GMP cryopreservation of large volumes of cells for regenerative medicine: active control of the freezing process

Abstract

Cryopreservation protocols are increasingly required in regenerative medicine applications but must deliver functional products at clinical scale and comply with Good Manufacturing Process (GMP). While GMP cryopreservation is achievable on a small scale using a Stirling cryocooler-based controlled rate freezer (CRF) (EF600), successful large-scale GMP cryopreservation is more challenging due to heat transfer issues and control of ice nucleation, both complex events that impact success. We have developed a large-scale cryocooler-based CRF (VIA Freeze) that can process larger volumes and have evaluated it using alginate-encapsulated liver cell (HepG2) spheroids (ELS). It is anticipated that ELS will comprise the cellular component of a bioartificial liver and will be required in volumes of ∼2 L for clinical use. Sample temperatures and Stirling cryocooler power consumption was recorded throughout cooling runs for both small (500 μL) and large (200 mL) volume samples. ELS recoveries were assessed using viability (FDA/PI staining with image analysis), cell number (nuclei count), and function (protein secretion), along with cryoscanning electron microscopy and freeze substitution techniques to identify possible injury mechanisms. Slow cooling profiles were successfully applied to samples in both the EF600 and the VIA Freeze, and a number of cooling and warming profiles were evaluated. An optimized cooling protocol with a nonlinear cooling profile from ice nucleation to -60°C was implemented in both the EF600 and VIA Freeze. In the VIA Freeze the nucleation of ice is detected by the control software, allowing both noninvasive detection of the nucleation event for quality control purposes and the potential to modify the cooling profile following ice nucleation in an active manner. When processing 200 mL of ELS in the VIA Freeze-viabilities at 93.4% ± 7.4%, viable cell numbers at 14.3 ± 1.7 million nuclei/mL alginate, and protein secretion at 10.5 ± 1.7 μg/mL/24 h were obtained which, compared well with control ELS (viability -98.1% ± 0.9%; viable cell numbers -18.3 ± 1.0 million nuclei/mL alginate; and protein secretion -18.7 ± 1.8 μg/mL/24 h). Large volume GMP cryopreservation of ELS is possible with good functional recovery using the VIA Freeze and may also be applied to other regenerative medicine applications.

FIG. 1.

The VIA freeze. The VIA Freeze and controller on the bench top (A), and (B) with lid open showing sample plate and cryovial holder for quality assurance samples (on left-hand plate edge).

FIG. 2.

Impact of cooling profiles used for feasibility experiments on small volume cryopreservation of encapsulated cell spheroids (ELS). Five different cooling profiles (A) were used to cool ELS in 1.8 mL cryovials to −100°C. These profiles are fully described in the Methods section of this article. The effect of differing ice nucleation temperatures on ELS viability using these profiles in post-thaw cultures was also investigated (B). Cells were cooled by a “multi-step cooling” method (●),^, linear 1°C/min (□), linear 1°C/min with an isothermal hold at −8°C (▲), a controlled concentration profile (◊), or a constant change in ice fraction with time (■). These data demonstrate that regardless of cooling profile utilized, ELS viability in post-thaw cultures is strongly correlated with ice nucleation temperature. When ice nucleation occurs at higher temperatures (i.e., when cholesterol is present and there is only minimal supercooling), ELS viability in post-thaw cultures is typically high; conversely, if ice nucleation spontaneously occurs at lower temperatures, supercooling will result in poor ELS viability in post-thaw cultures.

FIG. 3.

Cooling profiles used for direct comparison of ELS recovery in small and large volumes. Three different cooling profiles were used to cool ELS in 1.8 mL cryovials using the EF600 and in 200 mL cryobags using the VIA Freeze to −100°C in parallel. A simple linear rate (A) was compared with nonlinear profiles (B, C). In large volumes, heat transfer occurs more slowly as a result of the increased mass. These nonlinear profiles were utilized as these are a more accurate representation of cooling profiles that would be experienced by ELS cooled in large volumes, where heat transfer would be altered compared with ELS cooled in small volumes. They were derived using mathematical modeling.

FIG. 4.

Cryoscanning electron microscopy of fractured cryovials after cooling at a linear rate of 1°C/min. Ultrastructure resulting from controlled ice nucleation (A, C, E) is compared to spontaneous ice nucleation (B, D, F). In all samples the spaces originally occupied by ice crystals are revealed as voids following sublimation of ice. In figures (B, D) sectioned alginate beads that are ∼500 μm in diameter are outlined. In (C, E, F), cell clusters (c), ice voids (i), and smooth freeze concentrated alginate and cryoprotectant (CPA) (*) have been labeled. Scale bars on (A, B) 1 mm, on (C, D) 200 μm, and on (E, F) 20 μm.

FIG. 5.

Freeze substitution of fractured cryovials after cooling at a linear rate of 1°C/min. Ultrastructure resulting from controlled ice nucleation (A, C) is compared with spontaneous ice nucleation (B, D). In spontaneously nucleated samples, the spaces originally occupied by ice crystals are revealed as voids following sublimation of ice, and can be seen throughout the ELS structure. These ice voids are absent within cholesterol-nucleated ELS, although there is evidence of shrinkage spaces. At higher magnification, it is difficult to distinguish cell organelles in cholesterol-nucleated ELS due to the extreme cell dehydration during cooling. Conversely, organelles can be seen in spontaneously nucleated ELS. Cellular material (c), ice voids (i), cell membrane (cm), extracellular matrix (ecm), mitochondria (m) shrinkage spaces (s), and alginate (*) have been labeled. Scale bars on (A, B) 2 μm and on (C, D) 500 nm.

FIG. 6.

Process parameters and measured temperatures within a CryoMACS bag during linear cooling on the VIA Freeze. (A) The entire cooling run over 1 h 45 min. Set point temperature (solid black line). Measured temperatures within the cryobag containing 200 mL ELS, at the bottom of the bag (dark gray line), in the middle (mid gray line) near to the top (light gray line). Power input (% of maximum electrical power) supplied to the VIA Freeze (dashed black line on right hand Y axis). (B) Detail around ice nucleation in the bags; note expanded X and Y axes: set point temperature (solid black line); measured sample plate temperature (dotted line); power input (% of maximum electrical power) supplied to the VIA Freeze (dashed line on right hand Y axis). Note the plate temperature shows a small inflection in response to the latent heat of ice formation.