Abstract
The primary responsibility of biobanks is to collect biospecimens that are true reflections of the local population, thereby promoting translational research that is applicable to the community. The Swedish Cervical Cytology Biobank (SCCB) was designed as a hospital-integrated biobank in 2011. The SCCB has now been implemented in 10 county councils scattered across the country. It is headquartered at Karolinska University Hospital in Stockholm. The SCCB now processes more than 60% of the liquid-based gynecological cell samples obtained throughout Sweden.
To improve the productivity of health care and research that rely on SCCB samples, a high level validation of the biobank system according to the principles of Good Laboratory Practices (GLP) is required. The performance of an entire high-throughput system validated by measuring the cell yield proved unsatisfactory after 1 year of sample collection and aliquoting. However, the results led to a number of high quality technical interventions for workflow enhancement. Subsequently, the improved process was applied to the system and led to a significant increase in cell yield. After the integration of the improved high quality methodology into the SCCB, the biobank services progressed more rapidly to serve the needs of personalized medicine and clinical studies. This enhancement was mainly due to the increased ability of the biobank to provide samples to research groups without any risk of leaving insufficient sample volumes for the care of the donor.
Introduction
T
The international trend to replace conventional cervical cytology with liquid-based cytology (LBC) provides a safer and faster diagnostic method that also preserves the cellular samples. This ensures the stability of large amounts of high quality DNA, RNA, proteins, and intact cells for new cytology research studies and clinical tests. Thus, LBC is the recommended method for screening and testing for human papilloma virus (HPV) in Sweden. 2
The SCCB has been developed and implemented in different county councils with financial support from the BBMRI.se (Biobanking and Molecular Resource Infrastructure of Sweden, www.bbmri.se). The BBMRI.se is funded by the Swedish Research Council and is situated at the Karolinska Institute in Stockholm.
The goal of the SCCB has been to: i) create a national platform for the biobanking of gynecological cell samples in spite of the regional differences; ii) ensure the performance of high-throughput sample handling for population-based cohorts; and iii) act as a hospital-integrated biobank for research purposes. Since August 2011, the SCCB has adopted a unified sample handling and storage process for processing and archiving liquid-based gynecological samples. 1 The cervical cells are collected into sample containers with PreservCyt®, a methanol-based medium (ThinPrep-TP). They are then cycled from clinical laboratories into biobanks for processing after a diagnosis has been determined. The biobank uses a robotic system in which barcoded sample jars are moved to a decapper and scanned. Four milliliters of each sample are aspirated in a bi-directional manner and dispensed into corresponding intermediate tubes. To ensure reliable, multipoint dispensation of the buffer, aspiration of a total of 10 aspiration points is performed for flat bottom sample containers (TP). A 30 minute sedimentation period, which is equivalent to gentle centrifugation, fits nicely into the workflow, preventing periods of inactivity. Finally, a 300 μL aliquot of cells is dispensed through a one-point aspiration into the final storage vials in a 96-well microplate (0.5 mL Tracker 2D in Loborack-96w low cover, MPW52337BC3, Nordic Biolabs AB, Tä by, Sweden). The barcodes are compatible with the automation software and the biobank IT system Cytology-LIMS (Laboratory Information Management System). This provides full sample tracking throughout the entire workflow. The final storage plate location is identified by the LIMS system, which controls a spacious electronic freezer. The cells are preserved at −25°C. 1 Cervical cells stored from −25°C to −30°C without freezing provide a significant advantage in cytodiagnostics, as slides can be generated repeatedly (using different staining methods if required) without exposing the sample to freeze-thaw cycling. This process maintains good cell morphology.
Deficiency of Sample Handling and Storage Processes
The high-throughput system and the process for handling the gynecological cell samples have been validated over two different periods over the course of nearly 1 year. During this time, deficiencies in handling and storage have become evident. These deficiencies contradict the guidelines of the SCCB, as the stored cytological samples in the biobank and the data derived from those samples will be used for the clinical care of patients and will be provided to multiple researchers for multiple purposes.
Unsatisfactory cell yields
In this study, cell yield was defined as the change in cell quantity expressed as a percentage. To estimate the performance of the entire high-throughput system, sample handling processes and cell yield were considered in three stages: i) the cell yield was calculated as the amount of cells in the intermediate tubes compared with the corresponding sample jars (TPs) before processing (tube to TP); ii) the cell yield in the storage vials was compared to the corresponding intermediate tubes after dispensing (vial to tube); and iii) the cell yield in the storage vials was compared with the corresponding sample jar before processing (vial to TP).
