Different treatment approaches to infectious diseases : from novel antimicrobials to T-cell therapy

Part I:
The antibiotic tolerance of biofilms and the acquisition of bacterial resistance to virtually all antibiotics highlight the need to develop novel antimicrobials. While traditional drug design methods mainly target rapidly growing planktonic bacteria, anti-biofilm antimicrobials have to be selected based on their activity against biofilm-embedded sessile microorganisms.
This study reports on the activity of EN4 (8-hydroxyserrulat-14-en-19-oic acid), a compound derived from Eremophila (Myoporaceae) plant species, for its possible application in implant-associated infections (IAI). IAI are mainly caused by biofilm-forming staphylococci, which hinders treatment using traditional antibiotics. EN4 acts against a variety of Gram-positive bacteria and Mycobacterium tuberculosis, but not against Gram-negative microorganisms. Its efficacy is similar against methicillin-susceptible and -resistant Staphylococcus (S.) aureus (MSSA and MRSA, respectively). The minimal inhibitory concentrations (MICs) of EN4 for logarithmic-phase S. aureus and S. epidermidis are 25 µg/ml and 50 µg/ml and the minimal bactericidal concentrations (MBCs) 50 µg/ml and 100 µg/ml, respectively. EN4 shows rapid and concentration-dependent killing of staphylococci, reducing bacterial counts by > 3 log10 colony-forming units (CFU)/ml within 5 min at concentrations above 50 µg/ml. Additionally, EN4 is bactericidal against stationary-phase and adherent staphylococci independently of polysaccharide intercellular adhesin (PIA)-mediated biofilm. The antimicrobial activity of EN4 comprises a general inhibition of macromolecular biosynthesis and membranolytic properties. Consistently with the membrane-targeted activity, Pseudomonas aeruginosa turns susceptible to EN4 upon destabilisation of the outer membrane. Cytotoxicity of EN4 on mouse fibroblasts is time- and concentration-dependent in the range of MBCs suggesting a narrow therapeutic window. In vivo in a mouse tissue cage model EN4 showed neither bactericidal nor cytotoxic effect. This in vivo inhibition was caused by interaction with albumin. These findings suggest that upon structural changes EN4 might serve as a pharmacophore for the future development of new antimicrobials.
The problems associated with the therapy of IAI have prompted many studies on engineering implant surfaces that could resist microbial colonisation. Within this project the antiadhesive and antibacterial properties of surfaces coated with poly-2-methyl-2-oxazoline (PMOXA), poly(2-hydroxyethyl methacrylate) (PHEMA) and poly(poly(ethylene glycol) methacrylate (PEGMA) were evaluated. These polymer brushes were further covalently functionalised with vancomycin. Vancomycin, a cell wall active antibiotic, was used as it has previously been shown to be effective when covalently coupled. Employing various methods, a great nonfouling potential of the coatings against biofilm-forming S. epidermidis 1457 was shown. However, the antibacterial effect of covalently coupled vancomycin was not observed, even for Bacillus subtillis 6633, which is characterised by a higher susceptibility to vancomycin. The determination of surface concentration of the antibiotic revealed less than 0.02 ng vancomycin per 1 cm2, which is below the assessed minimal killing concentration range (500 ng to 1 000 ng/1 cm2). Moreover, assessment of in-solution activity of vancomycin-PMOXA construct pointed toward a decrease in the antimicrobial properties of the antibiotic, especially when the amine group was involved in binding. Altogether, the investigated PMOXA, PHEMA and PEGMA-based surface coatings exhibit great antiadhesive properties. However, covalent functionalisation with vancomycin did not confer antimicrobial properties to the polymers due to decreased activity and too low surface concentration of the antibiotic.
Part II:
Epstein-Barr virus (EBV)-associated tumours that include post-transplant lymphoproliferative disorders (PTLDs) belong to the most serious complications of immunosuppression occurring in approximately 5% of all transplant recipients and are still associated with an exuberant morbidity and mortality. In these patients, the immunosuppression reduces the number and/or the function of the EBV-specific T-cells leading to uncontrolled proliferation of EBV-infected B-cells and tumour formation. Thus, reconstitution of antigen-specific T-cells of the patients has the potency to provide immediate and also long-term protection against PTLDs. Several studies have demonstrated safety and clinical benefit of adoptive transfer of antigen-specific CD4+ and CD8+ T-cells. Nevertheless, controversial data exist on the necessity of CD4+ and CD8+ T-cell content of T-cell product to control PTLDs. Moreover, most T-cell generation procedures are very laborious and time-consuming, which precludes its wide application. A more rapid approach has recently been introduced into clinics, by which T-cells are directly isolated based on interferon (IFN)-γ production upon stimulation with virus-derived antigens. Since the T-cell response to EBV is directed against a set of viral proteins and strongly depends on the human leukocyte antigen (HLA) types, we aimed at identifying immunogenic antigens to generate a T-cell product consisting of both CD4+ and CD8+ cells covering a broad range of HLA types and at investigating their role in controlling PTLDs.
