Yingxia Wen

Human cytomegalovirus gH/gL/UL128/UL130/UL131A complex elicits potently neutralizing antibodies in mice

Human cytomegalovirus (HCMV) is a member of the β-herpesvirus family that causes significant disease worldwide. Although evidence exists that neutralizing antibodies and cytotoxic T cell responses to HCMV antigens can prevent HCMV disease and/or infection, there are no approved vaccines to prevent HCMV disease. Over the past 10 years, multiple HCMV vaccines have been tested in man but only partial protection has been achieved in these studies. HCMV contains multiple surface-expressed glycoproteins that are critical to viral entry, including gB, the gM/gN complex, the gH/gL complex, and a pentameric gH/gL/UL128/UL130/UL131A complex. Recently we showed that viral replicon particles (VRPs) expressing the gH/gL complex elicited more potently neutralizing antibodies than VRPs expressing gB in mice. Here we compare the immunogenicity of VRPs encoding the HCMV gH/gL and pentameric complexes, as well as purified gH/gL and pentameric complexes administered in the presence or absence of the MF59 adjuvant. The results of these studies indicate that the pentameric complex elicits significantly higher levels of neutralizing antibodies than the gH/gL complex, and that MF59 significantly increases the potency of each complex. In addition, we show that animals immunized with pentamer encoding VRPs or the pentameric subunit produce antibodies that recognize a broad range of antigenic sites on the complex. Taken together, these studies support the utility of the pentameric complex in HCMV vaccine candidates.

Expression of the human cytomegalovirus pentamer complex for vaccine use in a CHO system

Human cytomegalovirus (HCMV) causes significant disease worldwide. Multiple HCMV vaccines have been tested in man but only partial protection has been achieved. The HCMV gH/gL/UL128/UL130/UL131A complex (Pentamer) is the main target of neutralizing antibodies in HCMV seropositive individuals and raises high titers of neutralizing antibodies in small animals and non‐human primates (NHP). Thus, Pentamer is a promising candidate for an effective HCMV vaccine. Development of a Pentamer‐based subunit vaccine requires expression of high amounts of a functional and stable complex. We describe here the development of a mammalian expression system for large scale Pentamer production. Several approaches comprising three different CHO‐originated cell lines and multiple vector as well as selection strategies were tested. Stable cell pools expressed the HCMV Pentamer at a titer of approximately 60 mg/L at laboratory scale. A FACS‐based single cell sorting approach allowed selection of a highly expressing clone producing Pentamer at the level of approximately 400 mg/L in a laboratory scale fed‐batch culture. Expression in a 50 L bioreactor led to the production of HCMV Pentamer at comparable titers indicating the feasibility of further scale‐up for manufacturing at commercial scale. The CHO‐produced HCMV Pentamer bound to a panel of human neutralizing antibodies and raised potently neutralizing immune response in mice. Thus, we have generated an expression system for the large scale production of functional HCMV Pentamer at high titers suitable for future subunit vaccine production. Biotechnol. Bioeng. 2015;112: 2505–2515.

Conformationally selective biophysical assay for influenza vaccine potency determination.

Influenza vaccines are the primary intervention for reducing the substantial health burden from pandemic and seasonal influenza. Hemagglutinin (HA) is the most important influenza vaccine antigen. Subunit and split influenza vaccines are formulated, released for clinical use, and tested for stability based on an in vitro potency assay, single-radial immunodiffusion (SRID), which selectively detects HA that is immunologically active (capable of eliciting neutralizing or hemagglutination inhibiting antibodies in an immunized subject). The time consuming generation of strain-specific sheep antisera and calibrated antigen standards for SRID can delay vaccine release. The limitation in generating SRID reagents was evident during the early days of the 2009 pandemic, prompting efforts to develop more practical, alternative, quantitative assays for immunologically active HA. Here we demonstrate that, under native conditions, trypsin selectively digests HA produced from egg or mammalian cell in monovalent vaccines that is altered by stress conditions such as reduced pH, elevated temperature, or deamidation, leaving native, pre-fusion HA, intact. Subsequent reverse-phase high pressure liquid chromatography (RP-HPLC) can separate trypsin-resistant HA from the digested HA. Integration of the resulting RP-HPLC peak yields HA quantities that match well the values obtained by SRID. Therefore, trypsin digestion, to pre-select immunologically active HA, followed by quantification by RP-HPLC is a promising alternative in vitro potency assay for influenza vaccines.

