Extracellular Antifreeze Protein Significantly Enhances the Cryopreservation of Cell Monolayers

The cryopreservation of cells underpins many areas of biotechnology, healthcare, and fundamental science by enabling the banking and distribution of cells. Cryoprotectants are essential to prevent cold-induced damage. Here, we demonstrate that extracellular localization of antifreeze proteins can significantly enhance post-thaw recovery of mammalian cell monolayers cryopreserved using dimethyl sulfoxide, whereas they show less benefit in suspension cryopreservation. A type III antifreeze protein (AFPIII) was used as the macromolecular ice recrystallization inhibitor and its intra/extracellular locations were controlled by using Pep-1, a cell-penetrating peptide. Flow cytometry and confocal microscopy confirmed successful delivery of AFPIII. The presence of extracellular AFPIII dramatically increased post-thaw recovery in a challenging 2-D cell monolayer system using just 0.8 mg·mL–1, from 25% to over 60%, whereas intracellularly delivered AFPIII showed less benefit. Interestingly, the antifreeze protein was less effective when used in suspension cryopreservation of the same cells, suggesting that the cryopreservation format is also crucial. These observations show that, in the discovery of macromolecular cryoprotectants, intracellular delivery of ice recrystallization inhibitors may not be a significant requirement under “slow freezing” conditions, which will help guide the design of new biomaterials, in particular, for cell storage.


■ INTRODUCTION
Cell-based therapies are emerging as next-generation treatments for intractable and complex diseases (especially in oncology) which remain unresponsive to traditional "molecular" therapies. 1 However, mammalian cell cryopreservation, for the long-term storage of cells and tissues, remains an essential part of the manufacturing process and has been identified as a potential bottleneck in the future development of complex cellular therapy products. 2 Dimethyl sulfoxide (DMSO), a cell permeable cryoprotectant, remains the most widely used cryoprotective agent (CPA) for the cryopreservation of mammalian cells and tissues in cell suspension. 3 During typical "slow freezing" approaches, DMSO enters cells and minimizes injury through reducing electrolyte concentration in residual unfrozen solution within and surrounding cells at any given temperature, thus reducing intracellular ice growth, cell shrinkage and osmotic shock during freezing. 4,5 However, cell survival rates decrease due to DMSO cytotoxicity and inhibition of internal signaling. 6,7 Furthermore, prolonged cryopreservation of stem cells approved for the treatment of various blood and immunological diseases and for large-scale banking and manufacturing can result in differentiation induced by histone alterations and DNA methylation, creating difficulties in the cryopreservation of material routinely used in clinical applications. 8−11 Cryopreservation of cells in monolayer format is currently being investigated as a means to supply cells which can be readily used and do not experience extensive "phenotypic drift" due to time-consuming laboratory processes, such as inoculation and propagation, from frozen vials. Successful monolayer cryopreservation of cells would be revolutionary in minimizing batch-to-batch variation and for the development of 2-and 3-D cell models, tissue storage, viral diagnostics, and organ-on-a-chip applications. 12−15 However, the current standard (DMSO) approaches used in suspension freezing are not translatable to monolayer freezing, with evidence suggesting that cells within a 2-and 3-D network (monolayers and spheroids/organoids) experience different modes of cryoinjury. 16,17 Thus, development of novel cryoprotectants tailored toward the format of cryopreservation and to replace or reduce DMSO content is pivotal for the future development of cellbased therapies and diagnostics.