Thus, we included the commercially available overlapping peptide pools derived from EBV antigens, which are known to be expressed in PTLDs, i.e. EBV nuclear antigen (EBNA)1, EBNA3c, latent membrane protein (LMP)2a and the Bam HI Z restriction fragment of EBV, beginning with the Leftward OR Frame number 1 (BZLF1) 19-21. Moreover, to broaden the antigenic repertoire, we used a pool of 125 CD8+ and CD4+ T-cell epitopes with known HLA restrictions, consisting mainly of 9 and 15 amino acids derived from 15 EBV latent and lytic proteins (EBVmix). In the initial screening the EBVmix appeared to induce the highest T-cell response among a wide range of donors compared to EBNA1, EBNA3c, BZLF1 and LMP2a peptide pools. The isolation of specific T-cells from five healthy volunteers using the IFN-γ capture assay allowed to select the highest numbers of T-cells, characterised by the best purity of both CD4+IFN-γ+ and CD8+IFN-γ+ for EBVmix.
The expanded T-cell lines were highly specific to their cognate antigens. Moreover, upon encounter of the autologous EBV-infected lymphoblastoid cell lines (LCLs), the T-cells proliferated and produced IFN-γ, which demonstrated their ability to recognise and respond to endogenously processed viral antigens. This effect was especially pronounced for the T-cells specific to the EBVmix and LMP2a. We next proved the safety of these T-cell lines in a mixed lymphocyte reaction, in which they appeared to be devoid of alloreactive potential against third-party dendritic cells (DCs).
With regard to the functionality, all lines exhibited moderate cytotoxic properties of 0%-60% against LCLs, which is in line with previous reports. Due to this relatively low cytotoxicity we questioned their in vivo efficacy against PTLDs. Therefore, we sought to assess their long-term control over LCLs using an outgrowth assay that revealed the highest efficacy of the EBVmix-specific T-cell lines to inhibit the outgrowth of LCLs. While the T-cells specific to LMP2a, EBNA3c and EBNA1 required higher effector to target ratios to control LCLs, the BZLF1 lines predominantly failed in the outgrowth assay.
The proliferation, the concomitant secretion of IFN-γ in response to autologous LCLs as well as the control of the outgrowth and lack of alloreactivity were promising findings in view of the in vivo potential of these T-cells. Moreover, we could show in epitope mapping that the T-cell lines specific to the EBVmix had an oligoclonal epitope signature, which minimises the risk of induction of tumour mutants not susceptible to T-cell therapy.
As the role of EBV-specific CD4+ T-cells in adoptive transfer is controversial, CD4+ cells were purified from the T-cell lines. The generated CD4 lines showed high specificities after re-stimulation with the cognate antigens, but secreted less IFN-γ and exhibited mostly lower cytotoxicity than the original T-cell lines upon encounter of autologous LCLs. Most interestingly and in contrary to some previous reports from clinical studies the CD4 lines, especially in case of EBNA3c and EBNA1, exhibited a profoundly decreased controlling capacity of LCL in comparison to the T-cell lines. In contrast, purified CD8 lines, consisting of CD8+ T-cells specific to the peptide pools, inhibited the outgrowth similarly to their T-cell lines counterparts, indicating that EBV-specific CD8+, but not CD4+, T-cell have a crucial function against LCLs in vitro and that the CD4+ T-cell helper function might be dispensable in our in vitro system.
We could further show that the observed reduced capacity of EBV-specific CD4 lines against LCLs was not caused by their decreased viability, exhaustion, or immune evasion by LCLs. Importantly, when co-cultured with LCLs, CD4+ T-cells produced less cytotoxic effector molecules (perforin, granzyme B and granulysin) over time than the EBV-specific CD8+ T-cells, which at least in part explains their inability to eradicate the LCLs.
Altogether, we could show that the EBV-specific T-cell lines generated using for stimulation the EBVmix, comprised of 125 CD8+ and CD4+ T-cell epitopes derived from 16 EBV latent and lytic proteins, demonstrated the highest efficacy of controlling the autologous LCL. None of the commercially available EBV-derived peptide pools exhibited characteristics similar to the EBVmix. These findings indicated a clear advantage of combining defined T-cell epitopes derived from different EBV proteins over using single viral antigens for the generation of EBV-specific T-cells for adoptive transfer. In support of this notion, during preparation of our study Miltenyi Biotec has introduced a new product optimised for efficient stimulation EBV-specific T-cells. The PepTivator EBV Consensus also contains T-cell epitopes derived from various lytic and latent EBV proteins and 29 (67%) of them are contained within the EBVmix. Our preliminary results however indicated that the EBV Consensus induces lower T-cell responses among healthy individuals.