Before the start of the robotic process, 15×4 consecutive TPs from four runs were randomly selected. The cell concentration per mL was measured using a Bürker chamber and the cells were counted microscopically. The total numbers of cells in suspension in the TP containers were calculated. Subsequently, the aliquoting process was initiated, in which 4 mL volumes of the equivalent cell suspensions from the bottoms of the TP containers were aspirated and transferred into the intermediate tubes. The process was then paused, and the numbers of cells in the 4 mL volumes were calculated, similar to the calculations of cell suspension in the TP containers. The tubes were then returned to the robot worktable. After 30 minutes of sedimentation, the samples were aliquoted and prepared for the storage process. Corresponding storage vials containing 300 μL of cell suspension were selected and the cell concentration per mL and the total cell number in each aliquot were counted, as described above. During the validation, the cell suspensions in the containers were gently aspirated a number of times through a fine Pasteur pipette to disrupt the cell clumps. The cell yields in the intermediate tubes and storage vials were calculated by dividing the total number of cells by the number of cells in the corresponding TP container.
Our estimations of cell yield were not promising, particularly in the storage volumes. Deficiencies in the cell yield have been shown to result from low amounts of cells in the storage volumes. The cell yields were insufficient for clinical laboratory tests, donor medical needs, and research. The results are shown in Table 1.
Results show a compression of the cell yield associated with the number of selected samples. All values were reported as the percentage of mean±standard deviation (mean±SD%). (★), shows the significant enhancement of cell yield in storage vial after procedural improvement (t=2.339, p<0.05).
Unsatisfactory storage volumes
The storage volume used in the destination vials was 300 μL. In combination with the low cell yield, the system failed to cover the needs of the applicants to the biobank. Consequently, this deficiency was not in line with the purpose of the SCCB.
Improvement of the Handling and Storage Process
New aspiration technologies were developed to improve the quality of the SCCB sample handling and storage protocol and to ensure that the modified procedure will be sufficient for its intended purpose. These innovations improved the throughput of the process and increased both the cell yield and the volume of the storage samples from each individual.
Type of dispensing tips
Different types, sizes, and volumes of disposable tips were evaluated for precise automated pipetting during liquid handling. Consequently, the former pipette tips (1000 μL disposable tips for liquid handling robot, Tecan Cat#310612512, Germany) were replaced with a model with an aperture that has a diameter three times larger than that of the previous model (Thermoscientific, pure 1000 μL, Genomic, Cat#904-251GLOT13201265). The application modules were modified for the pipetting requirements, leading to a reduction in the tip blockage due to cell clumps and other materials in the cell suspension. This improved the aspiration capacity of the system.
Improving aspiration patterns in the sample containers
A variety of multipoint aspiration patterns were adopted for dispensing the buffer from the flat bottom of the TP container into the intermediate tubes. These techniques were designed to reliably aspirate and dispense a high number of cells from the TP into the intermediate tubes. These techniques were assessed and evaluated by running the entire automatic aliquoting system. The bi-directional aspiration points were upgraded from 10 to 24 multipoints in a zigzag pattern in the flat bottom of the container. The application modules were then modified for the dispensing requirements of the 4 mL cell suspensions from flat bottom TP into the intermediate tubes.
Improving aspiration patterns in the intermediate tubes
Unique and innovative aspiration technologies were required to improve the cell yield limitations of the storage vials. The method that met these demands was based on a multipoint aspiration pattern from the bottom of the intermediate tubes, followed by transfer into the storage vials. A single three-point aspiration technique was incorporated into the liquid-handling systems, enabling highly accurate dispensation of high-density cell populations from the bottom of intermediate tubes into the storage vials of the microplates. The application modules were then modified for the high-throughput pipetting requirements.
Improvement of the high-throughput storage volume and storage format
To achieve high-performance handling and storage processes at the SCCB, we attempted to increase the final storage volume. This new approach involved a multipoint aspiration pattern from the bottom of the TP and the intermediate tubes to allow for a larger volume with a dense cell population to be accurately dispensed into the storage vials. A range of different destination volumes were assessed. The most suitable volume for high performance storage was 600 μL. Consequently, the enhancement of the target volume required the replacement of the 96-well storage microplate format with a corresponding format volume. A standard 96-well format served this purpose well (0.75 mL Tracker 2D in Loborack-96w low cover, MPW52020BC3, Nordic Biolabs AB, Tä by, Sweden). The application modules of the robotic system and the LIMS were then modified to meet the requirements of the system.
Improvement of time spent during the protocol
Turnaround time is one of the most noticeable aspects of clinical laboratory service and is often used as a key laboratory performance indicator. The SCCB is a component of the clinical laboratories. The SCCB aims to provide a consolidated source of benchmark data so that laboratories can ensure timeliness for clinical outcomes that are of high quality. Therefore, an automatic startup control program was integrated with the robot application module, allowing for the automatic startup of the sample handling process. The startup time was a regulated step in a sequence of connected stages according to the laboratory workflow. The turnaround time after intervention was lengthened by 10 min for the entire handling process.