Improving influenza virus backbones by including terminal regions of MDCK-adapted strains on hemagglutinin and neuraminidase gene segments

Reverse genetics approaches can simplify and accelerate the process of vaccine manufacturing by combining the desired genome segments encoding the surface glycoproteins from influenza strains with genome segments (backbone segments) encoding internal and non-structural proteins from high-growth strains. We have developed three optimized high-growth backbones for use in producing vaccine seed viruses for group A influenza strains. Here we show that we can further enhance the productivity of our three optimized backbones by using chimeric hemagglutinin (HA) and neuraminidase (NA) genome segments containing terminal regions (non-coding regions (NCRs) and coding regions for the signal peptide (SP), transmembrane domain (TMD), and cytoplasmic tail (CT)) from two MDCK-adapted high growth strains (PR8x and Hes) and the sequences encoding the ectodomains of the A/Brisbane/10/2010 (H1N1) HA and NA proteins. Viruses in which both the HA and NA genome segments had the high-growth terminal regions produced higher HA yields than viruses that contained one WT and one chimeric HA or NA genome segment. Studies on our best-performing backbone indicated that the increases in HA yield were also reflected in an increase in HA content in partially purified preparations. Our results show that the use of chimeric HA and NA segments with high-growth backbones is a viable strategy that could improve influenza vaccine manufacturing. Possible mechanisms for the enhancement of HA yield are discussed.

Trypsin pre-treatment corrects SRID over-estimation of immunologically active, pre-fusion HA caused by mixed immunoprecipitin rings

Influenza vaccines are the primary intervention to prevent the substantial health burden of seasonal and pandemic influenza. Subunit and split influenza vaccines are formulated, released for clinical use, and tested for stability based on their content of immunologically active (capable of eliciting functional antibodies) hemagglutinin (HA). Single-radial immunodiffusion (SRID), the standard in vitro potency assay in the field, is believed to specifically detect immunologically active HA. We confirmed that, with conformationally homogeneous HA preparations, SRID specifically detected native, pre-fusion HA, which elicited influenza neutralizing and hemagglutination inhibiting (HI) antibodies in mice, and it did not detect low-pH stressed, post-fusion HA, which was selectively removed from the SRID gel during a blotting step and was significantly less immunologically active. This selective detection was due to the SRID format, not a conformational specificity of the sheep antiserum used in the SRID, as the same antiserum detected non-stressed and low-pH stressed HA similarly when used in an ELISA format. However, when low-pH stressed HA was mixed with non-stressed HA, SRID detected both forms in mixed immunoprecipitin rings, leading to over-quantification of pre-fusion HA. We previously reported that trypsin digestion of antigen samples selectively degrade stressed HA, so that an otherwise conformationally insensitive biophysical quantification technique, reversed-phase high pressure liquid chromatography (RP-HPLC), can specifically quantify trypsin-resistant, immunologically active, pre-fusion HA. Here, we report that trypsin digestion can also improve the specificity of SRID so that it can quantify immunologically active, pre-fusion HA when it is mixed with less immunologically active, post-fusion HA.

Inactivated influenza vaccine stress can affect in vitro potency assay relationship to immunogenicity

Influenza vaccines are the most effective intervention to prevent the substantial public health burden of seasonal and pandemic influenza. The capability of hemagglutinin (HA), the main antigen in inactivated influenza vaccines (IIVs), to elicit functional neutralizing antibodies determines IIV effectiveness. When HA is subjected to environmental stress during manufacturing or while stored prior to administration, such as low pH and temperature excursions, the HA immunological activity can be affected. Single-radial immunodiffusion (SRID), the standard in vitro potency assay for IIVs, is believed to specifically detect immunologically active HA and has been applied to evaluate HA stability against stress. Here we report that transient low pH treatment and freeze/thaw cycles with HA in PBS abolish SRID-quantified in vitro potency for all HAs of multiple influenza strains. Raised temperature substantially decreases in vitro potency with more extensive HA structural changes. Chemical stress and mechanical stress moderately change SRID in vitro potency values in a strain-dependent manner. Trypsin digestion, which selectively degrades stressed HA, followed by RP-HPLC quantification as a candidate alternative in vitro potency assay yields results comparable to SRID. Mouse immunogenicity studies confirm that HA stressed by transient low pH treatment does not elicit functional antibodies in vivo, nor does it have a measureable SRID value. However, HA stressed by raised temperature elicits high titers of functional antibodies in vivo despite substantial loss of SRID in vitro potency. This discrepancy between SRID in vitro potency and vaccine immunogenicity suggests that SRID may not reliably indicate IIV potency under all conditions. Further efforts to develop alternate potency assays that can better predict in vivo immunogenicity should continue along with additional studies exploring HA conformation, SRID values and consequent immunogenicity.

