The disease we aim to target with !MPACT switched two times during our project journey based on discussions with different types of stakeholders bringing in different arguments. Academical healthcare organizations such as University Medical Centers Utrecht, Rotterdam and Maastricht, and the Catherina Hospital looked at the feasibility and desirability of !MPACT from a science and clinical point of view. All four academic stakeholders confirmed that our final decision, focusing on AAV, is suitable for targeting with !MPACT. Therefore, they shared multiple reasons:

• AAV is an autoimmune disease that is associated with pathogenic autoantibodies. These autoantibodies fluctuate along with flare-ups of the disease. This would make !MPACT responsive to disease activity.
• Patients diagnosed with AAV often experience disease relapses that could be prevented with !MPACT due to its responsive mode of action.
• Current treatments for AAV exist of immunosuppressive drugs such as prednisolone, cyclophosphamide, and Rituximab are experienced as burdensome by patients. This is also validated by the AAV patients we interviewed. Due to the temporal and local action of !MPACT, academic hospitals confirmed that !MPACT should have fewer side effects.
• The current treatment procedure for AAV is intense for both the patient and the hospitals as current treatments are accompanied by multiple hospitalizations and follow-up monitoring. Academic hospitals confirmed !MPACT would significantly reduce workload in hospitals since it is administered only once
• Patients are currently life-long dependent on medication, which would no longer be the case when !MPACT is used.

Industrial parties, such as pharmaceutical company Novartis, provided different arguments why AAV is a suitable disease to target with !MPACT:

• AAV is a rare disease that opens opportunities for an orphan disease status. This status is accompanied by many advantages in terms of laws, regulations, and patent applications.
• Research on rare diseases is most often inherent to less fierce competition.
• AAV has high morbidity, and even though AAV is classified as a rare disease (prevalence less than 1:10000), the market of new therapies for AAV is large enough for a sufficient turnover.
In conclusion, numerous stakeholders from different fields of expertise confirmed AAV as a feasible disease to treat with !MPACT.

IL-10 as a therapeutic agent is advised by Utrecht UMC and Novartis and subsequently validated in conversations with Erasmus MC, Maastricht UMC, and the Catherina Hospital. All stakeholders acknowledged the anti-inflammatory nature of IL-10. They confirmed that if !MPACT's IL-10 production is well-regulated, it should reduce AAV-associated inflammation of the small blood vessels. Moreover, they confirmed that IL-10 is currently being studied as a therapeutic strategy for a varity of auto-immune diseases, including rheumatoid arthritis, and IL-10 has a high potential for other autoimmune diseases.

For treatment with !MPACT, immune cells from patients are isolated from the patient and subsequently genetically engineered to produce IL-10, after which they are injected back into the patient (see Project Description). MedTech Company CiMaas and Erasmus MC advised two different cell types to "harvest" from the patient. CiMaas was convinced that memory cells were required for our therapy to be effective in the real world. Hence, the engineered cells of !MPACT remain in the body for an extended period of time, allowing them to respond to AAV disease flare-ups. Therefore, CiMaas recommended using B cells because memory B cells are formed when antigens bind. Erasmus MC suggested using T cells as they are already natural producers of IL-10, have large mobility, and good manufacturing practice facilities for engineered T cells are already available. CiMaas and Erasmus MC confirmed that the required genetic modification for !MPACT could be implemented in the genes of their respective immune cells. Moreover, they both expected their immune cell type to be suitable for !MPACT, and effective in treating AAV. In conclusion, both T and B cells could be viable options for !MPACT based on feedback from the experts. We assessed which immune cell type should be isolated from the patient based on the arguments of the experts and literature study in the proposed implementation.

The genetic modification and treatment procedure required for !MPACT is very similar to Chimeric Antigen Receptor (CAR) T cell therapy. CAR T-cell therapy is a treatment against various types of cancers such as lymphomas and leukemias based on genetically engineered T cells.⁹ Several stakeholders such as Johnson & Johnson, the RIVM (National Institute for Public Health and the Environment), and the European Medicines Agency (EMA) explained that the resemblance of !MPACT to CAR-T cell therapy significantly increases the likelihood of !MPACT reaching the real-world market.

