Utilizing Retroviral Particles as a Gene Delivery Platform for in vivo Imaging of Immune Cells upon Adoptive Cell Transfer and as a Model System for Developing New Means of Testing Viral Infections

Retroviruses (RVs) from the subfamily Orthoretrovirinae are enveloped RNA viruses with a diameter that ranges between 80 and 100 nm. They possess the ability of retrotranscribing RNA into DNA by the action of the reverse transcriptase. In addition, they can integrate the genetic cargo into the genome of mammalian cells in a semi-random fashion by the retroviral integrase. These characteristics enable the design of retroviral vectors as a genomic engineering tool for gene therapy purposes. Moreover, retroviruses resemble the morphology and size of retroviruses of other enveloped RNA viruses, such as SARS-CoV-2, which offers the possibility to be utilized as a model system to mimic the behavior of pathogenic enveloped RNA viruses.
Cancer is the sixth cause of death globally, and constitutes a challenge in the clinic. Conventional treatments such as tumor resection, chemotherapy, and radiotherapy can work in the short-term, but a remission of the tumor often occurs. Cancer immunotherapy has been in the spotlight in the recent years, due to the potential of being an improvement as a complementary therapy. Adoptive cell transfer (ACT) has gained traction as the most promising modality of cancer immunotherapy, as six chimeric antigen receptor (CAR)-T cell products have been FDA-approved as of July 2022. However, there is a lack of understanding on the kinetics and migration patterns of the infused cells upon the transfer, which would give a complete overview on the behavior of the cells to be able to further improve these therapies.
The outbreak of the coronavirus disease 19 (COVID-19) pandemic supposed a worldwide threat that compromised the public health systems. Conventional molecular testing tools such as qPCR and antigen rapid tests were a key component in the monitoring of the spread of the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) virus. However, these methods did not offer a direct quantification of the number and size viral particles present in the sample, since it was indirectly inferred from the threshold cycle (Ct) values in qPCR, and from the intensity of the band in rapid antigen tests. Moreover, the plaque assay methodology for assessing the number of infective particles was laborious and the results could take 24 hours to be received.
These challenges motivated the development of the two research projects that comprise the present PhD dissertation, which were divided in Part I, and Part II. The common link between both studies was the employment of retroviral particles for a specific purpose. Part I utilized retroviral particles for the genomic insertion of reporter genes in murine immune cells to enable long-term, non-invasive, and whole-body in vivo imaging of infused cells upon ACT. Part II used the morphological features of retroviruses to mimic the behavior of SARS-CoV-2 in the early phases of development of a new technique to achieve viral load quantification from saliva samples.
Part I was subdivided in Phase 1 and Phase 2. Phase 1 accomplished a successful optimization of production of lentiviruses pseudotyped with the vesicular stomatitis virus G (VSVG) envelope glycoprotein. In detail, a triprotein encoding the humanized Renilla luciferase (hRluc), the monomeric red fluorescent protein 1 (mRFP1), and the truncated thymidine kinase from the herpes simplex virus 1 (HSV1-tTK), was subcloned into different plasmids containing different promoters: cytomegalovirus (CMV), phosphoglycerokinase (PGK), elongation factor 1 alpha (EF1α), and the long-terminal repeat (LTR) of murine embryonic stem cell virus (MESV). Moreover, other reporter genes were designed linked to HSV-tTK1: red-red tandem heterodimer (RRvT), monomeric near-infrared fluorescent protein 680 (miRFP680), monomeric rhubarb 713 protein (mRhubarb713), small monomeric ultra red fluorescent protein (smURFP). The most efficient combination of transfer, packaging, and envelope plasmids was the molar ratio 3:2:1, respectively, with the transfection enhancer Fugene® HD. High-speed centrifugation and ultrafiltration were chosen as the most convenient concentration techniques, as a viral recovery of 91.92%, and 70.08%, respectively, could be obtained.