Naturally occurring CPAs, including trehalose, 18 proline, 17 sucrose, 19 and antifreeze proteins (AFPs), 20,21 as well as synthetic cryoprotectants, 22−24 have been studied in the attempt to replace or improve DMSO cell suspension cryopreservation. In particular, AFPs (and their mimics) have received attention due to their potent ice recrystallization inhibition (IRI) properties and potential ability to stabilize membranes or modify ice nucleation. 25−28 Ice recrystallization (growth) during thawing results in the formation of large ice crystals, at the expense of small crystals, causing cellular damage and is thus a major contributor to cell death. Although IRI active compounds have provided some benefits for cryopreserving erythrocytes, 29−33 nucleated cell lines, 34,35 and stem cells, 36 complete removal of DMSO is rarely achieved. Furthermore, the influence of IRI active compounds on cells frozen in a monolayer format is poorly understood, as the mechanism of cryoinjury is different. Cell death during monolayer freezing is postulated to be caused by the propagation of intracellular ice between adjacent cells initiated by multiple mechanisms including surface-catalyzed nucleations (i.e., extracellular ice interacting with cell membrane forming a nucleation site for intracellular ice growth), 37,38 cell−cell and cell−surface interaction with adjacent cells, 39−41 or gap junctions within the membrane. 16,42−44 Controlled slow freezing of suspension cells with DMSO (typically 5−10 wt %) removes the risk of intracellular ice formation as the CPA enters the cells, causing more water to efflux, and reducing the likelihood of intracellular ice formation. 45−47 However, in the extracellular media, ice formation and growth does occur, hence the potential benefits of ice recrystallization inhibitors. Conversely, monolayer freezing of cells using DMSO typically results in low post-thaw viability due to the intercellular contacts and limited diffusion between cells suggesting that monolayer freezing may benefit from intracellular IRI active macromolecules. 17,48 Therefore, raising the question, "does the difference in ice growth during slow-freezing of cells in suspension and monolayer formats affect the benefits of intraversus extracellular localization of IRI active macromolecular cryoprotectant?" Bailey et al., 48 Graham et al., 17 and Matsumura et al. 49 have previously shown that macromolecular cryoprotectants can have dramatic effects on cell monolayer cryopreservation. However, determining the ideal cellular location for macromolecular cryoprotectants would aid in the understanding and development of emerging cryoprotectants, especially since intracellular delivery of synthetic and natural macromolecules can be challenging. 50 Considering the above, the aim of this study was to evaluate whether intra-or extracellular localization of ice recrystallization inhibition active macromolecules is advantageous to the cryopreservation of cells in suspension, monolayer, or both. Using a type III AFP (AFPIII) and an efficient cell-penetrating peptide, Pep-1, we delivered AFPIII inside or outside of cells. Extracellular AFPIII lead to a dramatic enhancement in the cryopreservation of a 2-D cell monolayer system, resulting in a 3-fold increase in the total number of cells recovered compared to 10 wt % DMSO alone. In contrast, intracellular AFPIII only protected cells at a lower DMSO concentration, although with high variability, supporting the hypothesis that extracellular ice growth is a major cause of cell death. AFPIII failed to improve cell recoveries in suspension freezing regardless of localization. These results will help the rational design of more potent macromolecular cryopreservatives. Physical and Analytical Methods. Ice Recrystallization Inhibition Assay (IRI). IRI was assessed using a Linkam Biological Cryostage BCS196 controlled with a T96-Linpad system controller equipped with a LNP95-Liquid nitrogen cooling pump (Linkam Scientific Instruments UK, Surrey, UK). Images were acquired using an Olympus CX41 microscope UIS-2 20×/0.45/∞/0-2/FN22 lens (Olympus Ltd., Southend on sea, UK) and a Canon EOS 500D SLR digital camera. Image processing and mean grain size measurements were completed using ImageJ 1.49v.