Johnson & Johnson stated that production facilities and good manufacturing practices (GMP) for T-cell therapies are already available and optimized. This lowers the cost of new production facilities while ensuring high-quality cell therapies with a short vein-to-vein time. Johnson & Johnson has thus ensured the consistent and efficient production of !MPACT is secured. Moreover, the RIVM emphasized that obtaining licenses for the production of !MPACT and the use of !MPACT outside the lab is much easier due to its similarity to CAR T-cell therapies. Environmental Risk Assessments (ERA) for CAR-T could be used as a starting point for !MPACT and license applications are likely to be accepted faster. Moreover, the risks of the production of cell therapies, safety measures, and transport protocols are familiar and available. Lastly, the EMA expects that market authorization goes faster since the EMA has gained more experience in assessing the efficacy and safety of ATMPS and T-cell therapies due to the emerging CAR T-cell therapies. Besides, the preparation of clinical studies of these complex therapies is previously designed and could be used for clinical trials of !MPACT. In conclusion, the facilities to bring !MPACT in the real world are already available and optimized which according to Johnson&Johnson, RIVM and the EMA significantly increases the probability of successful market entrance.

The business model and financial plan are validated by pharmaceutical companies such as Novartis, Organon, and Thermo Fisher, and new ventures such as CiMaas, and Ambagon Therapeutics. Both the large multinationals as well as the small new ventures confirmed that the commercialization strategy, market analysis, and financial forecast are viable. More information about the business strategy and its viability can be found on the Entrepreneurship page.

To validate that the GEMS receptor construct could be altered to secrete a therapeutic protein, we first focused on the secretion of IL-10. The therapeutic protein IL-10 was suggested as output by University Medical Centre Utrecht and pharmaceutical company Novartis. Morever, literature research also backs up this choice (see Project description). IL-10 is an anti-inflammatory cytokine, and this characteristic suggests the therapeutic potential of IL-10 administration in autoimmune disease.² The modularity of the GEMS receptor system allows us to alter the expressed protein relatively easily. Substitution of the SEAP reporter protein with IL-10 in the expression vector pLS13, gives us the correct plasmid for IL-10 secretion. The SEAP plasmid (pLS13) is activated by the transcription factor dimerized STAT3. Hence why it is activated by the JAK/STAT signaling pathway and this plasmid can be used as a basis for the IL-10 secretion. To obtain the IL-10 secretion plasmid, a human IL-10 gene was designed and ordered via IDT. Next, the SEAP gene is swapped out with this gBlock using the restriction and ligation technique.

Experiments

To test if we could produce the therapeutic protein IL-10 by activation of the receptor, multiple cell experiments were performed. We transfected the receptor construct to detect RR120 molecules, the STAT3 producing plasmid, and the previously mentioned IL-10 secretion plasmid into human embryonic kidney cells (HEK293T). In this way, the system is directly tested in human cells.

After this transfection, different concentrations of RR120 ligand were added to the cells to activate the GEMS receptor and the pathway. Receptor induction by RR120 should result in the expression and secretion of IL-10 (Figure 3A). After an incubation time of around 48 hours, the samples were analyzed using an IL-10 ELISA. Successful IL-10 expression was observed as the samples turned from blue to yellow during the ELISA sample preparation. Moreover, the IL-10 concentrations were quantified, which can be seen in Figure 3B.

This verified that the cells are able to excrete IL-10 after activation of the GEMS receptor, which meets our first requirement.

Following the successful expression of IL-10 using the GEMS receptor, we wanted to determine if antibody-induced receptor activation of the GEMS platform was possible. To our knowledge, this is not yet described in the literature, and, therefore, it is an important research question. Our receptor must be able to get activated by ANCAs (antibodies) which are abundantly present in AAV. These ANCAs have an affinity for the protein proteinase 3 (PR3), making this PR3 protein a suitable affinity domain for our synthetic receptor.³

Design of the affinity domain

As no literature about antibody-induced GEMS receptor activation exists, we had to design our receptor constructs’ affinity domain based on estimations. The binding of the antibodies to the receptor should induce receptor dimerization. We, therefore, looked into introducing a linker between the receptor scaffold (EpoR) and the affinity domain (PR3). This allows the antibody to bind bivalently to the two PR3 proteins in its optimal position.

