Clearance of residual genome editing components used for ex vivo genome-editing of allogeneic cell therapy products.

BACKGROUND AIMS: Clustered Regularly Interspaced Short Palindromic Repeats (CRISPR)-associated genome editing (GE) components (e.g., nucleases, guide RNAs (gRNAs), and plasmids) are used to genetically modify cells during development of ex vivo genome-edited cell therapies. Prolonged presence of GE components may increase the risk of unintended genome modifications (e.g., off-target editing and chromosomal rearrangements). This risk is a function of the stability of the GE components, culture conditions (i.e., culture length, media changes, etc.), and the nature of the GE component (i.e., only plasmids can be integrated into a cell's genome). Testing for residual GE components on ex vivo genetically edited drug products is generally recommended in current regulatory guidance (CBER 2024). For allogenic cell therapies derived from induced pluripotent stem cells (iPSC), cells typically undergo clonal selection and extensive culturing following completion of genome editing. This post-engineering clonal selection substantially reduces the amount of residual GE components while the long-duration cell culture significantly reduces the presence of active residual GE components. Here we present a case in which the need for testing of the drug product for residual GE components has been eliminated. METHODS: In silico modeling was used to estimate clearance mechanisms across a variety of relevant assumptions, including disposition of extracellular GE components via media changes and dilution of intracellular GE components via cell expansion. Determining the ability of each GE component-alone or in complex with other GE components-to modify genomic material was assessed by a series of both in vitro and ex vivo (i.e., engineering cells) studies. For the in vitro studies, a DNA cutting assay was developed to assess the ability of the component to cut a representative DNA strand. For the ex vivo modification of cells, an assessment of the knock-out of the relevant gene was completed by flow cytometry specifically assessing the presence or absence of protein expression on the modified cells. The persistence and stability of GE components were examined under cell-mimicking conditions and in ex vivo modified cells. The components were stressed under multiple conditions mimicking a range of culture conditions and tested in the aforementioned DNA cutting assay. The presence of residual gRNA was directly assessed in the ex vivo modified cells via a gRNA-specific digital droplet polymerase chain reaction (ddPCR) assay. RESULTS: Simulations estimating genome editing residual clearance via dilution for extracellular residuals (via media changes) or intracellular residuals (via cell doubling) demonstrate clearance of measurable residuals within 28 days of cell culture. Studies simulating the stability of genome editing residuals estimate less than 7 days for the nuclease, gRNA and ribonucleoprotein (RNP) complex. gRNA was undetectable by 8 days post-engineering under actual engineering conditions. Additionally, without gRNA present, CRISPR Cas12a nucleases did not demonstrate evidence of cutting. In contrast, plasmid DNA can be randomly integrated into the genome and free plasmid is highly stable under cell culture-like conditions (50+ days). Additionally, plasmid DNA integrated in cells will propagate during mitosis, leading to the additional risk of expansion of an unintentional integration event. CONCLUSIONS: Both the gRNA and nuclease in the RNP complex are required for DNA cutting. Neither individual component nor the complex are stable beyond 7 days in culture-mimicking conditions. These findings suggest that the risk of unplanned genomic modification resulting from residual gRNA or nuclease is minimal for processes in which extensive culture is performed after the completion of genome editing and clonal selection. However, the risk of residual plasmid DNA integration is significantly higher regardless of the manufacturing process. The residual plasmid itself is quite stable (at least 50 days) and the risk of random, off-target integration is present. By establishing the stability of these components, we have demonstrated that testing for residual gRNA or nuclease is not warranted for clonally derived allogeneic cell therapies.

A risk-based approach for cell line development, manufacturing and characterization of genetically engineered, induced pluripotent stem cell-derived allogeneic cell therapies.

Advances in cellular reprogramming and gene-editing approaches have opened up the potential for a new class of ex vivo cell therapies based on genetically engineered, induced pluripotent stem cell (iPSC)-derived allogeneic cells. While these new therapies share some similarities with their primary cell-derived autologous and allogeneic cell therapy predecessors, key differences exist in the processes used for generating genetically engineered, iPSC-derived allogeneic therapies. Specifically, in iPSC-derived allogeneic therapies, donor selection and gene-editing are performed once over the lifetime of the product as opposed to as part of the manufacturing of each product batch. The introduction of a well-characterized, fully modified, clonally derived master cell bank reduces risks that have been inherent to primary-cell derived autologous and allogeneic therapies. Current regulatory guidance, which was largely developed based on the learnings gained from earlier generation therapies, leaves open questions around considerations for donor eligibility, starting materials and critical components, cell banking and genetic stability. Here, a risk-based approach is proposed to address these considerations, while regulatory guidance continues to evolve.

