitatively described and when possible scored, using: (i) distribution qualifiers (i.e., focal, multifocal, locally extensive or diffuse), and (ii) a five-scale severity grade, i.e., 1 : minimal, 2: mild, 3: moderate, 4: marked and 5: severe. In some cases, serial sections were prepared for immunohistochemistry (IHC) analyses. IHC was performed as previously described (Ku MW, et al. EMBO Mol Med, e14459, 2021). Rabbit anti-N _C ov- ₂ antibody (Novus Biologicals, NB100-56576) and biotinylated goat anti-rabbit Ig secondary antibody (Dako, E0432) were used in IHC. Slides were scanned using the AxioScan Z1 (Zeiss) system and images were analyzed with the Zen 2.6 software. Statistical Analysis

Statistical significance was assigned when P values were < 0.05. ELISA titers were log transformed prior to statistical analysis. For comparison of two groups, nonparametric Mann-Whitney test was used. To compare more than 2 experimental groups, Kruskal-Wallis ANOVA with Dunn’s multiple comparisons tests were applied. Differences in neutralizing activity of VoCs were analyzed by two-way ANOVA with Sidak test for multiple comparisons. Tests were performed using GraphPad Prism software (Version 9, Graphpad Software, La Jolla, CA, United States).

Results

Immunogenicity of LV encoding various Scov-2 forms

Non-integrative LV encoding stabilized conformers of Scov-2 under transcriptional control of the cytomegalovirus (CMV) immediate early promoter (Pcwvie) were constructed (Figure 12A). The first two Scov-2 conformers were derived from a human codon-optimized full- length membrane anchored ancestral WA1 Scov-2 {Ku MW, et al. Cell Host Microbe, 29(2), 236-249 e236, 2021). LV: :SWAI-2P encodes a SWAI which harbors two stabilizing K ⁹⁸⁶ P and V ⁹⁸⁷ P substitutions in the hinge loop of the S2 domain. LV: :SWAIAF-2P encodes a SWAI which, in addition to the two K ⁹⁸⁶ P and V ⁹⁸⁷ P substitutions, is deleted of the loop encompassing the S1/S2 furin cleavage site (675-QTQTNSPRRAR-685 of SEQ ID NO: 27) for further stability at the prefusion state McCallum et al., Nat. Struct. Mol. Bio!., 2020, 27, 942-949; Launay et al., EBioMedicine 2022, 75, 103810). Sseta-2P is from the Beta (B.1 .351 ) VoC and contains the two K ⁹⁸⁶ p _anc | v ⁹⁸⁷ P substitutions. Sseta differs from SWAI , notably by the N ⁵⁰¹ Y/K ⁴¹⁷ N/E ⁴⁸⁴ K mutations located in the RBD {Tegally et al., Nature 2021, 592, 438-443). Whereas pseudoviruses carrying SWAI were neutralized by sera from individuals vaccinated with the currently approved vaccines, those presenting these RBD mutations moderately-to-strongly resist neutralization {Kuzmina et al., iScience 2021, 24, 103467). This observation provided a rational for adapting the S sequence variant for further vaccination. Expression of Scov-2 immunogens in HEK293T cells transduced with the four LV was confirmed by Western blot on total cell lysates (Figure 12B). As expected, the S2 furin cleavage product was only detected in the cells transduced by LV encoding SWAI , SwAi-2p or Sseta-2P which harbor an intact furin cleavage site.

To compare the immunogenicity of LV::SWAI , LV: :SWAI-2P, LV: :SWAIAF-2P and Sseta-2P, hamsters {n = 4/group) were immunized by a single i.m. injection of 1 x 10 ⁸ TU of either LV. Five weeks (wks) later, high serum titers of anti-SwAi IgG antibodies were induced by all LV studied (Figure 12C). As no significant difference in immunogenicity between these LV was observed, LV: :SWAIAF-2P, hereafter referred to as “LV::S”, was selected for evaluation of protection against homologous SARS-CoV-2.