Antibodies with identical antigen binding domains bind bivalently to their corresponding antigen, if the antigen separation is between 30 and 170 Å.4 As seen in Supplementary Figure 2 of the lab results. The EpoR and PR3 have an estimated diameter of 55 Å and 45 Å respectively. We, therefore, wanted to introduce a linker with a size between 10 Å and 60 Å. Moreover, a linker which is too short, will also hamper receptor dimerization. Linker lenghts were calculated using an online calculator based on the worm-like model.⁴

Next, the PR3 affinity domain needed to be designed in such a way that it could bind ANCAs. The DNA sequence for the design of this PR3 affinity domain was obtained from UniProt (ID: P24158).⁵ Because we wanted to eliminate the original biological function of the PR3 protein (hydrolyzation of proteins), we chose to remove the peptide sequence in front of the PR3 protein. This peptide sequence normally ensures that the PR3 protein becomes active.⁶

Unfortunately, the affinity of ANCAs for PR3 is not described in the literature. We hypothesize that a low affinity of ANCAs to PR3 could hamper the dimerization step of the two affinity domains. To rule out the possibility, we introduced a hemagglutinin tag (HA-tag) next to the PR3 affinity domain later in the project. The HA-tag is expertly characterized and is often used as a method for tagged protein detection because of its high affinity to its antigen.⁷ A schematic representation of the receptor construct is shown in Figure 4.

To obtain the plasmids encoding for the receptor constructs 0_PR3, 8_PR3, 0_HA_PR3, and 8_HA_PR3 we can use the GEMS construct described in the paper by Scheller et al.. Due to the modularity of the GEMS platform the RR120 VHH affinity domain can be swapped out for the linker, (HA-tag), and PR3. This way the plasmids including the 0_PR3, 8_PR3, 0_HA_PR3, and 8_HA_PR3 receptor constructs are made. For more information on this, please refer to the Notebook.

Experiments

After the construction of the receptor constructs, we wanted to determine if antibodies can bind and activate GEMS receptor constructs. This can be validated by two kinds of experiments. First, the antibody-induced receptor activation is validated using a SEAP assay. Next, flow cytometry experiments are performed to determine if the antibodies can bind to the affinity domain.

Cell experiment

To validate that the receptor constructs, located on the cell membrane of the !MPACT cells, can be activated via antibodies, we transfected different receptor constructs (displaying PR3 and HA) in HEK293T cells. To determine the activation of the receptors, we made use of the reporter protein SEAP (see Figure 5). By measuring the activity of the SEAP, an estimation can be done on the activation of the receptor constructs.

We transfected 0_PR3, 8_PR3, 0_HA_PR3 and 8_HA_PR3, together with the STAT3 and SEAP producing plasmids into HEK293T cells. Antibodies, specific for the antigen displayed on the receptor construct, were added, after which the cells were incubated for 40 hours. SEAP activity was read out using a plate reader.⁸ As no change is seen in the SEAP activity between the ligand-induced and uninduced conditions, we can confirm that the antibodies were not able to activate the receptor. Possible reasonings for this phenomenon and figures of the results can be found on the Results page.

Flow cytometry

Because the unsuccessful receptor activation could be caused by a poor binding of the antibodies to their affinity domains, we wanted to investigate antibody binding via flow cytometry. Similar receptor constructs, as described before, were transfected and stained with labeled antibodies. From the FACS results we concluded that a binding percentage of around 10 % is present for all the different GEMS constructs we made. For a more in-depth analysis of these results, have a look at the Results page. We hypothesize that the binding percentage of 10% could explain the unsuccessful activation of the receptor, which is seen during the cell experiments.

Conclusion

In conclusion we have clearly showed that antibodies are able to bind the affinity domain, which is a great starting point for further research. We are confident that after the optimalization of the binding between the receptor construct and antibody, antibody-induced receptor activation can be achieved. Multiple different suggestions are given at the end of the results page, which go deeper into this optimalization.