Noninvasive Quality Control of Cryopreserved Samples

Abstract

We present a novel noninvasive technology for quality control in biobanking. We implemented a contactless optical in situ method with a remote detection unit. The method detects physical and chemical changes by emission spectroscopy. In the present study, ice formation in a vitrified sample is revealed by Raman scattering. The technology allows us to monitor sample quality during cold storage and to assess the sample state after preservation, storage, or transport without the need for thawing.

Cryopreservation is the only applicable technique for preserving life and function during long-term storage of biological samples such as cells and tissue. The demand for biobanking in reproductive and regenerative medicine, tissue engineering, (stem-) cell biology, and pharmaceutical research is in rapid growth. In particular, the long-term storage of human samples (other than blood and plasma) can be sophisticated.

Such samples are usually unique and irreplaceable. Hence, cells leaving long-term storage cannot be globally tested for physiological and genetic soundness before their medical application. Therefore, it is obligatory to ensure a damage-free long-term storage. Once aqueous samples have reached a temperature below −137°C, they are considered to be stably cryopreserved, as this is the glass transition temperature (T_g) of amorphous water and no ice crystal formation should occur below T_g. Furthermore, chemical and biological activity and diffusion virtually stop below this temperature. These aspects are particularly important with respect to vitrified samples which have solidified in a metastable, amorphous (glassy) state by sudden cooling. Ice formation in vitrified samples due to transient infringing of the T_g usually leads to loss of the complete sample.^1,2 Furthermore, mechanical and thermal incidents may occur without notice in biobanking practice (e.g., when moving single specimens or entire sample collections). An interrupted cold chain may have ruinous effects on a sample, but currently is not traceable before rewarming. Against this background, it is remarkable that still no quality monitoring of long-term stored samples is conducted at present, neither in situ nor after sample-withdrawal. ³ Technical solutions for monitoring the status of the sample in situ are not yet available and there is no established technique for retrospective proof of physical and chemical changes inside the sealed sample. Preliminary work ⁴ in the field of low-temperature characterization of frozen samples found optical methods to be significant, noninvasive, and insensitive to cryogenic conditions. In particular, Raman scattering techniques were powerful approaches for monitoring and investigating cryopreserved samples. Here we focus on the OH-stretching band to investigate whether the sample water is amorphous or crystalline. ⁵ A more phenomenological approach would be to record a characteristic fingerprint spectrum of a freshly preserved sample and use it as a reference to confirm sample integrity. Such fingerprint spectra include the spectral range of typical vibration frequencies of organic compounds are often used for characterization of biological materials. ⁶

We set up a fiber-based Raman probe for a proof-of-concept experiment. A Cryostage (CS) (Linkam MDS 600, Linkam Scientific, UK) served as a mimic for the sample (S) during cryogenic storage. The sensing setup consists of confocal probe interface (CPI), connected via fiberoptics to a dichroic fiber junction (DFJ), as well as a fiber-borne CCD-spectrograph (Fig. 1a). The confocal probe interface is a combination of a fiber coupler and a focusing optic. The dichroic fiber junction couples a free beam laser (532 nm, Coherent Compass, USA) via beam expander (BE), dichroic beam splitter (DBS), and a fiber coupler into the probe fiber (PF). The Raman light coming back from the probe fiber is collected by the same fiber coupler, separated from excitation light by the dichroic and a long-pass filter (LP, Razoredge, Semrock) and a coupled-into-the-detection fiber (DF). The collected Raman light is spectrally detected by a fiber-borne spectrograph assembly (iDus CCD-detector with Shamrock 303 polychromator (Andor Technology plc., UK).

FIG. 1.

(a) Low temperature fiber sensor setup: symbols are explained in the text; (b) low temperature scanning confocal setup for Raman spectral imaging. BE, beam expander; CPI, confocal probe interface; CS, Cryostage; DBS, dichroic beam splitter; DF, detection fiber; DFJ, dichroic fiber junction; LN2, liquid nitrogen; LP, longpass filter; PF, probe fiber.

For a microscopic investigation of the crystallization, the Raman technology was adapted to a sample-scanning confocal microscope with the Linkam stage as cryogenic microscopy chamber. Mounting the cryo-chamber on a xy-sample-scanner (Physik Instrumente, Germany) of 300 μm traveling distance enables Raman spectral imaging (Fig. 1b). Image reconstruction was performed by an in-house-developed Labview Software.

