Spinal cord injury (SCI) occurs as a result of damage to the functional units of the spinal cord, leading to symptoms ranging from chronic pain to complete loss of sensation, motor and autonomic functions (Alizadeh, Dyck and Karimi-Abdolrezaee, 2019).
SCI is characterised by primary and secondary injuries. The primary injury is caused by the initial traumatic event, typically from road accidents and falls. The secondary injury is caused by a series of biological and functional changes, which lead to nerve degeneration, growth inhibition and inflammation in the injury site (Bradbury and Burnside, 2019).
The severity of trauma experienced at the site of the primary SCI directly relates to the broad spectrum of symptom profiles present upon hospital admission. This not only increases the need for effective and immediate treatment, but also highlights the importance of emergency medicine protocols as a basis for quick SCI diagnosis and spinal cord stabilization protocols to reduce the likelihood of secondary injuries. Diagnosis of SCI begins with the assessment of physical symptoms to gain an understanding of the type of injury sustained, these symptoms include (Roberts, Leonard and Cepela, 2017):
The type and severity of injury sustained can then be classified using the American Spinal Injury Association (ASIA) impairment scale, which ranges from A-E, with A being the most severe impairment classification and E the least. This scale is instrumental in providing both a diagnosis and prognosis for the patient, as outlined in Table 1. This scale allows for the identification of complete and incomplete injuries, as well as the identification of subsequent paraplegia or tetraplegia (Ahuja et al., 2017). Complete SCI refers to the complete loss of motor or sensory function below the site of injury, while incomplete injuries have retention of some motor and sensory function. The identification of incomplete and complete injuries allows for medical staff to identify the extent of motor deficits sustained after SCI and thus determine whether the patient is suffering from paraplegia - the loss of motor and sensory function at the legs or tetraplegia - the loss of motor and sensory function at both the arms and legs. This then allows for the development of efficient and relevant medical and rehabilitation protocols (Alizadeh, 2019)
Table 1: ASIA impairment scale demonstrating the grading system for an SCI according to different types of injuries (Roberts, Leonard and Cepela, 2017).
As part of our Phase I project we realised the importance of ensuring our scaffold could provide sufficient functional recovery to improve quality of life. Given the relatively increased symptom severity following cervical SCI, this region provided the optimal target for our therapy. However, during Phase II of our project, further research and discussions with experts in the field, Dr Jerry Silver and Mr Gordan Grahovac, who suggested that surgical procedures at the uppermost regions of the cervical spinal cord (C1-C5) carry too great a risk due to its crucial role in controlling respiratory function. As such, we adapted our proposed design this year to incorporate C6-C7 cervical lesions to reduce the risk of impairing respiratory function in the case of surgical complications (Alizadeh, Dyck and Karimi-Abdolrezaee, 2019). Additionally, we recognised the need for a more holistic and inclusive therapy given our limited treatment options within the cervical region.
We discovered that the thoracic region is a valid treatment location as the risk of disrupting respiratory function is small in comparison to intervention within other regions, allowing us to provide a relatively safer treatment and treat a larger number of patients (Lujan and DiCarlo, 2020). As our scaffold is a personalised therapy printed according to the dimensions of the individual lesions, incorporating a new region of the spinal cord into our target patient pool will not require any additional changes to our project design.
Following identification of the type of SCI we wish to treat, it is important to consider the time period during which we would administer treatment. In this regard it is also important to understand how SCI progresses over time. SCI progresses over 3 stages: acute, followed by sub-acute and finally chronic (Table 2) (Alizadeh, Dyck and Karimi-Abdolrezaee, 2019). The progression of phases following injury all contribute to the formation of a glial scar at the site of injury. The glial scar consists of reactive astrocytes that act both as a physical and chemical barrier for axonal regeneration.
Table 2: Pathophysiological stages in SCI progression.
During the acute phase, macrophages present in the inflammatory microenvironment and the glial scar of SCI lesions come into physical contact with severed axons, releasing proteases to cause axonal dieback (Tran, Warren and Silver, 2021). Chondroitin sulphate readily interacts with negatively charged proteins in the ECM, binds and synapses with axonal end bulbs and disrupts the balance of the spinal cord microenvironment (Fan et al., 2018). Further information regarding SCI pathophysiology can be found on our SCI Biochemistry Page.
