Vitrification and Nanowarming of Kidneys

Abstract Vitrification can dramatically increase the storage of viable biomaterials in the cryogenic state for years. Unfortunately, vitrified systems ≥3 mL like large tissues and organs, cannot currently be rewarmed sufficiently rapidly or uniformly by convective approaches to avoid ice crystallization or cracking failures. A new volumetric rewarming technology entitled “nanowarming” addresses this problem by using radiofrequency excited iron oxide nanoparticles to rewarm vitrified systems rapidly and uniformly. Here, for the first time, successful recovery of a rat kidney from the vitrified state using nanowarming, is shown. First, kidneys are perfused via the renal artery with a cryoprotective cocktail (CPA) and silica‐coated iron oxide nanoparticles (sIONPs). After cooling at −40 °C min−1 in a controlled rate freezer, microcomputed tomography (µCT) imaging is used to verify the distribution of the sIONPs and the vitrified state of the kidneys. By applying a radiofrequency field to excite the distributed sIONPs, the vitrified kidneys are nanowarmed at a mean rate of 63.7 °C min−1. Experiments and modeling show the avoidance of both ice crystallization and cracking during these processes. Histology and confocal imaging show that nanowarmed kidneys are dramatically better than convective rewarming controls. This work suggests that kidney nanowarming holds tremendous promise for transplantation.

circulating chiller, Julabo F32 with HP controller (range, -20°C to 100°C) was purchased from Boston Laboratory Company (Woburn, MA). LabVIEW (National Instruments, Austin, TX) codes for controlling the pump speed and real-time pressure / temperature recording using the DAQ, were written in-house. Pressure / temperature data was recorded at a frequency of 1 data point/sec. Flow rates (mL min -1 ) were experimentally calibrated vs pump rpm and validated against calculated values for a given tubing inner diameter. For the hypothermic perfusion setup, no oxygenator is used and there is no recirculation of CPA. The hypothermic perfusion parameters are listed in table S4. Temperature regulation of perfusate and organ chamber to setpoint (e.g., 0°C for hypothermic), tubing connection, and circuit-priming are conducted before organ retrieval to minimize cold ischemic times (≤ 30 min). The shortest possible tubing length is used to maintain temperature and avoid kinks. For perfusion experiments the peristaltic pump is started, the circuit is primed, and the system is tested for leaks and air-bubbles while circulating buffer/carrier solution is flowing. Perfusate temperature is tested and perfusate flow rate is adjusted to low speed (< 1 mL min -1 ) prior to kidney connection. For cleaning between experiments, all glass components are perfused and washed in a soap-bath and rinsed 3x with bleach + DI water, 3x DI water and 3x with ethanol.
For cold perfusion, the circulating chiller is set to -4°C and coolant circulation through the waterjacketed circuit components is turned on. 50 mL cold Euro Collins (EC) is poured into the perfusate reservoir maintained at 0-4°C on ice and circulation through the circuit is started.
Leaks and air-bubbles are removed through tilting and tapping the lines. The program for pressure/temperature recording is started and a baseline pressure is acquired at 1.5 mL min -1 .
Next, the EC flowrate is reduced to < 1 mL min -1 and the aorta of the kidney is connected to the perfusion circuit outlet tubing ( Figure 2A) while keeping the cannula submerged in UW in the cold-storage container to prevent air bubbles. The kidney is quickly and carefully transferred to the cold organ bath chamber while maintaining connection to the circuit. The vena cava is checked for effluent flow and is connected to a drainage tube for collecting effluents periodically (no recirculation), Figure S4.