Scanning Raman microscopy was performed in order to identify the origin of spectral bands using the setup shown in Figure 1b. Each pixel of such a microscopy image contains an entire Raman spectrum. By subsequent data analysis, images of arbitrary spectral channels can be generated displaying isolated molecular vibrations (i.e., chemical compounds). Hence, Raman microscopy enables one to measure and map the concentration as well as the chemical state of the main constituents of a sample. To resemble typical slow freezing conditions, we cooled a single suspended human mesenchymal stem cell (hMSC) down to −120°C at a rate of −1 K/min in a PBS solution containing 10% DMSO. The respective Raman images of three isolated vibrations are shown in Figure 2a and the corresponding spectral data are plotted in Figure 2b. In Figure 2a, concentration maps of cellular matter (measured as hydrocarbon backbone vibration C₃), ice (measured as lattice OH stretching vibration), and DMSO (measured as C-S-stretching vibration) of a hMSC slowly frozen to −120°C are shown. Figure 2b displays the respective spectra of pixels from the cellular body, extracellular ice phase, and dendritic fluid channels (residual extracellular liquid phase), and further denotes the vibration bands used for the chemical mapping in Figure 2a. The contrast mechanisms of low temperature confocal Raman microscopy are described elsewhere in detail.^7,8 The O-H stretching vibration region beyond 3000 cm-1 differs significantly between the crystalline and amorphous states of water (Fig. 2b). This distinct difference in spectral shape was used to monitor recrystallization in vitrified samples as described in the next paragraph.

FIG. 2.

(a) Scanning Raman microscopy images (50×50 μm², 64×64 pixel²) of hMSC at −120°C cooled down with 1 K/min in PBS 10% DMSO. Each image maps the intensity of an isolated Raman band and hence the distribution of a specific compound; (b) Raman spectra from three image pixels (dashed line: a pixel from ice; dotted line: a pixel from DMSO-rich extracellular fluid; continuous line: an intracellular pixel).

In order to prove the concept of monitoring recrystallization of vitrified samples during cold storage, we mimicked a measured typical temperature profile of a sample vial while a neighboring vial was withdrawn from the cryocontainer. The process was monitored with a remote sensor setup (Fig. 1a) using 2 μL glassy vitrification medium (without cells) in the Linkam stage (vitrification medium: PBS (300 mM sucrose), 20% ethylene glycol, and 20% DMSO, Tg=−128.5°C, measured by DSC). ⁹ We started the monitoring process at −141°C and heated the sample at a rate of 2 K/min. Until a temperature of −127°C, the spectra show no significant changes. The crystallization onset was measured at about −127°C and completed at about −122°C. The ice formation in the sample mimic is displayed as a spectral series in Figure 3, correlated to the temperature rise. The characteristic “ice band” was reproducibly detected (n>3) on crystallization of samples of various compositions, even in cell suspensions. Differential scanning calorimetry of the sample composition used in Figure 3 verified the results of the Raman study.

FIG. 3.

A series of Raman spectra from a sample mimic heated from −141°C to −100°C, indicating the formation of ice at −127°C by the occurrence of the characteristic ice band at 3120 cm⁻¹. Exemplary Raman spectra of glassy water and ice are shown on the right. The findings of the sensor experiment are in accordance with DSC data.

Ice formation and the application of cryoprotective substances in cryopreserved samples can be unambiguously detected at cryogenic temperatures by fiber-based Raman spectroscopy. Furthermore, this technique allows a differentiation between CPA as well as freezing protocols used, such as vitrification or slow freezing. As this technique is moreover capable of detecting any other chemical change in principle, it turns out to be a powerful method for contact-less monitoring of the status of cryopreserved samples during storage. The technical implementation of the concept may lead to automated monitoring of samples under cryogenic storage in the future. However, the storage technology has to allow transport of samples to the optical interface inside the cryo-tank. The optical interface and sample container have to be designed in a way that allows a representative measure (i.e., a multisport or widefield measure). Such efforts will be made only for very valuable samples. A short-term implementation into biobanking practice is the spectroscopic pre- and post-storage sample characterization in cryogenic temperature workbenches to check for chemical changes during the conservation and for ice formation in vitrified samples. A quality assessment for sample transportation can also be performed this way. The confocal optics of the present fiber probe already allows measurements through turbid container walls, as was tested on Teflon straws. A multispot confocal probe for workbench application is in development at the moment. Raman spectroscopy could be used to develop catalogue references for quality management of cryopreserved samples. Comparison to catalogued spectra may enable retrospective sample analysis.