Engineered human heavy-chain ferritin to improve anti-tumour drug delivery performance

Ferritin exists ubiquitously in nature. It has been used in medical application for decades because of its unique and promising plastic structure, together with high biocompatibility and stability. The most widely adopted mammalian ferritin, is human heavy-chain ferritin (HFn), which has been employed as an anti-tumour drug loading platform. HFn comprises 24 subunits to form a 12 nm hollow sphere. Various kinds of anti-tumour drugs have been explored to be loaded onto HFn by diverse drug loading approaches. However, protein loss and undesirable drug loading ratio are bottlenecks to practical drug loading. In addition, HFn has shortcomings such as short half-life in circulation. Different half-life extenders or other functional characteristics such as enhanced tumour targeting ability, have been equipped to HFn to improve its efficacy and broaden its application. In this research thesis, focus is on improvement of HFn performance as an anti-tumour drug delivery platform through 1) functionalisation by fusion, and 2) drug loading approach. Functionalisation and drug loading investigation were initially conducted in 2 strategies and compared. The first strategy tries to extend the half-life of HFn, improve tumour targeting ability, and enhance the drug loading performance. It fuses 2 functional peptides, PAS and RGDK, to HFn. PAS is a 40 aa peptide composed of repetitive P, A, and S residues. It can bind to water molecules and enlarge the hydrodynamic volume to extent half-life in circulation. RGDK is a four-residue tumour targeting peptide. A thermally induced passive diffusion approach was explored to load small molecule doxorubicin hydrochloride (DOX) to HFn to improve drug loading performance. The second strategy fuses HFn with a peptide drug called pro-apoptotic peptide (P). The sequence of P is KLAKLAKKLAKLAK. It can selectively kill tumour cells by disrupting mitochondrion membrane. This strategy makes use of the alternative fusion approach to load drug to guarantee loading efficiency and simplify loading process. RGDK was also employed in this strategy to improve HFn tumour targeting ability. Eight HFn based proteins were designed for these 2 strategies. Two of them are controls, HFn and E-helix truncated HFn (sHFn). Three PAS functionalised HFns are HFn-PAS, HFn-GFLG-PAS-RGDK and HFn-PLGLAG-PAS-RGDK. GFLG and PLGLAG are two enzyme-cleavable sites leading to the removal of PAS-RGDK from HFn after reaching tumour tissues. They are 4 aa and 6 aa long, respectively. sHFn is not employed in strategy 1 and HFn is the only control. Another 3 proteins P-HFn, sHFn-P, and sHFn-P-RGDK. HFn and sHFn are used and they are both controls. All proteins were successfully expressed in Escherichia coli (E. coli). Transmission electron microscopy (TEM) and size-exclusion chromatography (SEC) results showed that 2 controls and 3 PAS functionalised HFns were expressed in soluble assembly. However, 3 P functionalised HFns failed to self-assemble. P-HFn was expressed in inclusion bodies (IBs), and sHFn-P and sHFn-P-RGDK in soluble monomers. Without the correct structure, 3 P functionalised HFns could not be applied as anti-tumour drugs. Therefore, strategy 2 failed and there was no need for the 3 P functionalised HFns to undergo further bioactivity evaluations. Molecular docking and molecular dynamic (MD) simulation were conducted to investigate the impacts of P peptide on HFn assembling. Results demonstrate that the high density of positive charge and the hydrophobicity of P were the main causes for the failure of P functionalisation. In strategy 1, the 3 PAS functionalised HFns and their control HFn underwent further experimental investigations. Following the expression, HFn and 3 PAS functionalised HFns were purified using heat-acid precipitation and chromatography. Purification of functionalised HFns were optimised and it is worth to note that the inserted functional peptides resulted in the conformation change and therefore impacted the purification process. Fusion of PAS peptide has led to an E-helix turnover phenomenon, resulting in a significant decline of the stability. In structural comparisons of HFn and 3 PAS functionalised HFns, PAS has enlarged the hydrodynamic volumes of assemblies due to its hydration ability. However, RGDK peptide did not make any significant impact on HFn conformation. Following preparation of high-quality HFn-based proteins, a thermally induced passive diffusion approach was introduced to load drug efficiently. Conditions in thermally induced passive diffusion approach was optimised for both HFn and HFn-GFLG-PAS-RGDK through an orthogonal trial. Variables were heating temperature, heating time and buffer pH. Condition of pH 7.5, 50 C, 6 h was found to be the optimal for all 4 HFn-based proteins. It achieved a significantly improved protein recovery yield and loading ratio in comparison with previous loading approaches. Protein recovery yield of functionalised HFns were lower in contrast with HFn, because of the stability decrease after fusion. Molecular docking combined with MD simulation revealed the mechanisms behind DOX loading process. Finally, the anti-tumour performances of DOX loaded 4 HFns (i.e. DOX loaded HFn and 3 PAS functionalised HFns) were assessed in detail in vitro and in vivo tests. Intracellular distribution assay demonstrated DOX carried by these 4 HFn-based proteins could accumulate in tumour cell nucleus to exert function. MTT assay showed that RGDK in 2 PAS-RGDK functionalised HFns significantly improved the cytotoxicity. Cellular uptake test showed that RGDK had significantly enhanced tumour cell uptake efficiency. In animal tests, biodistribution results proved the improvement of tumour targeting ability by RGDK in HFn-GFLG-PAS-RGDK and HFn-PLGLAG-PAS-RGDK, compared with HFn and HFn-PAS. Pharmacokinetic test showed an approximate 5 times extension of half-life in circulation resulting from PAS peptide fusion. In anti-tumour growth test, 2 PAS-RGDK functionalised HFns achieved the highest anti-tumour efficacy. Overall, this work has significantly improved HFn as an anti-tumour drug delivery platform. Functionalisation with PAS and RGDK peptide can improve anti-tumour drug delivery performance through half-life extension and improvement of tumour targeting. Thermally induced drug loading approach can be used in both HFn and functionalised HFns to obtain a desirable loading performance. Mechanism of HFn drug loading has been revealed by experiments aided with computation tools including molecular docking and MD simulation. The findings will be of immediate benefit and interest to a wide range of researchers and manufacturers for drug delivery using ferritin.Thesis (Ph.D.) -- University of Adelaide, School of Chemical Engineering and Advanced Materials, 202