Controlled Rate Freezer (CRF) and thermometry protocol for vitrification
CRF Program: The CRF is pre-programmed to start at a chamber temperature of T c =0°C and cool down to T c = -121°C at a ramp rate of -40°C min -1 (this rate was ascertained through previous in vitro thermometry tests on bags containing CPA only). A 25 min anneal step is introduced at -121°C (just above T g ) to allow the organ to equilibrate right before glass transition. A slower ramp rate of -10 °C min -1 was introduced to cool from -121°C to -150°C in the glassy phase to minimize thermal gradients. At T c = -150 °C, a 30-minute temperature-hold step is programmed to allow the whole organ to equilibrate to the storage temperature (-150 °C) before transfer to a storage freezer at the same temperature.
Kidney Vitrification: After insertion of fiber-optic probes, the kidney is placed in a VS55+sIONP-filled cryobag and submerged in wet ice. The temperature probes are bundled and directed out of the bag through a narrow opening at the top of the bag. The kidney and the surrounding VS55+sIONP occupy an approximate volume of 5.5 cm x 4.5 cm x 1.5 cm. Excess air is evacuated from the bag, the bag is sealed, and temperature probes are affixed with transparent tape (3M, Minneapolis). The bag is closely wrapped with Al foil, pressing any air out between the foil and the bag. The foil-covered bag is next placed on a custom holder, which is introduced inside the CRF at T c =0°C. Within the CRF, prior to starting the cooling program, the orientation of the bag/holder is adjusted so that the LN2 vapor flow is parallel to the flat bag faces, as shown in Figure S6. Temperature probes are directed out of the CRF through a vent and connected to the multi-channel temperature data-logger (Qualitrol T/Guard, Fairport, NY). The data-logger is programmed to record temperature at a 1 s interval and is started. The CRF is sealed closed, following which the kidney is allowed to equilibrate to 0°C. Once a stable baseline is achieved at 0°C (5 min), the cooling CRF program is initiated. The total ischemic time between disconnecting the kidney from the hypothermic perfusion circuit to onset of cooling for vitrification was 17.5±2.5 mins (Table S5). After the temperature hold step at -150°C is reached, the kidney is allowed to equilibrate for at least 30 min, during which the temperature at different points in the kidney are monitored in real-time to ensure equilibration. At the end of the protocol, when the kidney has equilibrated to -150ºC, the foil covered bag is quickly (2-3 s) transferred over LN2 vapor to a -150°C cryogenic storage freezer (MDF-C2156VANC-PA, Panasonic, IL). Here, the kidney is further stored for a total time of > 30 min (or theoretically indefinitely) to ensure equilibration, while the RF coil is prepared for nanowarming and the perfusion circuit for sIONP washout. The foil layer is peeled off from the bag, inside the storage freezer. Success or failure of vitrification of these kidneys was demonstrated by visual inspection and μCT analysis.

General considerations during cooling and rewarming
Successful vitrification is an optimization problem contingent on balancing cooling rates and thermal gradients. Therefore, the following general considerations are made in designing the cooling protocol: (i) Cooling rates > CCR: The minimum cooling rate required to out-run ice crystallization from T m to T g (or more appropriately, homogeneous nucleation temperature T h to T g ), termed as the Critical Cooling Rate (CCR), is a property of the CPA concentration and composition [7,27] . For 8.4 M VS55, the CCR is -2.5ºC min -1 (Table S3). Therefore, the chamber cooling rate is adjusted to -40ºC min -1 to allow rapid cooling through convection (rates determined through in vitro experiments in VS55 at volumes same as organ experiments).
(ii) Minimizing thermal gradients below glass transition: Adding an annealing step just above glass transition, which is long enough to allow temperatures across the kidney to equilibrate before entering the glassy state, will ensure minimal thermal gradient at the onset of cooling in glassy state. Mehl et al showed that, for VS55, ice nucleation occurs predominantly between -90ºC and -135ºC, whereas crystal growth is maximal above -85ºC [22] . Thus, in the design of cooling and rewarming protocols, annealing steps are limited to temperatures below -90ºC to reduce the probability of ice crystal growth. Additionally, a slow cooling rate (-10ºC min -1 ) within the glassy state to the storage temperature (-150ºC) will prevent gradients, and thereby thermal stress, from building up as the kidney is further cooled to the storage temperature.
(iii) Organ container and geometry: A cryobag capable of withstanding temperatures to -150ºC was chosen as the container to vitrify the kidney because: (a) the thermal expansion coefficient matches VS55 more closely (compared to a tube, or a hard container) [16] , thus reducing strain during cooling and rewarming; and (b) the shape flexibility of the bag allows use of smaller volumes of CPA for immersion of kidneys, and therefore enables more rapid cooling. Furthermore, the shape flexibility allows molding the geometry to a form that reduces stresses (aspect ratio ~ 1) [16] and allows the vitrified kidney and bag to fit the nanowarming RF coil.
Additionally, an Al foil covering over the bag was used during cooling to cool the bag uniformly through conduction between the foil and bag, compensating for the asymmetry of forced convection resulting from the unidirectional flow of LN2 in the CRF ( Figure S6). The foil also allowed molding the bag geometry to fit the RF coil.
As shown in Figure 3A (black dotted line) the chamber temperature profile meets all these performance requirements.
It is instructive to note that for T<-90ºC, nucleation of ice crystals is expected, especially during the 25 min annealing step at -121ºC, however ice crystal growth kinetics are too slow at these temperatures to have a to result in significant ice formation [22] . However, during rewarming, avoiding growth of ice crystals from pre-existing nuclei (formed during cooling), is a challenge, and therefore require rates of temperature change during rewarming that are even greater those during cooling in order to "outrun" ice growth. In Figure 4, perfusion of kidneys with VS55+sIONP (10 mg Fe mL -1 ) and surrounding it with VS55+sIONP (4mg Fe mL -1 (n=3) or 10 mg Fe mL -1 (n=4)) ensures sufficient rewarming rates (> CWR) to avoid ice expansion and low gradients in the glass region to prevent fractures. However, 10 mg Fe mL -1 in the solution surrounding the kidney can cause the surface of the kidney to reach 0ºC or higher, at which CPA toxicity is increased, while the medulla has not risen above melting (T m ). This can cause unintended toxicity on the surface of the kidney due to higher temperatures while exposed to CPA. Thus, while 10 and 4 mg Fe mL -1 were used in the thermometry cohort shown in Figure   4A, 4 mg Fe mL -1 was used in the suspension solution for all biological endpoints.