Flow Cytometry. Flow cytometry was performed on a BD Influx cell sorter (BD Biosciences) equipped with 355-, 488-, 561-, and 642 nm lasers and detecting up to 24 parameters (21 fluorescence channels, two forward scatter channels and one side scatter). BD FACS Sortware software was used to operate the instrument. Sample analysis required the use of the 488 nm excitation laser and 530/30 nm filter. A 100 μm nozzle was fitted, operating at a pressure of 20 psi (sheath) and 21.5 psi (sample). Stream and laser alignment was performed using BD Sphero Rainbow Calibration Particles (8 Peaks, 3.0−3.4 μm), and all sample measurements consisted of a minimum of 100 000 recorded events. Cell detachment was completed utilizing Accutase solution (Fisher Scientific) and passed through Fisherbrand Sterile Cell Strainers (Fisher Scientific) to ensure single cell analysis. Voltage settings applied ensured that nonfluorescent control cells appeared at low fluorescence emission intensities. FlowJo X 10.0.7r2 (Tree Star, Ashland, USA) was used for all statistical analysis and plotting of flow cytometry data.
Statistical Analysis. Data was analyzed with a one-way analysis of variance (ANOVA) on ranks followed by a comparison of experimental groups with the appropriate control group (Tukey's post hoc test). R (R Foundation for Statistical Computing, Vienna, Austria) was used for all statistical analysis.
AFPIII Cytotoxicity. A549 cells were plated in a 96-well plate (Corning Incorporated, Corning, NY) at a density of 2 × 10 5 cells· mL −1 with media supplemented with varying concentrations of AFPIII (0−0.8 mg·mL −1 ) and incubated in a humidified atmosphere of 95% air and 5% CO 2 at 37°C for 24 h. Alamar blue reagent (10% v/v in cell media) was added to both AFPIII treated and untreated (control) cells. Absorbance measurements were obtained at 570 and 600 nm using a BioTek Synergy HT microplate reader to monitor the reduction of resazurin to resorufin by viable cells. Cells were incubated for 4 h at 37°C and 5% CO 2 with readings obtained every 30 min/1 h. Total cell viability was reported relative to control cells grown solely in cell culture media alone.
Optimization of AFPIII Concentration for Vial Freezing. A549 cells were obtained from an adherent culture by detaching with 0.25% trypsin and 1 mM EDTA in balanced salt solution, for 10 min, in a humidified atmosphere of 5% CO 2 and 95% air, at 37°C. Viable cells were counted using a hemocytometer (Sigma-Aldrich), following treatment with 0.4% trypan blue solution (Sigma-Aldrich), and placed in cryogenic vials at a density of 200k cell·mL −1 (400k cells per vial) Biomacromolecules Article in F12-K media supplemented with 10% FBS, varying concentrations of AFPIII (0−0.8 mg·mL −1 ), and DMSO (0−10 wt %). Cryogenic vials were placed in a CoolCell LX freezing container (BioCision, LLC, Larkspur, CA), and frozen at 1°C·min −1 to −80°C. Vials were subsequently stored in the vapor phase of a liquid nitrogen dewar (−196°C) for 24 h. Cells were rapidly thawed at 37°C, transferred to warm complete FK-12 media (9 mL), and centrifuged at 2g for 5 min. The cell pellet was resuspended in complete media, transferred into a 24-well plates (Corning Incorporated, Corning, NY), and incubated in a humidified atmosphere of 5% CO 2 and 95% air at 37°C for 24 h. Finally, cell viability was assessed by counting the number of viable cells (i.e., with intact membrane), as described above, and dividing by the total number of cells initially frozen. Four biological repeats were conducted.