Phase 2 focused on the optimization of the retroviral transduction of both primary murine CD8+ T cells and lineage negative bone marrow cells (Lin- BM cells). The best modality for transduction was spinoculation with a combination of the transduction enhancer LentiBOOST with either polybrene or vectofusin, and leaving the virus in the medium for up to 24 hours. Moreover, spinoculation was required for an efficient lentiviral transduction with unconcentrated murine leukemia (MLV)-RVs, and concentration would render these particles inefficient. Last, hematopoietic reconstitution with engineered Lin- BM cells with the RRvT and HSV-tTK1 reporter genes was possible at the week 4 after lymphodepletion. The proportion of reconstituted mice was highest when the engineered Lin- BM cells were injected shortly after transduction and were not FACS sorted. The expression of the mentioned reporter genes was retrieved in CD8+ and CD4+ T cells, B cells and NK cells only at week 4. Nevertheless, the signal was lost in the following months. Moreover, the in vivo signal of the cells at week 4 before isolation was not strong enough to be detected in the U-CT scanner.
For the second project, retroviral particles were mixed with healthy saliva samples to develop a system to mimic the saliva from a person infected with SARS-CoV-2, since both are enveloped, ssRNA viruses. With this, the development of a novel diagnostic tool to quantify the size and number of viral particles from saliva, based on fluorescent z-stack imaging and a Python-based software was performed. The approach consisted on employing the dimeric fluorescent dye oxazole yellow 1 (YOYO-1) to bind double stranded motifs of RNA within the viral capsid, which is barely fluorescent in absence of binding but presents a bright fluorescence when bound. In this work, a protocol was obtained from the optimization of a series of steps that included buffer addition, mucus disruption, filtration, glycerol addition, and YOYO-1 staining. The resulting sample was taken to z-stack fluorescent imaging, and the collected data constituted the input of the Python-based software. The resulting output was the estimated size and number of particles contained in such sample. In parallel, the Python-based software was being updated as more data from the sample preparation was collected. To have a first view on the sensitivity of the methodology, several dilution series were performed with retroviral particles and heat-inactivated SARS-CoV-2. Thereafter, the protocol was transferred to Odense University Hospital (OUH) to perform further optimizations on SARSCoV-2, first from ex vivo cultures, and later from clinical samples from infected individuals. The best performing filters and buffers were, polyethersulfone (PES), and polyvinylidene difluoride (PVDF), and phosphate buffer saline (PBS) and 0.001% Pluronic F-68. The sensitivity of the current methodology was measured from the samples of 13 infected individuals, and a value of 76.9% was obtained. These data were compared with parallel data of qPCR and plaque assay. Unfortunately, the patients that were negative on the novel methodology, were positive in the qPCR and plaque assays.
Together, the two research projects showed the potential of utilizing a viral system for the benefit of the biomedical field: one as a gene delivery approach for reporter gene imaging in ACT, and another as a model to safely mimic similar viral pathogens for the development of new tools of testing.
Overall, this PhD thesis showed the optimal ways of producing retroviral particles and employing them successfully to achieve the integration of reporter genes for in vivo imaging purposes in ACT. Moreover, the potential of the hematopoietic reconstitution assay with engineered cells was highlighted, as different immune populations could be differentiated containing the same gene of interest. Nevertheless, this expression was not bright enough at 4 weeks after the transplantation, and the little signal was lost in the following months. Further work with imaging modalities other than fluorescence was needed, as well as with the other reporter genes that were also designed in this work.
In regards to Part II, a method for detection and quantification of particles in the size range of most viruses by z-stack fluorescent imaging was accomplished. As part of the process, a protocol for filtration of virus-containing saliva samples was successfully developed together with a Python-based software capable of quantifying the size and number of the contained particles. Samples from 13 SARS-CoV-2 infectedbindividuals were analyzed with the present methodology, and a sensitivity of 76.9% was obtained. However, the data did not correlate on its entirety with conventional molecular testing, such as qPCR and plaque assay. Thus, further work is needed to improve this technique and utilize it as a complementary tool for future pandemics.