Article pubs.acs.org/crystal
Quantitative Analysis of the Efficacy and Potency of Novel Small Molecule Ice Recrystallization Inhibitors Stephanie Abraham, Kerkeslin Keillor, Chantelle J. Capicciotti, G. Evan Perley-Robertson, Jeffrey W. Keillor, and Robert N. Ben* Department of Chemistry and Biomolecular Sciences, University of Ottawa, 10 Marie-Curie, Ottawa, Ontario K1N 6N5, Canada S Supporting Information *
ABSTRACT: The effects of ice recrystallization are well-recognized throughout the literature. This phenomenon is the major cause for cellular damage during thawing of cells, ultimately reducing post-thaw viability and function. In this paper, we describe a method for quantifying the inhibitory effect on ice recrystallization of novel small molecules that are cryoprotectants for red blood cells. The method is ideally suited to the splat-cooling assay, where high-ice volume fractions are present. Using our method, we have derived first-order rate constants for the increase in the average crystal size based upon a “binning” approach of ice crystals as a function of size and time. Using this reliable metric, dose−response curves were constructed to obtain IC50 values. Two very effective inhibitors of ice recrystallization, p-methoxyphenyl β-D-glucopyranoside (PMP-Glc) and pbromophenyl β-D-glucopyranoside (pBrPh-Glc), had IC50 values of 16.3 and 14.8 mM, respectively. Interestingly, the Hill slopes from these dose−response curves were 5.12 ± 0.81 for PMP-Glc and 3.12 ± 0.62 for pBrPh-Glc, suggesting that an element of cooperativity may be involved in the mechanism by which these compounds inhibit ice recrystallization. This is particularly interesting, as unlike antifreeze (glyco)proteins, these small molecules do not bind to the ice surface.
■
INTRODUCTION
increase in average ice crystal size and also a reduction in free energy.10−12 Ice recrystallization has been identified as a problem in the frozen food industry as well as during cryopreservation of biological materials. Frozen food products have a finite shelf life because of the fact that ice recrystallization directly affects the texture, taste, and overall quality of the frozen product.13,14 Ice recrystallization is also the cause of extensive cellular damage that occurs during freezing and thawing, ultimately causing reduced post-thaw viability and function.15,16 These problems can be addressed with substances that inhibit the process of ice recrystallization. Reliable methods for the assessment of ice recrystallization and the inhibition of this process include the capillary method assay, wide-angle X-ray scattering, differential scanning calorimetry, and the original splat-cooling assay.17−23 In the capillary assay and splat-cooling assay, recrystallization is assessed by quantifying the change in the size of ice crystals in a sample after annealing for a predetermined amount of time at specific sub-zero temperatures. The ice crystal size in the sample is compared to a positive control for recrystallization such as a phosphate-buffered saline (PBS) solution. The capillary assay is qualitative in nature, while a number of
Recrystallization is an extensively studied phenomenon, and its importance in the field of geology and metallurgy has been thoroughly documented.1,2 The phenomenon is defined as the nucleation and growth of large nondeformed grains in the presence of smaller ones. In deformed grains, the atom arrangement is not ideal, increasing the overall energy of the system. Thus, reorganization to nondeformed grains and an overall decrease in the energy of the system occur.3−5 As such, recrystallization is a thermodynamically driven kinetic process resulting in an overall reduction in the free energy of the system. In ice, this growth of larger ice crystals at the expense of smaller ones is thought to occur through either grain boundary migration or Ostwald ripening.6−12 Grain boundary migration occurs through the direct transfer of individual water molecules from smaller unstable ice grains or crystals to larger more favorable ones, resulting in an overall decrease in curvature in an ice crystal and a subsequent reduction in energy. However, grain boundary migration neglects the presence of bulk water or the quasi-liquid layer (QLL) between individual ice crystals.8 In the Ostwald ripening model, water molecules transfer from the surface of small ice crystals through bulk water and the QLL to larger ice crystals. Smaller ice crystals have a higher surface area to volume ratio, which gives them a higher surface free energy; thus, the Ostwald ripening process results in an overall © 2015 American Chemical Society
Received: July 14, 2015 Revised: September 4, 2015 Published: September 14, 2015 5034
DOI: 10.1021/acs.cgd.5b00995 Cryst. Growth Des. 2015, 15, 5034−5039
Crystal Growth & Design
Article
always 1.0 (as designed), and the final end point Arel ∞ was nearly always zero. Under these conditions, 60 min kinetic runs typically corresponded to 3−30 half-lives of a monophasic, monoexponential decrease. Dose−Response Curves. Rate constants for the decrease of bin 1 were determined over 60 min in the presence of six to seven concentrations of ice recrystallization inhibitors in PBS buffer, ranging from 1 to 100 mM (2 log units). We also measured this decrease in the absence of inhibitor (0 mM) to define our upper plateau. For each sample concentration, triplicate wafers were prepared and analyzed as described above. The average rate constant measured in triplicate at zero inhibitor concentration (i.e., in PBS buffer alone) was used to normalize the rate constants measured in the presence of inhibitor. This provided, for each inhibitor, a set of normalized rate constants, knorm, versus inhibitor concentration, [I], whose log values were used in dose−response fitting according to eq 2.
quantitative versions of the splat-cooling assay exist and have been commonly utilized.17,24−26 While ice recrystallization inhibition (IRI) activity can be quantified by comparing ice crystal sizes, the manner in which inhibitors of this process function is still largely unknown. As such, a number of reports describing the IRI activity of naturally occurring antifreeze (glyco)proteins [AF(G)Ps] have figured prominently in the literature over the past 8 years.27−35 More recently, the kinetics associated with inhibiting ice recrystallization with AF(G)Ps has been examined, and Koop et al. have made significant contributions and obtained “dose−response” curves to better quantify activity.12,36,37 In this approach, the proportion of ice to the unfrozen fraction is small and not representative of a fully frozen sample such as that observed during the cryopreservation of cells and tissues.38 In this paper, we describe a method for assessing the kinetics of ice recrystallization inhibition of novel small molecule inhibitors39,40 by looking at the proportion of smaller ice crystals with respect to sample size when large amounts of ice are present. Rate constants for crystal growth are obtained, and IC50 values for these novel small molecule ice recrystallization inhibitors are calculated.