Induction of robust humoral responses against SARS-CoV-2 by a single i.n. LV::S administration

The inventors recently showed that LV::S used in a prime (i.m.) - boost (i.n.) protocol significantly improved protection against SARS-CoV-2 compared to a single i.m. injection in the hamster model (Ku MW, et al. Cell Host Microbe, 29(2), 236-249 e236, 2021). Here, the inventors evaluated the protective potential of a single i.n. administration of LV::S against the ancestral WA1 SARS-CoV-2. Hamsters (n = 6/group) were immunized i.n. with a single injection of 1 x 10 ⁸ TU of LV::S at wk 0 or at wk 5 (Figure 13A). As a positive control, a group of hamsters was primed i.m. with 1 x 10 ⁸ TU of LV::S at wk 0 and then boosted i.n. at wk 5 with the same amount of LV::S. Control hamsters received, following the same regimen, equivalent amounts of an LV expressing a green fluorescent protein, as an irrelevant antigen (LV Ctrl). At wk 7, all animals were challenged i.n. with 0.3 x 10 ⁵ TCID50 of WA1 SARS-CoV-2 (Figure 13A). Before the challenge, pre-immune sera and those from the LV Ctrl group were tested negative for anti-SwAi and -RBDWAI antibodies (Figure 13B). Following a single LV::S i.n. injection, all animals mounted high titers of anti-SwAi and -RBDWAI IgG. These antibody titers are obtained as soon as 2 wks post-immunization as shown by hamsters vaccinated at wk 5. The serum IgG titers remained stable until wk 7. At wk 7, significantly lower anti-SwAi and anti-RBDwAi IgG titers were detected in the groups injected i.n., compared to the i.m. -i.n. group. The sero-neutralization activity was evaluated by use of pseudoviruses harboring SWAI . In agreement with the anti-RBD IgG titers, sero-neutralizing activities were lower in the hamsters immunized with a single i.n. injection, compared to the i.m. -i.n. group (Figure 13C). Despite comparable anti-S and anti-RBD IgG titers at 2 or 7 wks after i.n. injection, sera from hamsters vaccinated at the earlier time point exhibited a slightly higher neutralizing capacity, suggesting the requirement of an antibody maturation over time to reach efficient neutralizing potential. However, all vaccinated groups had equivalent neutralizing capacities in their total lung homogenates, four days post SARS-CoV-2 inoculation (4 dpi) (Figure 13C). The virus neutralizing activity in lungs can be a more relevant correlate of protection than that detectable in sera.

Protection against homologous SARS-CoV-2 challenge induced by a single i.n. LV::S administration In the lungs of LV::S-vaccinated individuals of either i.m.-i.n. or single i.n. groups, ~ 2- to-4 log decreases of viral contents were observed compared to the LV Ctrl group, as determined by qRT-PCR detecting the SARS-CoV-2 Envelop (E) RNA at 4 dpi (Figure 14A, left panel). Lung viral content measured by a sub-genomic E RNA (Esg) qRT-PCR is an indicator of active viral replication ( Wolfel et a!., Nature 2020, 581 (7809) :465-9). This analysis showed a complete absence of replicating virus in the three vaccinated groups versus a geometric mean ± SD of (5.4 ± 6.8) x 10 ⁸ copies of Esg RNA of SARS-CoV-2/lungs in the LV Ctrl group (Figure 14A, right panel). At 4 dpi, in accordance with the protection observed, only 2-3% weight loss was detected in the hamsters vaccinated, either by i.n. alone or by i.m.-i.n. prime-boost regimen, compared to 12% weight loss in the hamsters which received LV Ctrl (Figure 14B). As evaluated by qRT-PCR at 4 dpi in the total lung homogenates of the LV::S- vaccinated and SARS-CoV-2 challenged hamsters, marked decreases were detected in the expression of inflammatory IFN-y, TGF-a, IL-6 cytokines, anti-inflammatory IL-10 cytokines, and CCL2, CCL3, CCL5 and CXCL10 chemokines, as well as FoxP3, in comparison to their LV ctrl-injected and challenged counterparts (Figure 14C). Changes in inflammatory markers was particularly noticeable for the group receiving the i.n. administration 2 weeks before the challenge. In agreement with these results, a positive correlation was found between viral loads and inflammation (r = 0.46, p < 0.05), whereas the weights were inversely correlated with viral loads and inflammation (r = -0.6842, p < 0.001 and r= -0.56, p < 0.01 , Spearman test), respectively.