As mentioned in Table 1, a glial scar forms within the lesion cavity during the sub-acute phase and persists into the chronic stage. The glial scar is composed of fibroblasts, astrocytes, chondroitin sulphate proteoglycans (CSPGs) and oligodendrocyte progenitor cells (OPCs) (Silver and Miller, 2004). It has a dual role in the injured spinal cord, by both inhibiting axonal regeneration and protecting axons from further demyelination (Yang et al., 2020). As reactive astrocytes divide and surround the lesion core, the glial scar can stabilise fragile central nervous system (CNS) tissue post-injury. This repairs the blood-brain barrier (BBB), limits inflammatory response and cellular degeneration (Silver and Miller, 2004). Chronically, however, the chemical inhibitory molecules, CSPGs secreted by astrocytes in the glial scar, are potent inhibitors for axonal regeneration (Tran, Warren and Silver, 2021).
At the tissue level, haemorrhage, ischemia and the death of oligodendrocytes leads to the demyelination of lesioned axons despite the protective effects of the glial scar. These biochemical events, together with axonal regrowth inhibiting molecules, cause impaired signal transduction (Fan et al., 2018). At the cellular level, endogenous neural stem cells primarily differentiate into astrocytes instead of functional neurons, resulting in an imbalanced proportion of differentiation. Additionally, the post-SCI cellular microenvironment is unfavourable for M2 macrophages, which have anti-inflammatory and tissue repair properties, furthering inhibiting recovery, see Figure 1 (Fan et al., 2018). At the molecular level, the microenvironment following an SCI is characterised by the down-regulation of factors advantageous for nerve recovery, such as neurotrophic factors, and the up-regulation of factors which inhibit nerve outgrowth, including cytokines, and chemokines (Fan et al., 2018) prolonging an inflammatory response and limiting functional recovery.
Our therapy is tailored to tackle the inflammatory biochemical microenvironment of the lesion. We aim to achieve this through the implementation of a mussel foot protein (MFP)-coated 3D-printed polycaprolactone (PCL) scaffold at the lesion site during the chronic phase. Our design aims to promote a permissive environment for axonal regeneration
whilst maintaining a mechanically stabilised environment
Figure 1: The inhibitory environment posed by the glial scar. The inhibitory environment is characterized by astrocytes surrounding the scar core which secrete CSPGs. CSPGs interact with the ECM and approach axonal tips, which stabilises injured tissue but simultaneously limits axonal growth
Chondroitinase ABC (ChABC) is a bacterial enzyme from Proteus Vulgaris found to degrade CSPGs and thereby has the potential to overcome their inhibitory effects. ChABC functions by digesting glycosaminoglycan (GAG) chains of CSPGs in the ECM of the SCI lesion and has been shown to promote axonal regeneration, neuroplasticity and neuroprotection. Additionally, it has been observed to encourage locomotive functional recovery and improve autonomic function (Bradbury and Carter, 2011). The SCI Biochemistry Page provides more detail regarding the mechanism of ChABC in CNS therapeutics.
In collating our research into the actions of ChABC, we decided to implement ChABC microinjections at the rostral and caudal ends of the lesion, alongside the mussel foot protein-coated 3D-printed polycaprolactone (PCL) scaffold within the lesion site. Microinjections of ChABC during scaffold implantation surgery and post-surgery aim to improve the permissiveness of the SCI microenvironment for axonal regrowth by reducing the inhibitory effects of CSPGs. This facilitates axonal regeneration throughout the scaffold and across the lesion, enabling axons to reconnect with distal interneuron targets.
During our research, we identified several challenges in incorporating ChABC within our treatment, including its thermal instability at physiological temperature, restricted enzymatic activity, and poor sustained local delivery when implemented in therapies in vivo (Bradbury and Carter, 2011). After talking to Dr Bradbury and Dr Hettiaratchi, we decided to improve the thermostability of ChABC using computational mutagenesis and modelling to ensure the effectiveness of our therapy. See our Modelling Page for more information on how we improved the thermostability of ChABC.
Unfortunately, there is no treatment to reverse the impacts of SCI to this date, although many promising treatments are currently being developed, involving drug administration and prostheses to encourage axonal regeneration and functional recovery (Ziemba and Gilbert, 2017). Instead, current SCI treatments focus on preventing further damage post-SCI, and reducing secondary symptoms including bowel and bladder issues, pressure ulcers and respiratory infections.