μCT setup for imaging vitrified kidneys
The vitrified kidney in the cryobag and was held at LN2 vapor temperature (-150°C) in a styrofoam container during imaging. Separate tubes of water and air at RT were attached to the top of the container to serve as calibration references for determining Hounsfield unit radiodensity. A typical scan takes 30 min to finish, and temperature history monitored in the sample chamber during that 30 min ranged from -174 to -158°C, well below the glass transition temperature. The images were reconstructed to reduce the beam hardening artifacts and improve image quality (3D CT pro, Nikon Metrology, MI). The images were converted to unsigned 16-bit float images, post-processed (VGstudio Max 3.2, Volume Graphics, NC), and exported as DICOM images for a final analysis using MATLAB (MathWorks). The scale in HU was determined rom calibrated values based on the air and water sample densities, and the resulting images were used to document successful vitrification as frozen samples and vitrified ones have different density and appearance [44] .

Computational heat transfer modeling
The computational geometry consists of kidney domain and bag domain containing the 100% VS55+sIONP (4 mg Fe mL -1 ) solution. The bag is approximated to have a VS55+sIONP volume of 25 mL, modeled as 5.5 cm high, 4.5 cm wide and 1.5 cm thick with an ellipsoidal crosssection from top view and parabolic cross-section from side-view. The rat kidney is modeled as an ellipsoid with a major axis as 2 cm and the other two of its minor axes as 1 cm keeping the overall volume of kidney ~ 1 mL consistent with anatomical literature values [45 46] .

Equation S2
The thermal properties are chosen for the vitrified CPA (VS55) solution inside the bag and kidney assuming CPA equilibrated in the kidney tissue. In particular, k = 0.3 W mK -1 , C p = f(T) (specific heat for VS55 is used as a function of temperature for the computational domain [12,47] and ρ= 1100 kg m -3 [47] . Table S1 summarizes thermal properties. The numerical solution to the non-homogenous transient heat transfer equation (Equation 1) is solved in the commercial FEA code COMSOL Multiphysics 5.5. The mesh was created using the finer option in COMSOL meshing code with tetrahedron elements consisting 5955 mesh vertices and 31226 tetrahedra. Convergence of solution was verified using COMSOL in-built physics-controlled feature with relative error of 0.01%. Further, mesh convergence was evaluated using different mesh element sizes from the physics-controlled mesh submenu in the mesh node of study. The simulation is performed for 2 different cases during cooling i.e. cooling in CRF and LN2 plunge cooling and similarly 2 different cases during rewarming i.e. nanowarming and convective warming in water bath. The initial temperature of domain is at 4°C during cooling and -138°C during rewarming. At the boundary of computational domain i.e. bag, the heat transfer mode of interaction between the domain of interest with the surroundings is considered by only convection of LN2 vapors in the cooling chamber for case of convective cooling in CRF as well as LN2 plunge case with different heat transfer coefficients in both cases. In the rewarming cases, the heat flux across the bag during nanowarming case is approximated as natural convective boundary condition whereas forced convection of water (at 37°C) is the mode of heat transfer at the boundary in convective warming (WB) case. Detailed information regarding boundary conditions such as "h" values, T ∞ and additional details can be found in Table S6. Also, the details regarding the mesh size, number of elements, and convergence can be found in Figure S9. A simplified thermal shock equation Equation S3 was used to calculate the temperature difference corresponding to stress-to-fracture for VS55.
where the Coefficient of Thermal expansion,