Generation and characterization of a bivalent protein boost for future clinical trials: HIV-1 subtypes CR01_AE and B gp120 antigens with a potent adjuvant

The RV144 Phase III clinical trial with ALVAC-HIV prime and AIDSVAX B/E subtypes CRF01_AE (A244) and B (MN) gp120 boost vaccine regime in Thailand provided a foundation for the future development of improved vaccine strategies that may afford protection against the human immunodeficiency virus type 1 (HIV-1). Results from this trial showed that immune responses directed against specific regions V1V2 of the viral envelope (Env) glycoprotein gp120 of HIV-1, were inversely correlated to the risk of HIV-1 infection. Due to the low production of gp120 proteins in CHO cells (2–20 mg/L), cleavage sites in V1V2 loops (A244) and V3 loop (MN) causing heterogeneous antigen products, it was an urgent need to generate CHO cells harboring A244 gp120 with high production yields and an additional, homogenous and uncleaved subtype B gp120 protein to replace MN used in RV144 for the future clinical trials. Here we describe the generation of Chinese Hamster Ovary (CHO) cell lines stably expressing vaccine HIV-1 Env antigens for these purposes: one expressing an HIV-1 subtype CRF01_AE A244 Env gp120 protein (A244.AE) and one expressing an HIV-1 subtype B 6240 Env gp120 protein (6240.B) suitable for possible future manufacturing of Phase I clinical trial materials with cell culture expression levels of over 100 mg/L. The antigenic profiles of the molecules were elucidated by comprehensive approaches including analysis with a panel of well-characterized monoclonal antibodies recognizing critical epitopes using Biacore and ELISA, and glycosylation analysis by mass spectrometry, which confirmed previously identified glycosylation sites and revealed unknown sites of O-linked and N-linked glycosylations at non-consensus motifs. Overall, the vaccines given with MF59 adjuvant induced higher and more rapid antibody (Ab) responses as well as higher Ab avidity than groups given with aluminum hydroxide. Also, bivalent proteins (A244.AE and 6240.B) formulated with MF59 elicited distinct V2-specific Abs to the epitope previously shown to correlate with decreased risk of HIV-1 infection in the RV144 trial. All together, these results provide critical information allowing the consideration of these candidate gp120 proteins for future clinical evaluations in combination with a potent adjuvant.

Reference antigen-free and antibody-free LTD-IDMS assay for influenza H7N9 vaccine in vitro potency determination

Influenza vaccines are the most effective intervention to prevent the substantial public health burden of seasonal and pandemic influenza. Hemagglutinin (HA), as the main antigen in inactivated influenza vaccines (IIVs), elicits functional neutralizing antibodies and largely determines IIV effectiveness. HA potency has been evaluated by single-radial immunodiffusion (SRID), the standard in vitro potency assay for IIVs, to predict vaccine immunogenicity with a correlation to protective efficacy. We previously reported that limited trypsin digestion (LTD) selectively degraded stressed HA, so that an otherwise conformationally insensitive biophysical quantification technique could specifically quantify trypsin-resistant, immunologically active HA. Here, we demonstrate that isotope dilution mass spectrometry (IDMS), a method capable of quantifying the absolute HA concentration without reference antigen use, can be further expanded by adding LTD followed with precipitation to selectively quantify the active HA. We test the LTD-IDMS assay on H7N9 vaccines stressed by low pH, raised temperature, or freeze/thaw cycles. This method, unlike SRID, has no requirement for strain-specific reference antigens or antibodies and can generate potency values that correlate with SRID. Thus, LTD-IDMS is a promising alternative in vitro potency assay for influenza vaccines to complement and potentially replace SRID in a pandemic when strain specific reagents may not be readily available.

Influenza vaccine production technologies: past, present and future

Influenza is a constantly evolving global health threat that leads to substantial morbidity andmortality particularly in vulnerable populations at either end of the age spectrum. Society has responded by creating a global public-private system that involves constant surveillance, candidate virus generation, and release reagent generation linked toworldwide influenza vaccine manufacturing capabilities. It was initially recognised that influenza circulates as multiple antigenically distinct subtypes, which led to the generation of vaccines containing multiple influenza strains. The first and still current major process used for influenza vaccine production is infection of embryonated hen’s eggs with influenza virus. While this approach was a true advancement, some shortcomings such as lack of vaccine match to circulating strains due to egg adaptation and production capacity limitations have led to recent innovations inmammalian cell production and synthetic technologies aimed at further improving global influenza responses.