Reduced infection-driven inflammation in hamsters vaccinated with a single i.n. LV::S administration

On lung histopathological examination, vaccinated controls demonstrated lung infiltration (Figure 15A) and severe alveolo-interstitial inflammation (Figure 15B) leading to dense pre-consolidation areas (Figure 15C). These lungs also displayed bronchiolar lesions, with images of epithelial sloughing of individual or clustered cells (Figure 15D) and of hyperplastic epithelial growth producing papillary projections (Figure 15E) or intraluminal epithelial folds (Figure 15F). In vaccinated groups the interstitial (Figure 15A) and alveolar (Figure 15G) lesions were minimal to moderate. Immunohistochemistry analysis of the lungs of LV ctrl-treated and infected hamsters, with a SARS-CoV-2 nucleocapsid protein (Ncov-2)- specific polyclonal antibody, detected numerous clusters of N _Co v-2 ⁺ cells in the bronchial epithelial cells (not shown) and in the interstitium (Figure 15H, right panels). In contrast, the severity of inflammation was reduced in LV::S-vaccinated animals. When detectable, the inflamed zones still contained N _Co v- ₂ ⁺ cells, indicating that, although viral replication has been controlled (Figure 14A), infiltration and virus remnants have not yet been fully resorbed at the early 4 dpi time point.

These data collectively indicated that immunization with a single i.n. administration of LV::S was as efficient as an i.m. prime followed by a i.n. boost regimen and conferred strong protective immunity against an homologous SARS-CoV-2 infection.

LV::SB&ta-2P prime (i.m.) - boost (i.n.) vaccination cross-protects against Omicron variant Given the dynamics of the pandemic, an important question is the ability of vaccines to induce cross-protection against new VoCs. Based on a series of LVs encoding for S from various VoCs, the inventors recently selected LV::Sseta-2p as the best candidate to generate the broadest spectrum of cross-neutralization potential ( Vesin, B., et al. Mol Ther 30, 2984-2997, 2022). To evaluate the efficacy of LV::Sseta-2P in the hamster model against SARS-CoV-2 Omicron, hamsters (n = 4-5/group) were primed i.m. or i.n. at wkO with 1 x 10 ⁸ TU of LV::Sseta- 2p. At wk3, one group of each were boosted i.n. with the same dose of LV::Sseta-2P (Figure 16A). All groups were challenged at wk7 with 0.3 x 10 ⁵ TCID50 of SARS-CoV-2 BA.1 Omicron sub-variant (Planas et aL, Nature 2022, 602(7898):671 :5). Somicron harbors 32 mutations compared to SWAI . Among these mutations, 15 are located in the RBD. Infection of hamsters with this BA.1 Omicron strain, isolated from a patient, resulted in a significant decrease in weight (Figure 16B).

Robust cross-reactive serum IgG titers were detected by ELISA against Somicron and RBDomicron in all LV::SBeta-2P-vaccinated hamsters (Figure 16B). No significant differences in antibody titers between the groups were observed 3wks post-prime. Antibody levels remained stable after the single injections, while a significant increase in anti-Somicron titers was observed in the animals primed and boosted i.n.. By contrast, anti-RBD antibody titers continued to rise in all vaccinated groups over time (Figure 16B, lower panel).