Equation S4
g is a geometric coefficient = 0.5 approximated for the modeled geometry, ν is Poisson's ratio = 0.2, E is the modulus of elasticity = 1 GPa and σ is the tensile strength of VS55 = 3.2 Mpa. After inserting all values, the limiting value (i.e. minimum) ΔT max ~ 38°C. In Figure 5D, experimental CRF cooling rate (black) was computed by taking the average of all temperature probes within a kidney, and then averaging this data over n=7 kidneys. For modeling, CRF CC dT/dt (blue) was computed by taking the average dT/dt of the maximum and minimum temperature limits across the modeled kidney volume. In Figure 5I, experimental NW (black) was computed by taking the average of all temperature probes within a kidney, and then averaging this data over n=7 kidneys. For modeling, NW dT/dt was computed by taking the average of the maximum and minimum temperature limits across the modeled kidney volume. Fe estimation in the medulla and cortex were obtained through calculations and analytical measurements-assuming average vascular volume fraction in entire rat kidney as 20% (20-25% is generally the vascular volume fraction in a kidney [45,46] , calculated for limiting case 20%). cFe (total) = 0.2*10 mg Fe mL -1 * 1 mL = 2 mg Fe (vascular volume ~ 1mL). Alternately, based on ICP-OES [19] , total Fe in kidney = 0.012 (mg Fe mg -1 dry wt.) * 163.7 (mg dry wt.) = 1.966 mg Fe. Using this information, and (cFe) medulla /(cFe) cortex ~ 2 from MR imaging (Figure 2), cFe medulla ~ 2.15 mg Fe mL -1 (tissue concentration), cFe cortex ~ 1.075 mg Fe mL -1 .

ICP-OES for Fe quantification in kidney washout samples
Each kidney sample was dried in a vacuum oven (LT1495X2, Thermo Fisher Scientific,

MR imaging for IONP distribution
All kidneys were inserted into a 50 mL PET falcon tube and surrounded by the same solution that the kidney contained (EC or VS55) but without IONPs. The SWIFT 3D R1 map was

Diffusion time calculation for VS55 diffusion in tissue.
Diffusion time, t ~ x 2 /D, where, x = mean intervascular distance in a kidney = 200 μm [45,46] . D = VS55 diffusivity in tissue = 7x10 -11 m 2 s -1 was obtained by μCT characterization of tissue-VS55 concentration vs time [37] . Therefore, diffusion time, t ~ 9.5 mins. This implies that, in theory, 15 min perfusion time at each concentration step should be sufficient for VS55 to diffuse into the tissue parenchyma.

Johnson-Avrami estimation of approximate ice formed during nanowarming
It is worth noting that in Figure 4A and Figure S7, there are regions in the temperature vs time nanowarming plot, where the instantaneous warming rates drop slightly below the CWR (in the 40-50ºC min -1 range), especially in the cortex. Using the Johnson-Avrami model, originally developed for modeling ice crystallization kinetics during cryopreservation [22] , the ice-fraction can be estimated from summation of the following equation: , over the temperature range of -90ºC (T1) to -45ºC (T2), where X is the ice-fraction. According to Mehl et al [22] , for VS55, an Arrhenius plot of the kinetic constant k(T) in a logarithmic scale vs the inverse of temperature generated a fitting function for ln(k(T)) = 21.28-4661/T. In addition, an average estimate for the Avrami exponent over the temperature range of -90 to -45ºC was assumed based on experimental data in Mehl et al [22] , n = 2.198. Next, a discretized form of the integral of Eq. S5 was used to estimate the ice-fraction. This analysis provides a rough approximation for the volume fraction of ice expected, assuming an Arrhenius dependence of nucleation and crystal growth rates on temperature. From this analysis, nanowarming results in an average ice volume fraction < 5%. In the worst-case scenario during nanowarming, where the rewarming rates are lowest in the cortex relative to other regions in the kidney, the estimated ice fraction in the cortical mass would be ~ 7%. While this ice fraction can be significant depending on the location and damage caused, the measured hypothermic pressures during sIONP washout after nanowarming show that the pressure was within acceptable range (<100 mmHg) in at least 2/3 of the kidneys ( Figure S11) and responding to osmotic shifts similar to controls, suggesting minimal damage. Additionally, this ice fraction can be further reduced by increasing warming rates through one of the following means: (i) increasing volumetric concentration of sIONPs in the perfusate to c(Fe)>10 mg mL -1 (ii) using a nanoparticle with higher SAR under the available field conditions, (iii) increase the frequency or field magnitude of the applied RF field. In the current study, we limited c(Fe) in the perfusate to 10 mg Fe mL -1 because higher colloidal instability is more likely at higher concentrations which could result in a poorer washout of the sIONPs from the kidneys. Thus, in the current study, 10 mg Fe mL -1 was chosen as the concentration to achieve sufficient rewarming rates without adversely affecting washout.
However, the impact of higher sIONP concentrations can be studied further.