Over the course of improving and validating the performance of the new handling and storage protocols, a number of limitations in the automated high-throughput systems were revealed. These limitations resulted from the tradeoff between the turnaround time in the clinical laboratories and the goal of obtaining the maximum amount of cells in the storage vials. In our system, the robotic aspiration procedure failed to collect the highest number of cells from the bottom of TP and the intermediate tubes. Enriching the number of collected cells required extended aspiration times that disrupted the laboratory workflow. Consequently, a certain number of the cells were continually lost. An overview of the workflow involving the process for handling cytological cell samples from the biorepository is shown in Figure 1.
A summary of work flow that illustrates the cytological cell sample handling and storage process. (★), the points at which interventions have been implemented. (▭), indicating the process and location. (), sample entry and exit points. (⋄), directed sample aspiration and transfer. (→), information flow in direction of arrowhead.
Subsequent Validation Assays for Sample Handling and Storage
An improved sample handling and storage protocol was adopted by the SCCB and extensively tested in a series of validation studies. For each modification of the sample handling and storage process, we performed a high-throughput cell yield validation analysis. After process optimization, 15×4 consecutive TPs in four runs were randomly selected and the cell yield validation assay was performed as previously described. The difference in the cell yield in the storage vials (vial:TP) demonstrated the impact of the modification. Statistical significance was calculated before and after the modification. The paired Student's t-test was used for statistical analysis. Differences were considered significant if p was <0.05. All values were reported as the percentage of the mean±standard deviation (mean±SD%).
The results for cell yield before and after process modification are shown as the mean±SD% in Table 1. The standard deviation associated with the mean values is provided in the same table. The t-test confirmed that a significantly higher numbers of cells were transferred from the TPs to the storage vials after our modification to a 600 μL volume with 34±1.55% (mean±SD%) compared to 8±4.2% before (300 μL final volume) procedural improvement (t=2.339, p<0.05). The cell concentration (median 80×104/mL) in the 600 μL volume ranged from 392×104/mL to 6×104/mL. This wide range for final cell concentration may be explained by the influence of different cell population densities from clinical sampling and the existence of cell clump that blocked the dispensing tips, resulting in deficient aspiration.
To assess the quality of the cells after the modifications, we performed HPV genotyping analysis. In the HPV validation analysis, 100 LBC samples were handled using the improved aliquoting method. The samples were archived at −25C° for 6–11 months in the SCCB. The samples were then randomly selected and anonymized by the director of the biobank, at which time they were designated as biobank samples. Volumes of 100 μL were extracted from the biobank samples. These volumes were suspended in PreservCyt fixative (ThinPrep, 20 mL). All extractions were registered in the Cervical Cytology Biobank-LIMS. The test samples were transferred to the clinical virology laboratory at the Karolinska University Hospital at Huddinge-Stockholm, where they were assessed according to the laboratory procedures. The Cobas 4800 HPV Test (Roche Molecular Systems) was utilized. 3 The Cobas 4800 HPV is a multiplex assay based on the real-time PCR that identifies HPV-16 and HPV-18 with concurrent detection of twelve other HPV types (HPV-31, −33, −35, −39, −45, −51, −52, −56, −58, −59, −66, and −68). The biobank samples were analyzed for the fourteen oncogenic HPV subtypes by multiplexed real-time PCR. They were genotyped by a senior virologist. The results were then compared with the original patient samples that had been analyzed with the same method and archived in the laboratory. All 100 LBC biobank samples provided the same results as the original patient samples. Thus, validation of the preserved cells after improvement of the biopreservation procedures demonstrated that the stored cells were stable and confirmed that the changes to the workflow were not detrimental to the preserved samples. These results are representative of the quality of the biobanked samples only. Table 2 shows the quantification of the HPV genotyping results of the biobanked samples after the modifications to the storage and handling procedure.
All 100 LBC biobank's samples provided the same results as the original patient samples. The presented results are the outcome of preserved samples after improvement of biopreservation procedures. Negative; defined as the absence of HPV genotype according to the acquired results from Cobas 4800. Some samples contain more than one genotype.
Conclusion
After considering the protocol limitations and up-to-date-technical restrictions, performing a comprehensive review of related articles and engaging in discussions with researchers in the field, we designed an improved procedure using a high-throughput system to significantly enhance the cell content, integrity, and usability in biobanked preparations. The increase in the final cell concentration and the volume in the storage vials demonstrated the positive impact of the operational interventions and improvements to the biobank procedure. The enhancement of the biobank capacity can be explained by the fact that the biobank is capable of providing samples to multiple research groups while still retaining samples for the donor and preserving the donors' rights to access their samples when required.
The new sample handling and storage protocol was gradually disseminated to all of the clinical laboratory centers collaborating with the SCCB. These centers now use the automated high-throughput liquid-handling systems. This shows that the improved processes have been integrated into health care delivery.
Footnotes
Acknowledgment
The authors thank the BBMRI.se for the financial support during the improvement of the SCCB workflow.
Author Disclosure Statement
No competing financial interests exist.