However, the treatment options being investigated do not always encourage functional recovery in vivo due to the low efficacy of the drugs tested in clinical trials. This is because the administration strategy of most drugs; high dosage bolus injection followed by a continuous infusion is non-localised, non-continuous, invasive, and exposes patients to infection (Ziemba and Gilbert, 2017). As the cellular response of SCI progresses over a long time frame, together with its complex biochemical microenvironment and the presence of a physical glial scar barrier, there is a need for innovative measures to deliver pharmaceutical agents locally and sustainably.
Rehabilitation is often another main focus alongside treatment, in order to maximise physical and social independence of patients, with an emphasis on redeveloping and adapting motor skills, and patients’ mental health (Ziemba and Gilbert, 2017). We recognise the need to provide a full treatment plan from the point of patient admission to integration back into society and improved quality of life, we have developed a guide on the available rehabilitation protocols on our Education Page.
We, Renervate Therapeutics are working to provide a novel and holistic treatment for SCI implementing thermostabilized ChABC microinjections to target the SCI lesion site. This is combined with a 3D bioprinted PCL scaffold implantation in the lesion, with designed macro- and micro-architecture to promote axonal regeneration, coated in MFP.
Polycaprolactone (PCL) is an FDA-approved, synthetic polymer that is commonly used in tissue engineering scaffolds and wider regenerative medicine applications (Li and LaBarbera, 2017). This material was chosen for our scaffold based on the research carried out in Phase I. Specifically, PCL matches key material specifications including optimal biomechanical properties, and is biodegradable and bioresorbable, while generally being nontoxic (Díaz et al., 2014), with no release of acidic by-products (Valmikinathan et al., 2011). Furthermore, PCL can be easily tailored to degrade in a chosen time period by adjusting features such as its molecular weight and porosity – as demonstrated in the modelling performed in Phase I of our project. The material can be further optimised for regeneration by considering the architecture of the scaffold, as discussed in the subsequent sections. However, PCL has very limited cell-adhesion capability (Zhang et al., 2019), sparking the idea of incorporating MFPs (specifically PVFP-5) alongside the scaffold.
There is minimal coverage in literature discussing the impact of scaffold macro-architecture on regeneration within the spinal cord. However, Wong et al. (2008) demonstrated that macro-architecture alone can influence the success of a scaffold-based treatment. Five PCL scaffolds with different macro-architectures were investigated including: cylinder, tube, multichannel and open path designs with and without a central core. The open path design accommodates the lateral funiculi where the corticospinal tracts descend (Wong et al., 2008). Overall, scaffolds with open path designs had greater regeneration results, and those without had an adverse effect on the length of the defect. Each of these designs were simulated using Finite Element Analysis in Phase I of our project to investigate the mechanical integrity of each. In these tests, the open path without core design ranked second best – only having a 2.7% increase in maximum stress compared to the version with a core (KCL iGEM, 2020). Further, Wong et al. (2008) did not find statistical significance that the open-path with core performed better with respect to decreasing the defect length, and the former design was described as being ‘slightly’ better in terms of performance. Considering these facts, we decided to simplify our design from an open path with core, to one without. This simplification in design makes the scaffold more accessible - by reducing the complexity of the design it becomes easier to produce and print.
The micro-architecture of our scaffold consists of the structures incorporated to promote axonal regrowth and adhesion. The pore structure, interconnectivity and porosity have to be designed and implemented with careful consideration to ensure that the best possible environment is created for tissue regeneration (Mata et al., 2009).
The method we used to introduce pores into our scaffold design has changed from Phase I to simplify the bioprinting process. During this Phase, the proposed method for this was by creating unit cells whereby the macro-architecture scaffold would be split up into homogenous blocks called voxels. These voxels would be replaced by gyroid-shaped cells and the whole scaffold would be trimmed to our macro-architecture geometry. This ensures the scaffold pore structure is highly interconnected (Germain et al., 2018) but also meant it was very complex, exacerbating the printing process. To overcome this, we have now proposed the use of the log-pile design method, which is also known as the ‘cross-hatch’ method. This method consists of cross-hatching logs of the same diameter stacked layer by layer to the desired height and then trimming the structure to our macro-architecture geometry. This method creates voids surrounding each intersecting log which are the pores of the scaffold. This log-pile design incorporates an interconnected pore structure into our scaffold while ensuring it remains simplified (Kelly et al., 2018).