Washout characterization methods in stable vs unstable colloids
In contrast to other methods for measuring residual iron content following washout of sIONP from kidneys, specific observations of arterial pressure during perfusion can indicate failure from perfusion of unstable CPA+IONP colloids. Generally, colloid stability criteria are established prior to perfusion using in vitro methodseg. an unstable colloid does not pass freely through a 1-4 µm PES filter. In addition, observation of arterial pressure during ex vivo perfusion can indicate instability. For eg, during kidney perfusion, a non-linear increase in perfusion pressure (P sIONP ) vs time or a high dP sIONP /dt (>10 mm Hg min -1 ) ( Figures S3B and S14), could be a potential sign of blockages in the kidney vasculature from sIONP aggregates. High grade blockage can be tested by turning or pulsing the pump on and off while observing the pressure relaxation time to equilibrium. Blockage results in failure of the pressure to return to baselinegreater the blockage, longer the relaxation time to baseline following each pulse (gray shaded areas and blue dotted reference line in Figure S3B). Vascular occlusion from sIONP aggregates could result in localized high pressure and induce damage to the vasculature and endothelial lining as pressure exceeds physiological pressure limits, as evidenced in Figure S3I-K, S3N-U.
Indeed, in the few cases where high washout pressures (>>100 mm Hg) were observed during perfusion with an unstable colloid, kidney characteristics post-washout included visual and histological confirmation of sIONP retention in the glomeruli and vasa-recta ( Figure S3L-M) and significant damage to the endothelium in blood vessels, especially in the medulla ( Figure S3T-U). Additional characterization of a kidney using μCT, following sIONP loading at dP sIONP /dt > 100 mm Hg min -1 (Figures S3I-K and S14), showed features consistent with sIONP aggregation in the kidney (compare to Figure 2K). High intensity pin-point spots in the cortex corresponded to IONP localization in the cortex, very likely in the glomeruli. Additionally, high contrast lines were observed in the medulla where medullary rays (collecting ducts and straight tubules) and the vasa-recta were present. Prussian blue staining indicated iron retention in the glomeruli in the cortex and inter-tubular space in the medulla. This likely indicated damage of the capillaries in the vasa recta and leakage into the inter-tubular space, appearing to take the geometric pattern of the medullary rays. In some cases, a yellow colored residue in the glomeruli and vasa recta were observed post-washout ( Figure S5C), which stained negative for Congo Red and PAS, indicating very likely remnants of unstable sIONPs or coating left behind due to incomplete washout.
Additionally, no sIONPs were observed in the effluent from the ureter (clear), suggesting that the sIONPs are less likely to enter the tubules and the collecting duct. This was further verified by Prussian blue staining of the kidneys, where no blue stains for Fe were observed in the tubules or collecting ducts ( Figure S5B).