■
k norm =
(2)
In this two-parameter sigmoidal equation, IC50 is the concentration of inhibitor that gives 50% antagonism (i.e., knorm = 50) and n is the Hill slope.
■
RESULTS AND DISCUSSION Ice recrystallization is typically monitored in either (a) a frozen wafer, generated by a splat-cooling assay, or (b) a slurry of ice crystals formed in a supercooled solution. In the slurry method, characterized rigorously by Koop and co-workers, the fraction of ice volume is small. In contrast, the fraction of ice volume in the frozen wafer method is very large and remains remarkably constant during the annealing process. The latter closely resembles that of a cryopreserved cell, whose cytosol is completely frozen or vitrified.38 Therefore, the splat-cooling assay serves as an excellent model for the evaluation of our ice recrystallization inhibitors, in light of their potential as cellular cryoprotectants. In the past, the effect of a given ice recrystallization inhibitor has traditionally been determined by its ability to reduce the average size of crystals growing in a frozen wafer, as measured at a single time point.24 However, this approach ignores two well-known aspects of the recrystallization phenomenon: the time dependence of crystal growth and the heterogeneous nature of crystal sizes that are observed during the recrystallization process. In this study, we have chosen to examine a number of novel IRIs that are among the first reported small molecules capable of inhibiting ice recrystallization. Specifically, we examine p-methoxyphenyl β-D-glucoside [PMP-Glc (Figure 1)] and p-bromophenyl β-D-glucoside
EXPERIMENTAL SECTION
Materials. D-Glucose was purchased from Sigma-Aldrich. pMethoxyphenyl β-D-glucopyranoside (PMP-Glc) and p-bromophenyl β-D-glucopyranoside (pBrPh-Glc) were synthesized according to a literature method.40 Crystal Measurement. Ice recrystallization was measured using a PBS solution in a splat-cooling assay.17 A 10 μL drop of each solution was dropped from a height of 2 m through a shielding tube onto a polished aluminum block, which was precooled to −80 °C on dry ice. The “splat cooling” of the solution landing on the block instantly produced an ice wafer ∼20 μm in thickness and 1 cm in diameter. The wafer was then quickly moved onto a cooled microscope cover glass and transferred to a cryostage. The sample was annealed at −6 °C for 60 min, during which a section in the middle of the wafer was photographed using a microscope equipped with an LMPlanF1 20×, 0.40 objective. The cryostage temperature was maintained with a programmable Peltier unit (S3 Series 800 temperature controller, Alpha Omega Instruments). The image recorded at each time point was analyzed using ImageJ. Specifically, all crystals within the field of view were circled by hand, excluding those only partially visible at the image boundary (see the Supporting Information). The area of each circled crystal was calculated using ImageJ and corrected for the appropriate magnification factor of the objective lens. These crystals were then sorted into discrete bins using Excel. Bin sizes were assigned in increments of 0.001 mm2, as it was observed that at time zero, all ice crystals could be just contained within this bin. In this way, subsequent crystal growth would result in larger crystals moving out of bin 1 and into higher bins. The relative importance of each bin was determined by summing the area of each crystal within that bin and dividing by the sum of the areas of all crystals within the field of view. In this way, the proportionate area of each bin was calculated for every sample wafer. Kinetics. Each sample wafer was allowed to develop over 60 min, with images recorded and analyzed typically at nine time points, namely, 0, 5, 10, 15, 20, 25, 30, 45, and 60 min. The proportionate area of each bin was thus followed as a function of time. The decrease in the relative area of bin 1 was found to decrease monoexponentially and was fit, using nonlinear least-squares analysis, to eq 1. rel A trel = A 0rel (e−kobst ) + A∞
100 1 + n × 10 logIC50 − log[I]
Figure 1. Chemical structures of PMP-Glc and pBrPh-Glc.
(pBrPh-Glc) that are effective inhibitors of ice recrystallization at low millimolar concentrations and also function as effective cryoprotectants for human red blood cells (RBCs). In this application, reduced concentrations of glycerol (only 10−15% vs 40%) are utilized with slow freezing rates.40 Reducing the quantity of glycerol during the freezing process results in a decreased post-thaw processing time during which glycerol is removed prior to transfusion, making frozen RBC units more
(1) Arel t
is the measured relative area In this monoexponential equation, of bin 1 at time t, Arel 0 is the fitted relative area at time zero, kobs is the fitted rate constant for the decrease of the relative area of bin 1, and rel Arel ∞ is the fitted final end point. The fitted value for A0 was nearly 5035
DOI: 10.1021/acs.cgd.5b00995 Cryst. Growth Des. 2015, 15, 5034−5039
Crystal Growth & Design
Article
equally sized “bins”, where the incremental bin size was determined by the smallest area into which all crystals could be fit at time zero. The summed area of all the crystals in each bin was then converted to a proportion of the area of the total sample, to emphasize the relative importance of each bin. Representative data are shown in Figure 3, obtained using the splat-cooling assay in the presence of 22 mM PMP-Glc. At time zero, 277 crystals were measured, and their individual areas were determined to be