Optimization of AFPIII Transfection Using Pep-1. A549 cells were plated in a 24-well plate at a density of 800k cell·mL −1 (400k cell per well), to achieve a confluent monolayer, and left to attach for 24 h in a humidified atmosphere of 5% CO 2 and 95% air at 37°C. Pep-1 (1 mg·mL −1 ), dissolved in ultrapure H 2 O, and AFPIII-Alexa Fluor 488 (0.25−2 mg·mL −1 ), dissolved in PBS, were mixed in a 1:1 ratio and incubated at room temperature for 30 min. Cells were incubated with the AFPIII-Alexa Fluor 488/Pep-1 mixtures (100 μL) for 30 min in a humidified atmosphere and for a further 1 h after adding F12-K media without an additional supplement (100 μL). Complete F12-K media (100 μL) was added, and cells were further incubated for 1 h. Cells were imaged using the Olympus CX41 microscope described above (Supporting Information). The procedure was repeated with the optimized final concentration of AFPIII-Alexa Fluor 488 (0.8 mg· mL −1 ) and Pep-1 (0.5 mg·mL −1 ). Fluorescence quantification was achieved using flow cytometry, and confocal images were obtained following fixation with 4% paraformaldehyde (15 min) and staining with NucBlue Live Cell ReadyProbes. Cells untreated and treated solely with AFPIII-Alexa Fluor 488 (0.8 mg·mL −1 ) or Pep-1 (0.5 mg· mL −1 ) in PBS:water (1:1) were also analyzed to determine nonspecific binding, Pep-1 fluorescence, and cell autofluorescence.

■ RESULTS AND DISCUSSION
To evaluate the intracellular versus extracellular impact of macromolecular cryoprotectants, a type III antifreeze protein from ocean pout (AFPIII) was selected. The rationale for using an antifreeze protein, as opposed to polymeric cryoprotectants, was their high IRI activity (capable of inhibiting all ice growth well below 0.1 mg·mL −1 ) 23 and the ability to use established methods (see below) to enable cellular delivery of proteins. Recombinant AFPIII was expressed in Escherichia coli (E. coli) and purified according to previously published methods (Supporting Information). 51 Site-directed mutagenesis of the terminal Ala present within the AFPIII encoding plasmid was used to produce a mutant cysteine variant of AFPIII (AFP-Cys), enabling site-specific conjugation to Alexa Fluor 488 Maleimide (AFPIII-Alexa Fluor 488) for fluorescence-based quantification and imaging (Supporting Information).
A549 (adenocarcinomic human alveolar basal epithelial) cells were chosen as a model cell line for cryopreservation studies. To assess the role of intra-versus extracellular AFPIII during mammalian cell cryopreservation, translocation of AFPIII into the cells was essential. The cell-penetrating peptide Pep-1 was selected for this purpose, as it is known to be a broadly applicable cell-uptake enhancing peptide via a nonendocytic mechanism. 52−56 Due to the extensive literature available on Pep-1 mediated translocation of protein cargoes, the mechanism of complexation with AFPIII was not studied here. There is evidence to suggest that Pep-1 forms a complex with protein cargos through hydrophobic and electrostatic interactions and adopts a helical structure upon interaction with the plasma membrane lipids, allowing insertion into the lipid bilayer. The complex is then released into the cytoplasm with partial "decaging" of the cargo. However, the exact mechanism of complexation was irrelevant to the cryopreservation of mammalian cells and the overall aim of this study.
To determine the minimum concentration of AFPIII required for cellular delivery, a range of AFPIII-Alexa Fluor 488 concentrations (0.125−1 mg·mL −1 ) were incubated with a fixed concentration of Pep-1 (0.5 mg·mL −1 ) for 30 min, to form protein/peptide cargos, and subsequently applied to A549 monolayers (2.5 h). AFPIII-Alexa Fluor 488 uptake was monitored using fluorescence microscopy (Supporting Information, Figure S1). Internalization of AFPIII was observed at concentrations of 0.5 and 1 mg·mL −1 , so an intermediate concentration of 0.8 mg·mL −1 was selected for the intracellular delivery of AFPIII using Pep-1 in all subsequent experiments, with a 1:1.65 molar ratio of AFPIII to Pep-1. Higher Pep-1/ cargo ratios are often used (20:1 or 10:1); however, previous reports suggest that such high concentrations can result in the formation of aggregates that are poorly taken into the cell and accumulate on the cell surface or endosomes. 53,55 Therefore, to minimize these issues, while still ensuring AFPIII internalization, a 1:1.65 molar ratio was optimal. Further (quantitative) uptake studies were also explored (below). With a working range established, the cytotoxicity of AFPIII toward A549 cells was evaluated to ensure that any cryopreservative effects, either positive or negative, were not skewed by intrinsic cytotoxicity. A549 cells were incubated with AFPIII (0−0.8 mg·mL −1 ) for 24 h and cell viability was assessed using the resazurin reduction assay to determine any changes to cell metabolic activity. Over 88% cell viability was obtained, even at the highest concentrations, demonstrating minimal effect to cellular metabolic activity ( Figure 1) and an advantage over cytotoxic CPAs.