Development of an induced pluripotent stem cell-specific microRNA assay for detection of residual undifferentiated cells in natural killer cell therapy products.

Most clinically evaluated chimeric antigen receptor (CAR)-based cell therapies are generated from autologous immune cells. However, there are several limitations to autologous cell therapy, including low availability, poor quality of starting cellular material and limited expansion capability. Recently, induced pluripotent stem cell (iPSC)-derived allogeneic cell therapy platforms have gained popularity, as they seek to overcome many of the challenges inherent to current autologous cell therapies. However, teratoma risk associated with residual undifferentiated cells (i.e., iPSCs) in the drug product may restrict potential clinical applications if left unaddressed. To ensure the safety of the final cell therapy product, there is a need to develop quality control assays to detect residual iPSCs. Combining microRNA (miRNA) sequencing data with publicly archived miRNA microarray datasets, we demonstrated that miRNAs belonging to the 300 family (miR-302a-5p, miR-302c-3p and miR-302d-5p) and 500 family (miR-518f-5p and miR-519-3p) were highly expressed in iPSCs (both periperal blood mononuclear cell- and T cell-derived iPSCs) compared with a number of differentiated cell types. We developed and validated a sensitive digital droplet polymerase chain reaction (ddPCR) assay targeting these miRNAs to detect low levels of residual iPSCs in differentiated cell samples. The miRNA ddPCR-based method with primers for miR-302a-5p, miR-302c-3p and miR-302d-5p detected as few as 5, 3 and 10 undifferentiated iPSCs, respectively, in the background of 106 iPSC-derived natural killer (iNK) cells. These results suggest that our method targeting identified iPSC-specific miRNA transcripts is specific and sensitive for the quality assessment of NK cell therapy products derived from iPSCs.

Water structure in vitro and within Saccharomyces cerevisiae yeast cells under conditions of heat shock.

The OH stretch mode from water and organic hydroxyl groups have strong infrared absorption, the position of the band going to lower frequency with increased H-bonding. This band was used to study water in trehalose and glycerol solutions and in genetically modified yeast cells containing varying amounts of trehalose. Concentration-dependent changes in water structure induced by trehalose and glycerol in solution were detected, consistent with an increase of lower-energy H-bonds and interactions at the expense of higher-energy interactions. This result suggests that these molecules disrupt the water H-bond network in such a way as to strengthen molecule-water interactions while perturbing water-water interactions. The molecule-induced changes in the water H-bond network seen in solution do not translate to observable differences in yeast cells that are trehalose-deficient and trehalose-rich. Although comparison of yeast with low and high trehalose showed no observable effect on intracellular water structure, the structure of water in cells is different from that in bulk water. Cellular water exhibits a larger preference for lower-energy H-bonds or interactions over higher-energy interactions relative to that shown in bulk water. This effect is likely the result of the high concentration of biological molecules present in the cell. The ability of water to interact directly with polar groups on biological molecules may cause the preference seen for lower-energy interactions.

Mirror image forms of snow flea antifreeze protein prepared by total chemical synthesis have identical antifreeze activities.

The recently discovered glycine-rich snow flea antifreeze protein (sfAFP) has no sequence homology with any known proteins. No experimental structure has been reported for this interesting protein molecule. Here we report the total chemical synthesis of the mirror image forms of sfAFP (i.e., L-sfAFP, the native protein, and D-sfAFP, the native protein's enantiomer). The predicted 81 amino acid residue polypeptide chain of sfAFP contains Cys residues at positions 1, 13, 28, and 43 and was prepared from four synthetic peptide segments by sequential native chemical ligation. After purification, the full-length synthetic polypeptide was folded at 4 degrees C to form the sfAFP protein containing two disulfides. Chemically synthesized sfAFP had the expected antifreeze activity in an ice recrystallization inhibition assay. Mirror image D-sfAFP protein was prepared by the same synthetic strategy, using peptide segments made from d-amino acids, and had an identical but opposite-sign CD spectrum. As expected, D-sfAFP displays the same antifreeze properties as L-sfAFP, because ice presents an achiral surface for sfAFP binding. Facile synthetic access to sfAFP will enable determination of its molecular structure and systematic elucidation of the molecular basis of the antifreeze properties of this unique protein.