Characterization of sIONP size and magnetic properties before and after washout
Dynamic Light Scattering (DLS, Brookhaven Zeta PALS instrument, Holtsville, NY) was used to measure the mean hydrodynamic diameter of sIONPs in VS55 before loading and after washout. 400x diluted samples of 10 mg Fe/ml in VS55 from three separate kidney perfusion experiments were used for the measurement. Histogram depicting intensity-weighted contributions of size ranges (G(d)) and the cumulative percentage contributions (C(d)) were plotted vs size (d) ( Figure S12).
Vibrating Sample Magnetometry (MicroMag VSM, Lake Shore, Princeton, NJ) was used to measure the DC hysteresis properties of sIONPs. Samples from kidney perfusion experiments were measured at room temperature to determine if they retained their superparamagnetic properties ( Figure S13). A liquid helium cryostat allowed measurements of sIONP in VS55 at cryogenic temperatures. A stepper motor that rotates the vibration head was used to center the sample. Vibration amplitude of 0.2 and averaging time of 0.1s were used.    Figure S14). This is evidenced by the increase in relaxation time to equilibrium (light blue dot-dash line), following each pump ON-OFF sequence and a failed washout as shown in (D). C) Washout pressure, P wash , from perfusion and washout of a stable VS55+sIONP colloid, where P wash < 100 mm Hg through the washout and pressure decreases with time as the CPA concentration is reduced to baseline (buffer). D) Failed washout from an unstable VS55+sIONP colloid, where either (i) P wash increases continuously, beyond the pressure sensor saturation (~400 mm Hg) under constant flow conditions (~0.5 ml min -1 ) and physically ruptures the kidney (data not shown) or, (ii) P wash does not relax to baseline pressure when pump is switched OFF to maintain P wash <100 mm Hg, indicating the venous end (at atmospheric pressure) is blocked off from the arterial end. E) ICP-OES measurements (n=3 per group) comparing iron retention in mg Fe mg -1 dry weight for successful (stable colloid) vs failed (unstable colloid) washed out kidneys. 2.5-3x higher retention is observed in kidneys blocked from an unstable colloid. *p<0.05, **p<0.005, ***p<0.0005 (One-way ANOVA with Tukey's posthoc multiple comparisons test). For reference, LOD in the measurement is 1 ng mg -1 . It was previously shown that a kidney fully perfused with sIONP in VS55 (10 mg Fe mL -1 ) displayed a c(Fe) ~ 0.012 mg Fe mg -1 dry weight [19] . Thus, in washed out kidneys, 20% or lower Fe is retained, whereas failed washout in kidneys (unstable colloid) corresponded to approximately 2.5x higher Fe than washed out kidneys (stable colloid). The red dotted line indicates a concentration of Fe previously observed (0.0024 mg Fe mg -1 dry weight) [26] in kidneys of mice which were i.v. injected with IONPs (0.18 mg Fe g -1 mouse), 24 hours post injection. This concentration appeared to be well tolerated by the mouse for at least a month. The ICP-OES data is generally in agreement with observation of gross hemi-sections of the kidney where (G) shows kidneys with a darker contrast in the medullary region after perfusion and wash out from an unstable VS55+sIONP colloid, compared to a washed-out kidney (stable colloid) (H). I) μCT image of a kidney perfusion loaded with VS55+sIONP (unstable colloid) at a rate, such that rate of increase in P sIONP > 100 mm Hg min -1 .    Figure S14. (A-C) sIONP loading pressure (P sIONP ), rate of increase of sIONP loading pressure (dP sIONP /dt) and initial washout pressure (P wash-out ), respectively for a kidney optimally loaded with a stable sIONP+VS55 colloidal system. (D-F) sIONP loading pressure (P sIONP ), rate of increase of sIONP loading pressure (dP sIONP /dt) and initial washout pressure (P wash-out ), respectively for a kidney loaded with an unstable aggregation-prone IONP+VS55 colloid (clogs a 4 µm PES filter). Failure cases (D-F) are characterized by high P sIONP (>100 mm Hg), high dP sIONP /dt (>>10 mm Hg/min), and high initial washout pressures (>100 mmHg). (G) Correlation plot of P wash-out vs dP sIONP /dt for stable (black squares) and unstable colloids (red triangles).

J, K) show zoomed in pictures
Green-shaded region indicates 95% confidence intervals for a linear correlation between P wash-out vs dP sIONP /dt, which is expected for stable loading as the resistance at the end of sIONP loading is equal to the resistance during the beginning of washout. The unshaded white boxed region in the lower left corner in (H) indicates the region of optimal loading conditions. The circled ROI indicates the outside probe at a higher steady-state temperature compared to other probe locations when the kidney is equilibrating to the CRF chamber at -150ºC. (B) Rate of decrease in temperature measured at different probe locations in the temperature range of -121 to -150ºC (circled region in (A)). The rates indicate that the outside probe is cooling at a similar rate compared to other probe locations. (C) Offset measured in two fiber-optic probes in the -20 to -150ºC temperature range, when calibrated against a Pt100 RTD. The difference in offsets between the two fiber-optic probes used here increases as the temperature decreases. (D) Temperature vs time plot as in (A) after applying probe-specific offset correction. Perfusion loading Step-loading VS55 (table S2) CPA flow rate 1.5 ml min -1 (constant) P.V. pressure (mm Hg) 0-100 mm Hg (variable with steps)

Cryopreservation Step Cold Ischemic Time (mins)
Cold storage  Perfusion ≤ 20 Perfusion  onset of ramp cooling in CRF 17.5 ± 2.5