Furthermore, we altered the size of the pores in our scaffold. Our literature reviews from Phase I dictated that the ideal range for pore sizes of a scaffold would be between 100µm and 400µm (Bayram et al., 2019) and we chose our pores to lie at the centre of this range, between 200µm and 300µm. This year we noted that our chosen range may be too small for the actual scaffold design, as the post-processing of the scaffold after initial printing may cause the size of the pores to decrease. Furthermore, the coating of the mussel foot protein bioadhesive can cause a decrease in the size of the pores. To accommodate this, we have designed the pores of the printed scaffold to be 500µm. This means that following the processing of the printed scaffold the pore size would fall within the aforementioned ideal range.
Our scaffold’s porosity still remains at approximately 60% following the changes to our pore structure and size, which would give our scaffold an elastic modulus of approximately 1.48 MPa (Guarino et al., 2007), matching the average elastic modulus of a spinal cord.
During Phase I, 3D printing was chosen as the fabrication method over electrospinning and solvent casting due to its accuracy and replicability, as well as material resourcefulness. This year we decided to further investigate the field of 3D printing to determine the most optimal method of additive manufacturing for our scaffold design.3D printing encompasses several methods including selective laser sintering (SLS), fused deposition modelling (FDM), multi jet fusion (MJF) and several others.
Figure 2: An illustration of a Fused Deposition Modelling (FDM) printer
SLS uses a high power laser to selectively sinter polymer powder particles. The powder bin is heated up nearly reaching the powder melting temperature followed by a laser scanning the contour of a layer. Every time a layer is completed the platform moves downward and a blade recoats the bin, allowing for a new layer to be lasered. This method allows for printing without supports, since every new layer is supported by the underlying powder making it suitable for very detailed and intricate prints. However, shrinkage needs to be taken into account. As every new layer cools down it decreases in size and internal stresses build up, meaning that a 3-3.5% increase in the original CAD file should be introduced to get a final print of the desired size. Excess powder can also get stuck in the part, which needs to be removed after the part has cooled down.
FDM is based on selectively depositing melted filament material in a predetermined path layer by layer. First, the printer nozzle is heated up to reach the filament melting point which is then extruded onto the printing bed. The nozzle moves along the X, Y and Z axis building up the part layer by layer. To speed up the cooling process some printers use cooling fans within or around the nozzle. FDM has a lower resolution and accuracy in comparison to other 3D printing methods meaning complex structures may not be able to be successfully printed. The lower resolution also means the printed items are more likely to have clearly visible layers meaning they require post-processing to be smooth. However, this method is the most cost-effective and highly accessible method of 3D printing. It is also able to achieve faster printing speeds than SLS.
After thorough literature research and guidance from Professor Trevor Coward, we decided on FDM printing was chosen due to the wide availability of printers in hospitals as well as the best compatibility with the chosen PCL material and cost-effectiveness. For further SLS vs FDM comparison visit Scaffold Engineering.
To further validate our approach, we decided to pursue two different avenues of modelling - computational fluid dynamics (CFD) to simulate the scaffold’s response to the in vivo spinal cord system, and a PVFP-5/Polycaprolactone binding model to assess the amount of protein needed for each scaffold.
Deemed a ‘necessity’ for scaffold design (Ali and Sen, 2017), computational fluid dynamics (CFD) allows for the investigation of the spinal cord’s fluid behaviour and interaction with the scaffold as an effective alternative to complex and expensive experimental tests (Ali and Sen, 2018). Specifically, CFD encompasses the quantitative prediction of fluid-flow phenomena based on the conservation laws (i.e. conservation of mass, momentum and energy) that influence fluid motion—using digital computers (Hu, 2012). Using these simulations, it is possible to predict the permeability of and the wall shear stress (WSS) imposed on the scaffold by the local microenvironment (Ali and Sen, 2017; Ali and Sen, 2018).