Following challenge, hamsters vaccinated by a single i.m. injection of LV::Sseta-2P or those who received LV Ctrl gradually lost weight (Figure 16C). Hamsters vaccinated i.n. with LV::Sseta-2P exhibited less than 5% of weight loss, without signs of morbidity (Figure 16D). At 4 dpi, viral contents in the lungs and in nasal turbinates analyzed. High viral contents were detected in both organs of the LV ctrl-injected group (Figures 16E, 16F). In contrast, no Esg RNA was detected from the lungs of the i.m. - i.n. group and significant reductions of ~2 log were observed in the other groups (Figure 16E). Of note, the i.m. vaccinated hamster which did not control viral replication had the highest weight loss. Although also significantly reduced, active viral replication was still detectable in the NT of all hamsters, indicating that LV-based i.n. vaccination, despite its strong efficacy in the protection of the lungs, does not fully prevent nasal infection (Figure 17F). However, an i.n. boost, regardless of the route of prime led to a better efficacy over a single vaccine administration in the control and the spread of infection in the respiratory tract tissues.

Decrease of virus content as determined by immunohistochemistry in LV:: SBeta-2P- vaccinated hamsters.

At 4 dpi, histopathological analysis of the lung sections in the Ctrl group showed similar lesions detailed in the Figure 15H (Figure 17). Immunohistochemistry images displayed a generally less abundant N _Co v-2 staining in mice boosted i.n or i.m, relative to the primed-only and LV ctrl-injected animals, although there was a relatively high degree of intra-group variation (Figure 17). In addition, the inventors did not observe a tight correlation between the extent of the IHC signal and the Esg qRT-PCR quantifications, indicating that part of the immunostained antigen corresponds to non-replicating virus remnants.

Altogether, these results showed that while a single i.n. immunization with LV can be enough to control the infection, an LV-based i.n. administration will be better adapted to boost a previously induced anti-COVID-19 immunity.

Induction of cross-reactive antibodies in LV::SwAi-2P-primed and LV::SB&ta-2P-boosted hamsters

The inventors then evaluated the efficacy of an LV::Sseta-2P i.n. boost in animals previously exposed to SWAI . Hamsters (n = 4/group) were primed i.m. at wkO with 1 x 10 ⁸ TU of LV::SwAi-2por LV::Sseta-2P. At wk5, both groups were boosted i.n. with 1 x 10 ⁸ TU of LV::Sseta- 2P (Figure 18A). Robust serum IgG titers were detected against S and RBD proteins at any post-prime time point tested, in all vaccinated hamsters (Figure 18B). After the prime or after the boost, comparable kinetic profiles and intensities of SWAI- or Somicron specific antibody responses were observed (Figure 18B, upper panels). Either the homologous or the heterologous i.n. boost marginally increased the anti-S antibody titers by 1.8- or 2.5-fold, respectively. By contrast, ~4- to 10-fold lower serum IgG responses against RBDomicron were measured compared to RBDWAI (Figure 18B, lower panels). However, anti-RBDomicron IgG titers were significantly better improved by heterologous boost than by homologous boost with a 3.8- versus 1 .7-fold increase, respectively.

Anti-Scov-2 antibody imprinting in LV::S-primed and LV::SBeta-2P-boosted hamsters

Five weeks post i.m. injection, both LV: :SWAI-2P and LV::Sseta-2P induced high sero- neutralizing activities against pseudoviruses harboring SDGUG or SAipha (Figure 19A). Cross- neutralizing activity against SDGUG, SAipha, and Soeita was similar in the two groups of immunized hamsters. Of note, after a single i.m. injection, only LV::SBeta-2P-immunized hamsters exhibited sero-neutralization activity against all S variants, although weaker against Somicron.