X-ray structure of snow flea antifreeze protein determined by racemic crystallization of synthetic protein enantiomers.

Chemical protein synthesis and racemic protein crystallization were used to determine the X-ray structure of the snow flea antifreeze protein (sfAFP). Crystal formation from a racemic solution containing equal amounts of the chemically synthesized proteins d-sfAFP and l-sfAFP occurred much more readily than for l-sfAFP alone. More facile crystal formation also occurred from a quasi-racemic mixture of d-sfAFP and l-Se-sfAFP, a chemical protein analogue that contains an additional -SeCH2- moiety at one residue and thus differs slightly from the true enantiomer. Multiple wavelength anomalous dispersion (MAD) phasing from quasi-racemate crystals was then used to determine the X-ray structure of the sfAFP protein molecule. The resulting model was used to solve by molecular replacement the X-ray structure of l-sfAFP to a resolution of 0.98 A. The l-sfAFP molecule is made up of six antiparallel left-handed PPII helixes, stacked in two sets of three, to form a compact brick-like structure with one hydrophilic face and one hydrophobic face. This is a novel experimental protein structure and closely resembles a structural model proposed for sfAFP. These results illustrate the utility of total chemical synthesis combined with racemic crystallization and X-ray crystallography for determining the unknown structure of a protein.

Temperature excursion infrared (TEIR) spectroscopy used to study hydrogen bonding between water and biomolecules.

Water is a highly polar molecule that is capable of making four H-bonding linkages. Stability and specificity of folding of water-soluble protein macromolecules are determined by the interplay between water and functional groups of the protein. Yet, under some conditions, water can be replaced with sugar or other polar protic molecules with retention of protein structure. Infrared (IR) spectroscopy allows one to probe groups on the protein that interact with solvent, whether the solvent is water, sugar or glycerol. The basis of the measurement is that IR spectral lines of functional groups involved in H-bonding show characteristic spectral shifts with temperature excursion, reflecting the dipolar nature of the group and its ability to H-bond. For groups involved in H-bonding to water, the stretching mode absorption bands shift to lower frequency, whereas bending mode absorption bands shift to higher frequency as temperature decreases. The results indicate increasing H-bonding and decreasing entropy occurring as a function of temperature, even at cryogenic temperatures. The frequencies of the amide group modes are temperature dependent, showing that as temperature decreases, the amide group H-bonds to water strengthen. These results are relevant to protein stability as a function of temperature. The influence of solvent relaxation is demonstrated for tryptophan fluorescence over the same temperature range where the solvent was examined by infrared spectroscopy.

Hydrogen bonding and the cryoprotective properties of glycerol/water mixtures.

Molecular dynamics simulations and infrared spectroscopy were used to determine the hydrogen bond patterns of glycerol and its mixtures with water. The ability of glycerol/water mixtures to inhibit ice crystallization is linked to the concentration of glycerol and the hydrogen bonding patterns formed by these solutions. At low glycerol concentrations, sufficient amounts of bulk-like water exist, and at low temperature, these solutions demonstrate crystallization. As the glycerol concentration is increased, the bulk-like water pool is eventually depleted. Water in the first hydration shell becomes concentrated around the polar groups of glycerol, and the alkyl groups of glycerol self-associate. Glycerol-glycerol hydrogen bonds become the dominant interaction in the first hydration shell, and the percolation nature of the water network is disturbed. At glycerol concentrations beyond this point, glycerol/water mixtures remain glassy at low temperatures and the glycerol-water hydrogen bond favors a more linear arrangement. High glycerol concentration mixtures mimic the strong hydrogen bonding pattern seen in ice, yet crystallization does not occur. Hydrogen bond patterns are discussed in terms of hydrogen bond angle distributions and average hydrogen bond number. Shift in infrared frequency of related stretch and bend modes is also reviewed.

