Purification of viable peripheral blood mononuclear cells for biobanking using a robotized liquid handling workstation

The foundation of research biobanks after completion of the human genome project has been a critical step for entry in the postgenomic era. Indeed, biobanks provide high-quality samples for the setting up of research protocols dedicated to the translation of important discoveries coming from basic research laboratories to clinical practice [1]. One example of their usefulness comes from the generation of The Cancer Genome Atlas (TCGA) [2]. Because of human specimens with associated clinical data provided by biobanks, it is possible to analyze large cohorts of over 30 tumors with high-throughput genome sequencing. The TCGA led to the identification of several novel molecular alterations in cancer, allowing the classification of tumor subtypes according to distinct genomic alterations and, consequently, a precision medicine approach for patient care [3]. Despite the general consensus regarding the critical role of biobanks in modern research and the growing number of newly founded institutional biobanks, there is a high level of heterogeneity between biobanks, with particular reference to the procedures used for sample processing [4]. In 2018, the International Standardization Organization (ISO), in association with international biobank scientific organizations, such as the International Society for Biological and Environmental Repositories (ISBER) [5] and the European Biobanking and BioMolecular Resources Research Infrastructure (BBMRI) [6], provided the ISO/DIS 20387 standard [7], which is specifically dedicated to the biobanking process. The aim of the ISO20387 is to standardize the processing of biosamples across countries by defining important requirements for correct sample processing. In this context, our work aimed to contribute to the standardization of sample processing by presenting an innovative automatic workflow setup on the Freedom EVO 150 automatic liquid handling workstation. To date, different companies, such as Hamilton (https://​www.​hamiltoncompany.​com), Qiagen (https://​www.​qiagen.​com/​us/​products/​instruments-and-automation/​?​akamai-feo=​off), Brooks Life Sciences (https://​www.​brookslifescienc​es.​com/​about-brooks-life-sciences) and Tecan (https://​www.​tecan.​com/​), provide solutions for the customization of automatic procedures for biological sample processing. Due to their versatility, automated liquid handling systems are commonly used for different laboratory applications [8‐10], particularly when the pipetting steps must be as accurate and precise as possible, as in the cases of biopharmaceutical products preparation [11, 12], serial dilutions, and reagent transfers [13, 14]. In this paper, we propose these complex technological systems for the automatic purification of peripheral blood mononuclear cells (PBMCs) from whole blood collected in ethylenediaminetetraacetic acid (EDTA) tubes. Because the density gradient purification of PBMCs is a procedure with high operator-dependent variability, we believe that the methodology presented in this paper could be useful for the standardization of PBMC isolation in biobanking.

The isolation of mononuclear cells from peripheral blood is a protocol routinely used in biobanking [17]. Notably, PBMCs include a heterogeneous population of cells that are mainly involved in the body’s immune defense, as in the case of B, T, NK and NKT lymphocytes, monocytes and dendritic cells [18]. The study of these immune cell subsets can be of interest in different fields of medical research. For example, T and B cells have been shown to be activated during neuroinflammatory autoimmune diseases such as Parkinson’s disease [19] and multiple sclerosis [20]; circulating tumor cells have been detected in cryopreserved PBMCs from patients affected with sarcoma [21] and are intensively studied in breast, lung, prostate and colorectal cancers [22]. Despite the broad field of applications, the isolation of PBMCs from peripheral blood is a laboratory procedure still affected by high operator-dependent variability. The dilution, stratification, and collection of mononuclear cells from the density gradient medium are some of the critical steps that affect standardization. Over time, different attempts have been made to standardize the PBMC isolation procedure; however, few reports were able to present the automatization of the entire procedure. In 2012, Mathay et al. [23] developed an automated process with the Tecan pipetting robot Freedom EVO 200 for extraction of the buffy coat (BC) from three subjects. The comparison with the automated method was performed by means of a complete blood count and DNA extraction. The authors combined BC extraction with upstream automated plasma aliquoting and downstream DNA extraction from BC in the same run. Regardless of the low number of subjects (n = 3) examined in their work, they were able, for the first time, to optimize the manual procedure for BC extraction on the Tecan system. In 2015, Hamot et al. [24] improved the workflow using the Tecan system again for the automated isolation of PBMCs from whole blood collected in BC CPT vacutainer tubes (citrate anticoagulant). The authors validated the automated procedure in terms of cell yield, viability, recovery, white blood cell subpopulation distribution, gene expression, and lymphoblastoid cell line transformation. Interestingly, they compared the performance of CPT tubes in isolating PBMCs with respect to the manual procedure performed using Ficoll^® and Leucosep^®.

