Systematic studies into uniform synthetic protein nanoparticles

Nanoparticles are frequently pursued as drug delivery carriers due to their potential to alter the pharmacological profiles of drugs, but their broader utility in nanomedicine hinges upon exquisite control of critical nanoparticle properties, such as shape, size, or monodispersity. Electrohydrodynamic (EHD) jetting is a probate method to formulate synthetic protein nanoparticles (SPNPs), but a systematic understanding of the influence of crucial processing parameters, such as protein composition, on nanoparticle morphologies is still missing. Here, we address this knowledge gap by evaluating formulation trends in SPNPs prepared by EHD jetting based on a series of carrier proteins and protein blends (hemoglobin, transferrin, mucin, or insulin). In general, blended SPNPs presented uniform populations with minimum diameters between 43 and 65 nm. Size distributions of as-jetted SPNPs approached monodispersity as indicated by polydispersity indices (PDISEM) ranging from 0.11–0.19. Geometric factor analysis revealed high circularities (0.82–0.90), low anisotropy (<1.45) and excellent roundness (0.76–0.89) for all SPNPs prepared via EHD jetting. Tentatively, blended SPNPs displayed higher circularity and lower anisotropy, as compared to single-protein SPNPs. Secondary statistical analysis indicated that blended SPNPs generally present combined features of their constituents, with some properties driven by the dominant protein constituent. Our study suggests SPNPs made from blended proteins can serve as a promising drug delivery carrier owing to the ease of production, the composition versatility, and the control over their size, shape and dispersity.


Scanning electron microscopy
The SEM micrographs presented in this article are cropped and magnified from larger fields of view for clarity and ease of particle observation. Figure S1 provides full field of view of the typical images utilized to build population statistics throughout the study.

Dynamic light scattering
The particle size distributions were measured with dynamic light scattering (Malvern ZSP ZEN-5600). The protein material was chosen with refractive index of 1.45 and absorbance of 0.001.
Phosphate-buffered saline was used as the dispersant with refractive index of 1.332 and viscosity of 0.9074. Measurements were done at 25 °C, at 3.00 mm position in a disposable microcuvette. (ZEN0040).

Analysis software
FIJI (a distribution of ImageJ v1.53c) was used for all image analysis. OfficeLibre Calc was used for all data manipulation and for the generation of summary statistics. Graphpad Prism 9.0 was used for presentation of distributions, scatter plots, and violin plots. Origin 9 was used for peak extraction.

SEM Analysis via ImageJ/FIJI
SEM images are collected as lossless TIFF files and were processed via ImageJ.

Data processing via open source spreadsheet software
In order to obtain results that will allow for comparison of the data from imaging with DLSbased PDI results for the systems, histograms and individual data were calculated as volumebased values in order to generate a calculated SEM-based PDI value (denoted here as PDI SEM ) and to present an intensity-based analog (iSEM) for comparison to intensity-based DLS. iSEM distributions, which were calculated from nSEM × (d/2) 3 and then normalized. This was done to provide dry particle (as manufactured) analysis that is comparable to the nDLS and iDLS results (as used in solution after post-processing). The entirety of the analysis for this step was done utilizing OfficeLibre Calc. The various expressions of diameter populations were evaluated using statistical analysis including t-test, IQR, and ANOVA.

Two-dimensional analysis
To understand how the geometric characteristics for blended SPNPs are influenced by their constituents, we quantitatively and qualitatively described the SPNPs. The diameter was compared to other geometric attributes (min. diameter, anisotropy, circularity, and roundness).
The linear regressions of the paired x-y data sets were created and a scoring factor was used (from 0 to 10) to describe the extent of similarity to the monospecies SPNPs. For example, to understand how a blend of HSA and transferrin SPNP resemble their constituents, this scoring system can be applied. The scoring factor represents a convolution of the relative agreement of the blended regression slope and the agreement of the regression strength (⟨r 2 ⟩) when compared to the monospecies SPNPs. This is done by treating the slope of the blend regression as a linear combination of the slopes of the constituent regressions, scaling based on the extent of agreement between the strengths of the regressions.   Figure S2: Two-factor individual analysis for the HEM series. Scatter plots of minimum diameter, anisotropy, circularity, and roundness vs diameter.      Figure S5: Two-factor individual analysis for the INS series. Scatter plots of minimum diameter, anisotropy, circularity, and roundness vs diameter. S13