LV::Sseta-2P i.n. boost increased the cross sero-neutralization potential against all VoCs in both groups (Figure 19B). Although the levels of neutralizing antibodies were improved in the sera from the LV::SwAi-2P-primed and LV::Sseta-2P-boosted hamsters, they were barely able to cross-neutralize pseudoviruses harboring Sseta and totally unable to cross-neutralize pseudoviruses harboring Somicron (Figure 19B). Lung homogenates exhibited a similar profile with no cross-neutralizing activities against Sseta or Somicron following the heterologous primeboost (Figure 19C). In net contrast, sera and lung homogenates from LV::SBeta-2P-primed and LV::SBeta-2P-boosted hamsters were much better able to cross-neutralize pseudoviruses harboring Sseta or Soamma and - to a lesser extent - those harboring Somicron. However, this primeboost regimen was sufficient to confer protection against SARS-CoV-2 ₀ micron challenge, as observed above (Figure 16). Therefore, an LV::Sseta-2P boost improves cross sero- neutralization much better in LV::SBeta-2P-primed hamsters than in their LV::SwAi-primed counterparts. These results showed a clear imprinting effect in anti-S humoral immunity, and particularly obvious against the Omicron and Beta variants.

Discussion

LV-based platform has emerged recently as a powerful vaccination approach against COVID-19. The inventors notably demonstrated its strong prophylactic capacity at inducing protection in the lungs against SARS-CoV-2 infection when used as a systemic prime followed by mucosal i.n. boost (Ku MW, et al. Cell Host Microbe, 29(2), 236-249 e236, 2021). In the present study, as a further step toward a clinical trial, the inventors used LV encoding stabilized forms of SWAI or Sseta- This choice was based on data indicating that stabilization of viral envelop glycoproteins in their prefusion forms improves the yield of their production as recombinant proteins in industrial manufacturing of subunit vaccines. Moreover, it also increases the efficacy of nucleic acid-based vaccines, by raising availability of the antigen under its optimal immunogenic shape (Hsieh et ai., Science, 2020, 369(6510):1501 -5).

In the first part of this report, the inventors demonstrated that a single i.n. administration of an LV encoding SWAI confers, as efficiently as an i.m. -i.n. prime-boost regimen, full protection of lungs in the highly susceptible hamster model, as evaluated by virological, immunological and histopathological parameters. The hamster ACE2 ortholog protein interacted efficaciously with Scov-2, readily allowing host cell invasion by SARS-CoV-2 with high replication rates. With rapid weight loss and development of severe lung pathology subsequent to SARS-CoV-2 inoculation, outbred hamsters provided a sensitive model to evaluate the efficacy of drug or vaccine candidates (S/a et al., Nature 2020, 583(7818) :834-8). Hamsters represented a more challenging model than Rhesus macaques which developed only a mild COVID-19 pathology. The strong protection of the lung conferred by a single i.n. administration against homologous challenge in the hamster model was therefore an asset of primary importance. This protection most likely resulted from the development of a mucosal immunity. Induction of antigen-specific secretory dimeric IgA that could block the interaction of the virus at the mucosal level had been shown to reduce the viral shedding and to correlate with protection (Halfmann et a!., Cell Rep. 2022; 39(3):110688; Munoz-Fontela et al., Nature 2020, 586(7830) :509-15; Wang et al. Sci Transl Med. 2021, 13(577)). Although infectious virus was still detected in the nasal turbinates of i.n. -immunized hamsters, the significant reduction in infectious viral titers could lead to reduced transmission and dissemination as recently described by Langel SN et al. providing a means of disease control (Langel etal., Sci Transl. Med. 2022, 14(658)). Indeed, the inventors have previously shown in the mouse model that anti-S IgG and secretory IgA antibodies were generated in lungs, together with lung resident memory B and T cells following i.n. LV administration. The presence of IgA induced by LV-based SARS-CoV-2 vaccines correlated with complete pulmonary protection against the virus ( Vesin, B., etal. Mol Ther30, 2984-2997, 2022). Unfortunately, the mucosal immunity could not be assessed in the hamster model because of the lack of immunological tools, including anti-lgA antibodies and antibodies to activation/memory T-cell markers. Meanwhile, there was growing evidence that i.n. immunization provided a better protection, not only against SARS-CoV-2 ancestral strain, but also against newly emerged VoCs (Afkhami et al., Cell 2022, 185(5):896-915; Bricker et al., Cell Rep. 2021, 36(3): 109400). The studies exploring this domain so far used chimpanzee adenoviral vectored vaccines that are known to be pro-inflammatory, and thus risky for use in mucosal vaccination (Coughlan et al., Mol. Ther. 2022, Adenovirus-based vaccines-a platform for pandemic preparedness against emerging viral pathogens). In net contrast, LVs are non- cytopathic and very weakly inflammatory (Ku et al. Vaccines 2021:1-16, 1988854) and much more suitable for mucosal vaccination. The fact that a single i.n. LV-based vaccine administration, either 2 or 7 wks before homologous SARS-CoV-2 challenge, elicits protection is valuable in setting clinical trials with LV-based vaccines. This platform can provide remarkable advantages for mass vaccination, with the major advantage of mucosal immunization in the reduction of viral transmission.