In the present paper, in contrast to Hamot et al. [24], we developed an automated script for the isolation of PBMCs from whole blood collected in K3 EDTA vacutainer tubes using density gradient media (HiSep). In comparison to the work of Hamot et al., we did not use CPT tubes for the setup of the automated procedure to make the entire procedure cost-saving and feasible in all laboratories equipped with a robotized workstation similar to the Tecan Freedom EVO 150. The automatic procedure was validated by comparison to the manual procedure. Under our experimental conditions, we demonstrated that automation guarantees the same yields as the manual procedure in terms of the number of recovered lymphocytes and monocytes and the reduced contamination of neutrophils and red blood cells.

Furthermore, no significant differences were found between the two methods in terms of the percentages of lymphocyte subpopulations. The automatic separation of PBMCs also guarantees no operator-related variability, the identification and tracking of the sample in all protocol phases, and the recording of all instrument-specific parameters during isolation. Additionally, we ascertained that automatic processing guarantees sterile biological samples. In fact, after the automated procedure, isolated PBMCs were grown in complete medium for up to 10 days, and no contamination was observed (data not shown). Finally, it is important to consider that using the automatic workflow, it is possible to develop additional upstream or downstream procedures, such as the generation of plasma aliquots and the isolation of nucleic acids, respectively. The generation and optimization of automatic protocols using robotized liquid handling systems will be useful for ameliorating the standardization of biobanking procedures [25].

In our study, we presented an automatic process related to the isolation of PBMCs from peripheral blood in the workflow of our biobank. It has considerable advantages as mentioned above; however, technical and practical limitations are present. The first aspect concerns the lack of a blood level detection unit. Our automatic protocol was developed on the basis of spatial coordinates x, y, and z, in which the use of fixed volumes determine the specific spatial coordinate of the phase in which the PBMCs are present. For the right set-up of the x, y and z coordinates we were assisted by the manufacturer technical support both on-site (for the instrument pipeline set-up) and online (for the script development). During the blood stratification process, it was important to generate a homogeneous cellular suspension to be gently stratified over the density gradient media. This difficult problem was resolved by programming the instrument for doing the following operations: (1) add 1 mL of DPBS; (2) mix up and down for five times; (3) add a second 1 mL volume of DPBS; (4) mix up and down for five times; (5) immediately distribute over density gradient media. Also, during blood stratification, the flow rate was imposed to be as slow as possible (20 µL/s–1 mL Tip) for avoiding mixing of the diluted blood with the density media, especially during the first steps of stratification. Importantly, additional instrument components (still not available in our case) for script optimization could be useful. For example, the installation of a laser-based inspection unit for a more accurate determination of the x, y and z coordinates could be helpful, especially for the blood stratification as well as for identification and picking-up of the PBMC layer after centrifugation. The setting-up of this technical procedure required about eight months of internal test and close collaboration with the specialized instrumentation technician provided by the manufacturer. Subsequently, a two-week training period was necessary for all biobank operators to learn practical information (relating to the workstation such as replacing the tip slots, reagents, correct placement of blood samples in the loading position) and information relating to the machine management software. Currently, all the operators of our biobank are able to manage the instrument easily. Another critical aspect concerns the time needed for processing the blood samples; currently, our automatic script takes 45 min to process (from dilution to stratification) 12 blood samples.

Finally, although automation represents an approach used by many laboratories around the world, the costs related to instruments and consumables remain inaccessible to infrastructures with limited funds, especially in developing countries.

Our manuscript demonstrates that the automated isolation of PBMCs is feasible with an automated system and comparable to traditional manual procedures; we believe that automation is becoming essential in the biomedical and diagnostics research field because it will allow biobank networks to standardize biological sample processing, allowing a fair comparison of biological data analyses downstream. In this way, it will be possible to reach the overall goal of an international research framework that aims to facilitate access to human biological materials. The life cycle stages of biosamples [26] (intended as the collection, accession, acquisition, identification, preservation, long-term storage, quality control (QC), transport, and disposal of biomaterials) are a major source of heterogeneity, and process automation could represent a potential means of addressing these obstacles.