To further demonstrate the extent of intracellular incorporation of AFPIII with Pep-1, confocal images were obtained using the optimized molar ratio of AFPIII-Alexa Fluor 488 and Pep-1 (1:1.65) shown in Figure 2A. The peptide/protein cargo was allowed to complex for 30 min and then applied to confluent monolayers of A549 cells for 2.5 h, mimicking the suspension and monolayer pretreatment conditions used later in the freezing protocol. Confocal images clearly indicated enhanced uptake (green channel) when both Pep-1 (0.5 mg· mL −1 ) and AFPIII-Alexa Fluor 488 (0.8 mg·mL −1 ) were present, compared to AFPIII-Alexa Fluor 488 alone, suggesting that AFPIII internalization is minimal without the use of Pep-1. These results further validate our use of AFPIII for investigating intracellular versus extracellular impact of macromolecular cryoprotectants on mammalian cell cryopreservation. This was further corroborated using flow cytometry analysis, with Pep-1 leading to approximately 5 times more uptake of AFPIII ( Figure 2B). Confocal z-stack images revealed fluorescence localized throughout the cytoplasm, which was essential when assessing the importance of intraversus extracellular AFPIII (Supporting Information, Figures  S2−S5). In addition, there were no visible changes to A549 cell morphology or membrane integrity when transfecting AFPIII with Pep-1.
Following optimization of intracellular AFPIII delivery, determining the optimum concentration of extracellular AFPIII (in solution) was essential to maximize post-thaw cell recovery. A549 cells were cryopreserved in a suspension containing DMSO (0−10 wt %) and AFPIII (0−0.8 mg·mL −1 ). Cells were allowed to recover for 24 h post-thaw, before determining cell recovery (Figure 3). [Note: the 24 h incubation period post-thaw is vital to avoid false positive results associated with short post-thaw times, where any additive can give the impression of cryoprotection if apoptotic processes are not allowed time to complete]. 57 Cell recovery was determined using the trypan blue exclusion assay for cell viability and the total number of viable cells were compared to an unfrozen control. This method provides total cell recovery, as damaged cells are not counted, compared to methods that only measure the viability of the recovered cells, which typically ignores highly damaged cells and can overestimate cell recovery values by not taking into account the number of cells recovered postthaw (e.g., 50% of cells recovered, however 100% are viable). A549 cells were also frozen in equivalent DMSO solutions (0− 10 wt %) to provide a comparable standard freezing approach for the assessment of post-thaw cell recovery. Unsurprisingly, A549 cells frozen in 10 wt % DMSO alone outperformed lower concentrations for this cell line, with an average post-thaw recovery of 67%. These results are consistent with those reported by Bailey et al. 48 Addition of AFPIII at 0.8 mg·mL −1 to the media resulted in a dramatic loss in post-thaw cells at all DMSO concentrations. The decrease in cell viability witnessed  at high AFPIII concentrations is consistent with previous work, where AFPIII's ice shaping capabilities result in the formation of destructive "needle-shaped" ice which can pierce cell membranes and vital internal structures during freezing. 20,33,58,59 Lower concentrations of AFPIII (0.1 and 0.05 mg·mL −1 ) lead to almost identical results to DMSO alone, with slight differences in cell recovery being statistically insignificant. During freezing, AFPIII adsorbs onto prismatic and bipyramidal ice faces without attaching to basal faces, allowing the development of bipyramidal ice crystals, which "burst" along the c-axis when the freezing point is reached forming "needle-shaped" ice crystals. Vorontsov et al. 60 and Ishiguro and Rubinsky 61 revealed that lower AFPIII concentrations form dendritic ice growth whereas higher concentrations result in the development of cell damaging spicular ice crystals. Thus, to avoid cell injury and maximize recoveries post-thaw, 0.05 mg·mL −1 AFPIII was used as the external CPA concentration for further suspension cryopreservation studies.