The permeability of a scaffold is impacted by its porous properties—such as the number of interconnecting pores—and indicates how well fluid flows through the material (Singh et al., 2018). Specifically, a sufficiently permeable scaffold facilitates diffusion of nutrients, gases and waste disposal, which directly impacts cell proliferation (Ali and Sen, 2018). However, this value should be well tuned to match the target site; where a higher permeability indicates a high flow rate (causing cell washout), and lower permeability results in low nutrient supply (Singh et al., 2018). The accepted value for the spinal cord is approximately in the range of ∼2 × 10−13 m2 for grey matter, and around ∼5 × 10−14 m2 for white matter (Venton et al., 2017). Conversely, the flow-induced WSS provides an insight into the formation of desired tissues throughout the scaffold—where the WSS is a mechanical stimulant of cell growth (Ali and Sen, 2017). Therefore, calculating both of these parameters allowed us to evaluate the scaffold’s potential success, and decide whether the design needs to be altered. The scaffold was simulated using ANSYS software; more in-depth methodology and results may be found on our model page here.
In order to measure the amount of MFP required per scaffold, we did theoretical calculations to determine the area of the scaffold that was accessible. Then we computationally measured the area of the adhesive side of the protein. That way we could determine how many units we would need to cover the scaffold with one layer. Using Avogadro’s constant we then converted that number to micrograms of protein needed. This extrapolation provided us with roughly 1.3 micrograms of protein being necessary to cover a 10 mm scaffold.
In recent years, mussel foot proteins (MFPs) have been extensively investigated for use in tissue bioadhesives, adhesive hydrogels and various other medical devices. Several studies have reported their biocompatibility and non-cytotoxicity in vivo , thereby making them ideal for use in medicine (Santonocito et al., 2019; Choi et al., 2012) . Furthermore, the ability of mussels to adhere to underwater surfaces in turbulent coastal conditions has provided promising routes for the use of MFPs as therapeutic agents (Forooshani and Lee, 2017). During Phase I, we took inspiration from these proteins after learning about ‘underwater adhesives’ from the GreatBay SCIE 2019 iGEM project, and reading the paper ‘Recombinant mussel protein Pvfp-5β: A potential tissue bioadhesive’, (Santonico et al., 2019). We, therefore, decided to research the use of PVFP-5, secreted by the Asian green mussel (Perna viridis) for our project.
Whilst our PCL scaffold provides the basis for our therapy, it is limited by its lack of adhesive ability, which is imperative for cell growth and axonal regeneration (Cheong et al., 2019). This dilemma formed the foundations of our 2020 project in Phase I, where we looked towards using a mussel-inspired bioadhesive to coat our scaffold and promote axonal regrowth.
Pvfp-5 is the first protein to be secreted in the mussel foot of Perna viridis, with emphasised adhesive capabilities due to its high content of the catechol amino acid, L-3,4-dihydroxyphenylalanine (L-DOPA) (Santonocito et al., 2019). From literature, we discovered that L-DOPA is the main residue involved in adhesion in the mussel foot (Nicklisch and Waite, 2012), and we, therefore, identified understanding L-DOPA chemistry as a fundamental aspect of our research this year.
Mussel foot proteins contain an abundance of tyrosine residues, naturally modified to the post-translationally modified form, L-DOPA (Burzio et al., 2000). Along with collagen, MFPs make up the core of the mussel byssus, a bundle of threads secreted by the mussel foot, ending in adhesive plaques (Forooshani and Lee, 2017). L-DOPA residues in PVFP-5 are secreted into these byssal threads, where they are involved in both adhesive and cohesive forces with marine surfaces.
L-DOPA residues form a variety of interactions with different surfaces via hydrogen bonding and coordination bonding (Forooshani and Lee, 2017). They are also involved in sclerotisation, a form of tanning involving the cross-linking of proteins (Burzio et al., 2000). This makes up the cohesive forces of the protein and therefore the structural integrity of the byssal thread. In nature, this cross-linking is achieved through the auto-oxidation of L-DOPA to form DOPA-quinone, which then polymerises into cohesive chains (Burzio et al., 2000). During Phase I, we determined that it is possible to mimic this process in vivo by using the enzyme tyrosinase, which catalyses the hydroxylation of tyrosine residues to L-DOPA, followed by the oxidation of L-DOPA to DOPA-quinone.
An important consideration for us was the binding mechanism between PVFP-5 and PCL, to ensure successful adhesion of our protein to the scaffold. After analysing the chemical structures of both materials, we predicted two main binding modes: hydrogen bonding and hydrophobic interactions. The first one occurs between the oxygen of the ester groups of PCL and the diphenol of L-DOPA.