Carbohydrate intramolecular hydrogen bonding cooperativity and its effect on water structure.

Molecular dynamics (MD) simulations combined with water-water H-bond angle analysis and calculation of solvent accessible surface area and approximate free energy of solvation were used to determine the influence of hydroxyl orientation on solute hydration and surrounding water structure for a group of chemically identical solutes-the aldohexopyranose sugars. Intramolecular hydrogen bond cooperativity was closely associated with changes in water structure surrounding the aldohexopyranose stereoisomers. The OH-4 group played a pivotal role in hydration as it was able to participate in a number of hydrogen bond networks utilizing the OH-6 group. Networks that terminated within the molecule (OH-4 --> OH-6 --> O-5) had relatively more nonpolar-like hydration than those that ended in a free hydroxyl group (OH-6 --> OH-4 --> OH-3). The OH-2 group modulated the strength of OH-4 networks through syndiaxial OH-2/4 intramolecular hydrogen bonding, which stabilized and induced directionality in the network. Other syndiaxial interactions, such as the one between OH-1 and OH-3, only indirectly affected water structure. Water structure surrounding hydrogen bond networks is discussed in terms of water-water hydrogen bond populations. The impact of syndiaxial versus vicinal hydrogen bonds is also reviewed. The results suggest that biological events such as protein-carbohydrate recognition and cryoprotection by carbohydrates may be driven by intramolecular hydrogen bond cooperativity.

Tryptophan interactions with glycerol/water and trehalose/sucrose cryosolvents: infrared and fluorescence spectroscopy and ab initio calculations.

In order to correlate how the solvent affects emission properties of tryptophan, the fluorescence and phosphorescence emission spectra of tryptophan and indole model compounds were compared for solid sugar glass (trehalose/sucrose) matrix and glycerol/water solution and under the same conditions, these matrices were examined by infrared spectroscopy. Temperature was varied from 290 to 12 K. In sugar glass, the fluorescence and phosphorescence emission spectra are constant over this temperature range and the fluorescence remains red shifted; these results are consistent with the static interaction of OH groups with tryptophan in the sugar glass. In sugar glass containing water, the water retains mobility over the entire temperature range as indicated by the HOH infrared bending frequency. The fluorescence of tryptophan in glycerol/water shifts to the blue as temperature decreases and the frequency change of the absorption of the HOH bend mode is larger than in the sugar glass. These results suggest rearrangement of glycerol and water molecules over the entire temperature change. Shifts in the fluorescence emission maximum of indole and tryptophan were relatively larger than shifts for the phosphorescence emission-as expected for the relatively smaller excited triplet state dipole for tryptophan. The fluorescence emission of tryptophan in glycerol/water at low temperature has maxima at 312, 313, and 316 nm at pH 1.4, 7.0, and 10.6, respectively. The spectral shifts are interpreted to be an indication of a charge, or Stark phenomena, effect on the excited state molecule, as supported by ab initio calculations. To check whether the amino acid remains charged over the temperature range, the infrared spectrum of alanine was monitored over the entire range of temperature. The ratio of infrared absorption characteristic of carboxylate/carbonyl was constant in glycerol/water and sugar glass, which indicates that the charge was retained. Tryptophan buried in proteins, namely calcium parvalbumin from cod and aldolase from rabbit, showed temperature profiles of the fluorescence spectra that were largely independent of the solvent (glycerol/water or sugar glass) and temperature whereas the fluorescence and phosphorescence yields were dependent. The results demonstrate how the rich information found in tryptophan luminescence can provide information on the dipolar nature and dynamics of the matrix.

Infrared spectroscopy used to evaluate glycosylation of proteins.

Infrared (IR) spectroscopy is used for studying the carbohydrate moieties of glycosylated proteins. IR spectra of mono- and disaccharides in the fingerprint region are specific to each sugar and to the environment of the sugar molecules (i.e., aqueous solution or anhydrous glass phase). The IR spectra of glycosylated proteins (mucin, soybean peroxidase, collagen IV, and avidin) were compared with those of the constituent sugars and cytochrome c (a protein with no glycosylation). Our results demonstrate that the IR absorption spectra of glycosylated proteins show distinct absorption bands for the sugar moiety, the protein amide group, and water. Therefore, IR can be used to detect glycosylation.