As a final control, it was also necessary to confirm if the presence of Pep-1 had any impact on the ice recrystallization inhibition (IRI) activity of AFPIII, a crucial cryoprotective function of this protein, as any changes could give rise to false positive cryopreservation results. The "splat" assay was used to assess changes in IRI activity, whereby small ice crystals are nucleated in a wafer and annealed at −8°C for 30 min. 62,63 Following this, the mean grain size of the crystals was determined, and reported relative to a negative control (PBS). AFPIII demonstrated potent IRI activity, as shown in Figure 4, inhibiting all ice growth at concentrations below 0.01 mg· mL −1 . This value is consistent with previous reports, demonstrating superior and consistent IRI capabilities compared to other cryopreservants. 51 This recrystallization inhibition effect is maintained, if not slightly enhanced, in the presence of 0.5 mg·mL −1 Pep-1, which is the working concentration of the peptide used in this study. Pep-1 alone was found to present negligible IRI activity, eliminating the possibility of results being enhanced or depleted by Pep-1 interactions. These results indicate that Pep-1 translocation of AFPIII was an ideal choice to ensure AFPIII's IRI activity was retained, allowing a direct comparison between the benefits of localizing IRI active macromolecules intra-versus extracellularly.
Guided by the above data, the impact of intra/extracellular AFPIII on cryopreservation was evaluated. A549 cells were cryopreserved in suspension using 0−10 wt % DMSO alone, or with AFPIII either complexed with Pep-1 or AFPIII alone. Cell recovery was determined 24 h post-thaw using the trypan blue exclusion assay. As can be seen in Figure 5, Pep-1 alone lead to a small decrease in post-thaw recovery in 10 wt % DMSO. A resazurin reduction assay was completed of Pep-1, Figure S7, revealing that uncomplexed Pep-1 at the concentrations used to complex with AFPIII, results in a significant (P ≤ 0.05) decrease in cell viability. Therefore, this slight decrease in cell recovery can be attributed to uncomplexed Pep-1 cytotoxicity. The presence of AFPIII intracellularly, extracellularly, or both, surprisingly lead to minimal changes in post-thaw cell recovery. A small decrease in cell recovery was noted in the presence of intracellular AFPIII at all DMSO concentrations, but this was found to be statistically insignificant. There have been previous reports of intracellular ice recrystallization inhibitors enhancing the cryopreservation of red blood cells using glycerol (red blood cells are not stored in DMSO); 32 however, in that particular case, the conditions (low glycerol and transient cooling/warming) were deliberately chosen to promote intracellular ice growth. Here, by using slow freezing rates lethal intracellular ice growth should be prevented/reduced however the formation of extracellular ice growth remains an issue. Extracellular AFPIII only increased post-thaw cell viability by approximately 6% at equal DMSO concentrations between 2 and 10 wt %. This data supports the hypothesis that, for optimized suspension cryopreservation systems (such as this one), inhibiting ice recrystallization does not significantly increase viability. The highest percentage increase occurred

Biomacromolecules
Article when 2 wt % DMSO was used, with an increase from 13.1% to 21.0%, a nonideal concentration for these cells. Unfortunately, percentage recovery values still remained low despite the presence of AFPIII as the lack of DMSO most likely introduced additional mechanisms of cell death which IRI active compounds do not account for including exposure to excessively high concentrations of electrolytes ("solute effect") causing irreparable damage to the cell membrane and osmotic volume excursions exceeding cell tolerance. 4 However, extracellular AFPIII (or other IRIs) appeared to show application in the cryopreservation of cells requiring low DMSO storage, due to susceptibility to cytotoxicity or induced differentiation, and where ice recrystallization is a more significant contributor to cell death.