The continued emergence of SARS-CoV-2 VoCs prompted the inventors to expand their study by assessing the protective potential of a heterologous antigen booster which could, in terms of anti-S antibody response, mimic some aspects of a previous infection or earlier vaccination with the first-generation vaccines, mainly based on SWAI . Numerous breakthrough SARS-CoV-2 infections have been observed in vaccinated individuals, showing the incomplete cross-efficacy of these vaccines (Abu-Raddad LJ, et al. N Engl J Med. 2021 ;385(2):187-9; Kuhlmann C, et al. Lancet. 2022 ;399(10325) :625-6). Recently, it has been reported that mucosal booster vaccination is needed to establish robust sterilizing immunity in the respiratory tract against SARS-CoV-2 (Tang J, et al. Respiratory mucosal immunity against SARS-CoV-2 following mRNA vaccination. Sci Immunol. 2022:eadd4853). In LV-immunized hamsters, the inventors did not detect striking differences between the ability of SWAI and Sseta- 2P antigens to induce cross-reactive serum IgG responses against Scov-2. However, a clear distinction should be made between the protective capacity of vaccines and their ability to induce neutralizing antibodies, since T-cell responses are also major effector players against SARS-CoV-2 infection. In particular, the effectiveness of LV-based protection is not only dependent on the capacity to induce neutralizing antibody responses but also, and to a large extent, on their T-cell immunogenicity. It is noteworthy that an almost complete protection of lungs is achieved in pMT KO mice that are totally devoid of mature B-cell compartment and antibody response (Ku MW, et al. EMBO Mol Med. 2021 :e14459). In addition, mucosal resident memory T cells, as well as IFNy ⁺ IL-2 ⁺ TNF ⁺ triple positive CD8 ⁺ T cell effectors, are readily detectable in the lung of LV::S-primed (i.m.) and boosted (i.n.) mice [26]. Furthermore, findings obtained following natural infection largely suggest that specific T-cell immunity, which is generally less affected by mutations occurring in the S antigen of emerging SARS-CoV-2 variants, are largely effective against viral replication (Altmann DM, et al. Cell Rep Med. 2021 ;2(5):100286 ; Mazzoni A, et al. Front Immunol. 2022;13:801431). T-cell mediated protection is also certainly operating in hamsters. However, as mentioned above, the lack of immunological tools prevented the characterization of T-cell responses in the present study.