To confirm our prediction we consulted Professor Herbert Waite, who validated our hydrogen bonding model, but also proposed an additional binding mode in the form of cysteine-dependent covalent chemistry. On the other hand, hydrophobicity was suggested by Dr Barry and is likely to occur between the non-polar residues of PVFP-5 and the aliphatic straight chain of polycaprolactone. As the binding of PVFP-5 to PCL has not previously been reported in the literature, these predictions are still subject to speculation. However, this meeting gave us confidence that the binding of these two materials was likely to be successful via the aforementioned methods.
In combination with the micro-architecture of the scaffold and ChABC, our hope for using PVFP-5 as a bioadhesive coating is that axons will be able to adhere to, navigate and migrate along the scaffold and across the lesion to reconnect with remaining host axons, promoting functional recovery. Furthermore, by coating the scaffold the reduced likelihood of a foreign body reaction is reduced, compared to if the scaffold were to be implanted directly into the spinal environment.
As a result of using a new protein sequence for PVFP-5, we decided to generate a structural model to allow full structural characterisation of our protein before we initiated our lab work. Our new sequence contains a high number of cysteine residues (18) - all of which we predicted to be involved in disulphide bond formation. We initially used our modelling approach from Phase I, which involved full structure homology modelling, then a molecular dynamics simulation to settle the atomic irregularities within the structure. However, this year with the large number of disulphide bonds required in our structure (9), this approach was not successful. Following PFAM analysis of our protein sequence, we learnt that our protein is composed of three EGF-like domains. Following the advice provided by Dr Andrew Beavil, we altered our modelling approach and designed a new one in which we (homology) modelled each EGF-like domain of our protein individually, then ran molecular dynamic simulations on the whole structure. We did this after bringing the domains together to ensure the correct formation and orientation of our desired disulphide bonds. To further validate our new modelling workflow, a structural prediction simulation using the highly acclaimed AlphaFold2 artificial intelligence software was used, which confirmed the order of our predicted bonds. Further details on our structural model can be found on our modelling page, and a guide to troubleshoot structural models with a high disulphide bond content using our modelling workflow can be found on our contributions page.
To validate our research from Phase I, our aim for Phase II was to synthesise our own recombinant PVFP-5, as well as design and implement a co-expression system with PVFP-5 and a tyrosinase enzyme (secreted by Bacillus megaterium) to allow for the post-translational modification of tyrosine residues into L-DOPA. Our inspiration for this co-expression system was a study conducted by Santonocito et al., who reported a successful synthesis of PVFP-5 in vivo (Santonocito et al., 2019).
Furthermore, from contacting Dr Byeongseon Yang, we developed a significant understanding of the benefits of using a co-expression vector in place of in-vitro methods. However, limitations to laboratory access due to the COVID-19 pandemic resulted in us being unable to complete our co-expression of PVFP-5 with tyrosinase. Despite this, we were able to synthesise and express recombinant PVFP-5 using a variety of competent E. coli cell lines to investigate the most optimal for production of our protein in its soluble form. Due to the numerous disulphide bonds in our protein structure, we specifically used the Rosetta-Gami B (DE3) and SHuffle (DE3) due to their enhanced ability to form disulphide bonds in the target protein (Lobstein et al., 2012).
Although the oxidation of L-DOPA to DOPA-quinone is required for the polymerisation process, a problem we encountered during our research was the reduced adhesive ability of DOPA-quinone in comparison to L-DOPA (Forooshani and Lee, 2017). As PVFP-5 forms the bioadhesive component of our therapy, an important consideration for us was finding a way to prevent this auto-oxidation and maintain the adhesive properties of our protein. After an extensive literature search, we found that the most promising prevention methods include the use of borate, catechol protecting groups for L-DOPA, or maintaining a low pH during expression and purification (Forooshani and Lee, 2017; Choi et al., 2012). More detail on the different oxidative prevention methods can be found here.
As L-DOPA is highly pH-sensitive, we decided that the most favourable route would be lowering the pH of our buffers during the protein purification stage. As described by Professor Waite, an acidic pH helps stabilise the catechol substituent of L-DOPA, thereby protecting the adhesive properties of PVFP-5. He also suggested the use of either acetic or hydrochloric acid to achieve this, and we, therefore, decided to implement his advice when devising a purification protocol. Although we were unable to complete this updated purification in the lab due to COVID-19 restrictions, we have drafted protocols for future PVFP-5 expression and purification, which can be found here.