The above results suggested that the intra/extracellular localization of AFPIII was not sufficient for enhancing mammalian cell cryopreservation in suspension freezing. While this is the most standard cell cryopreservation format, there is a real need to develop technologies and methods for 2and 3-D cell models, for tissue storage and organ-on-a-chip applications. 13−15 The monolayer freezing protocol is illustrated in Figure 6A. A549 monolayers were frozen in DMSO alone, or with AFPIII (0.8 mg·mL −1 ), AFPIII (0.8 mg·mL −1 ) plus Pep-1 (0.5 mg·mL −1 )), or Pep-1 alone (0.5 mg·mL −1 ). Cells were incubated with these cryoprotectants for 2.5 h, except for DMSO alone where incubation with media was completed, followed by incubation with DMSO (0−10 wt %) concentrations for 10 min. This procedure allowed time for intracellular translocation of Pep-1 with AFPIII and maintained the same AFPIII, Pep-1, and DMSO exposure times under all relevant conditions. Excess solution was removed, and cell monolayers were frozen at a controlled freezing rate of 1°C·min −1 to −80°C using a CoolCell MP plate. Confocal images revealed that incubation with DMSO did not affect the cells' ability to retain transfected AFPIII (Supporting Information, Figures S2−S5).
As expected, cryopreservation in DMSO alone lead to limited cell recovery, with 25% cell recovery achieved using 10 wt % DMSO, substantially lower compared to suspension cell cryopreservation due to the intercellular contacts and limited diffusion between cells, minimizing DMSO's capabilities of reducing intracellular ice growth. Lower DMSO concentrations provided less recovery. Surprisingly, Pep-1 alone resulted in a significant decrease in cell recovery. As previously discussed, uncomplexed Pep-1 is cytotoxic at the concentrations used; however, resazurin reduction was still 80% relative to control cells, Figure S7. Thus, cytotoxicity cannot be the only contributing factor. Uncomplexed Pep-1 may be interacting with membrane proteins and thus weakening membrane integrity and making it more susceptible to ice damage. Intracellular AFPIII lead to an increase in cell recovery at all DMSO concentrations, however with high variability. Compared to all other treatments, intracellular AFPIII provided the highest increase in cell recoveries, from 0.3% to 10.5%, when 2 wt % DMSO (i.e., lower DMSO concentrations) was used. Poisson et al. 64 demonstrated that cell permeable small molecules, aryl glycosides, possessing IRI activity can reduce intracellular ice recrystallization content and increase post-thaw cell viability of red blood cells and human umbilical vein endothelial cells frozen with low glycerol. DMSO and glycerol both provide cryoprotection using the same mechanism of action, by reducing the  Cell recovery was determined by trypan blue exclusion assay relative to unfrozen controls. The data represents the mean value ± SEM (n = 3; n.s. p > 0.05, *p ≤ 0.05, **p ≤ 0.01 and ***p ≤ 0.001 compared to percentage cell recovery values at equivalent DMSO concentrations with no AFPIII).

Biomacromolecules
Article concentration of either CPAs the risk of intracellular ice growth increases. This suggests that intracellular AFPIII is likely reducing intracellular ice recrystallization, thus reducing the chances of larger crystals forming during freeze−thaw and cell death. Intracellular ice is particularly problematic in monolayer freezing due to intracellular ice propagation from cell−cell and cell−surface interaction with adjacent cells, or gap junctions within the membrane. 41,43,44 Therefore, intracellular localization of IRI active compounds, such as AFPIII, may find use in situations where lower DMSO concentrations are beneficial.