In the LV::S-primed and LV::SBeta-2P-boosted hamsters, despite the enhanced sero- neutralizing potential against D614G, Alpha and Delta variants, largely statistically reduced cross-neutralization activities were observed against Beta and Gamma variants and no crossneutralization activity was observed against the Omicron variant. Likewise, in the cases of influenza A viruses, a first exposure to a serotype can affect future responses to its variants (Gostic KM, et al. Science. 2016;354(6313):722-6). This raises concerns about immune imprinting effects of previous infections or vaccinations on antibody responses, which will need to be considered when designing vaccines (Roltgen K, et al. Immune imprinting, breadth of variant recognition, and germinal center response in human SARS-CoV-2 infection and vaccination. Cell. 2022). The present study indicates that pre-exposure of the immune system to an early S variant has a negative impact on the neutralizing antibody response, measured after a late booster with a heterologous S variant. The results of the inventors corroborate recent data showing that healthcare workers infected either by SARS-CoV-2 ancestral or Alpha variant exhibit a reduced neutralizing immunity against Omicron (Reynolds CJ, et al. Science. 2022;377(6603):eabq1841). Moreover, using mRNA vaccines, Kalnin et al. also showed that heterologous boosting provided inferior neutralizing antibody titers compared to homologous boosting (Kalnin KV, et al. Vaccine. 2022;40(9):1289-98). The hypothesis can be put forward that additional injections of the variant S sequence could be required to counteract this negative effect and to reach sufficient levels of cross-neutralization against VoCs.

Collectively, the results of the inventors demonstrated the ability of the LV as an effective vaccine delivery platform. LV was an effective and promising strategy to elicit a strong protective immunity against SARS-CoV-2 VoCs and possessed the advantage to be noninflammatory and thus suitable for use in mucosal i.n. vaccination. The inventors have recently demonstrated the safety of LV::Sseta-2P i.n. administration in mice in which the high dose of 1 x 10 ⁹ TU of LV had been injected ( Vesin, B., et al. Mol Ther 30, 2984-2997, 2022). No adverse effects had been detected by lung histopathological analyses.

• Full cross-protective capacity of LV::SBeta-2P against SARS-CoV-2 Omicron variant

The inventors evaluated the protective efficacy of LV::Sseta-2P in B6.K18-hACE2 ^{IP THV} transgenic mice, which were prone to SARS-CoV-2 infection in the lung, and in addition displayed unprecedented brain permissiveness to SARS-CoV-2 replication (Ku MW, et al. EMBO Mol Med, e14459, 2021). B6.K18-hACE2 ^{IP THV} mice (n = 5/group) were primed intramuscularly (i.m.) with 1 x 10 ⁸ TU/mouse of LV::Sseta-2p or an empty LV (sham) at wk 0 and then boosted intranasally (i.n.) at wk 3 with the same dose of the same vectors (Figure 20A). Mice were then challenged (i.n.) with 0.3 x 10 ⁵ TCID50 of SARS-CoV-2 Omicron variant at wk 5.

Lung and brain viral RNA contents were then determined at day 5 post infection by using a sub-genomic E _Co v-2 RNA (Esg) qRT-PCR, which is an indicator of active viral replication (Chandrashekar et al., 2020, Science, 369, 812-817; Tostanoski et al., 2020, Nat. Med. 26, 1694-1700; Wolfel et al., 2020, Nature, 581, 465-469). LV::Sseta-2P vaccination conferred sterilizing protection against SARS-CoV-2 Omicron in the lungs, i.e., undetectable Esg viral RNA in the vaccinated mice versus (5.83 ± 6.22) x 10 ⁹ copies of viral RNA/lungs in their sham- vaccinated counterparts (Figure 20B, left). In parallel, Esg qRT-PCR quantification of viral RNA contents in brain detected no copies of viral RNA in the vaccinated mice versus (6.41 ± 1.29) x 10 ⁹ copies in the brain of the sham-vaccinated controls (Figure 20B right). Of note, even if 2 mice out of 5 did not show cervical infection with Omicron, the three others showed very high amounts of viral replication in the brain.

Therefore the LV:Sseta-2P displayed a full cross-protective capacity against the Omicron variant, which was fully comparable to its efficiency against the ancestral or the Delta variant that we previously demonstrated (Ku MW, et al. EMBO Mol Med, e14459, 2021 ; Ku MW, et al. Cell Host Microbe, 29(2), 236-249 e236, 2021; Vesin, B., et al. Mol Ther 30, 2984-2997, 2022).