Extracellular AFPIII gave rise to significant increases in postthaw recovery for all DMSO concentrations, especially at higher DMSO concentration (5−10 wt %). In fact, 10 wt % DMSO increased the recovery rate from 25% to 60% upon addition of AFPIII. Extracellular AFPIII combined with 5 wt % DMSO also demonstrated a large increase in cell recovery (1.6fold), compared to the equivalent DMSO alone cell recoveries, and demonstrated the highest statistically significant increase compared to all other combination of treatments. In addition, extracellular AFPIII at 5 wt % DMSO also obtained cell recoveries (38%) greater than 10 wt % DMSO alone, demonstrating that external AFPIII (in solution) could be used to reduce the total DMSO concentration required (which is biomedically important) while enhancing post-thaw cell recovery. Acker et al. demonstrated that, although intracellular ice formation is regarded as an inherently lethal event, innocuous intracellular ice exists and may even increase postthaw cell viability of multiple cell lines using slow freezing. 65,66 Therefore, it is possible that extracellular ice growth recrystallization is the dominant cause of cell death and IRI active compounds are hence more potent when extracellular. Alternatively, nonspecific uptake of AFPIII could inhibit intracellular ice recrystallization sufficiently to avoid crystal sizes which have been deemed harmful in previous studies. 67 AFPs are known to stabilize fatty acids located throughout the membrane during freezing and can cause membrane leakage, however this is highly dependent on lipid composition of the bilayer and thus individual cell types. 68−70 Membrane leakage would allow the efflux of water, and most likely exchange with DMSO, to occur rapidly and prevent extensive intracellular ice growth. Regardless of the exact mechanism, these results confirm that, at least when DMSO is used on cell monolayers, that extracellular localization of antifreeze proteins provide significant benefit to monolayer cryopreservation of cells which, according to previous literature, is likely correlated to their ice recrystallization inhibition activity or membrane interactions. These effects were clearly less important for suspension cell cryopreservation suggesting that the primary mechanism of cell death are different in the two systems, and the cryoprotectants required will vary depending on the complexity of the cellular system being cryopreserved.

■ CONCLUSIONS
Here we present a study into the role of intra-versus extracellular antifreeze proteins, on the cryopreservation of a model cell line, showing how this localization has a dramatic effect on post-thaw cell recovery. The Pep-1 cell-penetrating peptide was employed to deliver recombinantly expressed type III antifreeze protein into A549 cells. Intracellular delivery was confirmed using confocal microscopy and flow cytometry, showing 5 times more internalization of AFPIII. Using conventional suspension cell freezing methods, with DMSO as the cryoprotectant, the antifreeze protein demonstrated minimal impact on post-thaw cell recovery, whether it was intra-or extracellular. This suggests that the primary mode of cell death in this system was not due to unwanted ice recrystallization, and hence the antifreeze protein offered limited protection. The same delivery strategy was subsequently applied to 2-D cell monolayers, a more challenging cell cryopreservation system where DMSO alone fails to enable significant cell recovery. Without the antifreeze protein, only ∼25% of monolayer-frozen cells were recovered post-thaw. Insignificant increases in post-thaw recovery was observed when the antifreeze protein was delivered into the cell; however, when applied extracellularly, the post-thaw recovery of cells rose from 25% to over 60%, representing a remarkable increase. These results provide evidence that extracellular macromolecular cryoprotectants result in superior cryoprotective effects when supplemented to DMSO cryopreservation and provides evidence that antifreeze proteins, and potentially additional macromolecular cryoprotectants, do not need to be designed with cellular uptake as a criterion. This work also shows that the requirements to enhance cell monolayer recovery are distinct from suspension, so a "one size fits all" approach might not always be suitable. Finally, these results show that the mechanisms of cell death during cryopreservation vary between systems. In summary, this work provides new insight into the requirements of a macromolecular cryoprotectant and provides a step forward toward being able to rationally design new materials to enable new therapies and biotechnological tools.