Increased circulating tumor cell capture from leukapheresis material - Getting the most out of it
Michiel Stevens is a PhD student in the department Medical Cell Biophysics. (Co)Promotors are prof.dr. L.W.M.M. Terstappen and dr.ir. F.A.W. Coumans from the faculty of Science & Technology.
Circulating Tumor Cells, or CTCs, can be used to monitor disease progression or treatment response. However, due to their rarity, the detection and characterization of CTCs is challenging. In most cases, an enrichment of tumor cells is used, after which CTCs are identified among the enriched cells using immunofluorescent markers. The CellSearch system is the most prominent example of this approach, where immunomagnetic particles targeted against the EpCAM antigen are used to selectively capture the CTCs present in a 7.5 mL sample of blood. In 64% of 7.5 mL blood samples from metastatic carcinoma patients, the number of CTCs found is <1 [1], which hampers their utility in research, indicating an increase in the number of available CTCs is needed. In an immunomagnetic enrichment procedure, the characteristics of the used magnetic particle in combination with the expression of the targeted antigen determine the number of particles that are attached to the tumor cell. As the EpCAM expression of CTCs is generally low, an increase in CTCs detection could be made by allowing the
capture of cells with low numbers of particles attached. Another way to increase the number of CTCs found is by interrogating a much larger volume of blood, using Diagnostic LeukApheresis (DLA) [2]. In this procedure the mononuclear cell fraction of the blood (also containing the CTCs) is separated, while the other blood components are returned to the patient. In this way, the complete blood volume of a patient can be processed. The resulting DLA samples are about 80 mL in volume and contain approximately 180-fold more white blood cells (WBCs) than a 7.5 mL blood sample, making the CellSearch unfit for their processing. In lieu of better options, researchers have used small aliquots consisting of 2-5% of the total DLA product reaching a sixfold increase in the number of detected CTCs. The use of larger aliquots of DLA product leads to an excessive number of WBCs being co-enriched, which impedes both efficient (semi-)automated CTC identification as well as the needed single-cell isolation for CTC characterization. Processing many tests is unfeasible due to cost and time constraints.
In a regular tube of blood, 99% of the cells are red blood cells, while their number is tenfold lower in a leukapheresis sample. We have shown that as a result, also the used sample volume and amount of magnetic particles used can be reduced by tenfold without any loss of CTCs, leading to a 2.7-fold reduction in co-enriched WBC. The efficiency of an immunomagnetic separation, be it from blood or DLA material, is determined by the number and magnetic properties of the particles attached, as well as the used magnet configuration. With the use of a flow chamber, larger samples size can easily be processed. Additionally, the cells can be brought into close proximity to the magnets, where the magnetic force is the strongest. By optimizing the used magnetic array, the exerted force can be maximized, resulting in an improved ability to separate low EpCAM-expressing cells. With the use of this flow chamber and optimized magnetic array, we performed the separation of CTCs from DLA product using a tenfold lower concentration of magnetic particles, resulting in a cost reduction as well as a 3.2-fold reduction in co-enriched WBCs without any CTC loss. We have also shown that the
often-seen cell clumping in the resulting samples can be prevented by incubating the sample with DNase prior to immunomagnetic enrichment. With these improvements, using the reagents from a single CellSearch test, we have processed 10-fold larger aliquots of DLA material, amounting to ~20% of the total sample. This is the product obtained from more than 1 liter of blood, and resulted in a 44-fold increase in obtained CTCs compared to processing a tube of blood.
Despite a relative decrease in co-enriched WBC, the number of co-enriched WBC obtained from these large samples is still high, making segmentation and (semi-)automated identification challenging. Optimization and testing of the StarDist image segmentation method show that in DLA samples processed by CellSearch, an additional 20% of CTCs can be detected that are initially missed due to faulty segmentation. For the characterization of the found CTCs, these need to be extracted as single cells. We have in this work shown that for this step the magnetic labeling present on the cells from the used magnetic enrichment can be used. For this we make use of a very local magnetic field generated by a magnetic microneedle, with which we can pick up selected single cells. This approach can be used to pick-up single CTCs for isolation, or to precisely place them in a desired location. The improvements presented in this work, especially those allowing the enrichment of CTCs from large aliquots of DLA material, using low amounts of reagents while preventing cell clumping, create new possibilities for the (epi)genomic characterization as well as functional analysis. For these possibilities to be used in a clinical setting, further development and automation as well as availability of the technology is needed, which will need to be achieved through commercial application.
1. Allard, W.J.; Matera, J.; Miller, M.C.; Repollet, M.; Connelly, M.C.; Rao, C.; Tibbe, A.G.J.; Uhr, J.W.; Terstappen, L.W.M.M. Tumor Cells Circulate in the Peripheral Blood of All Major Carcinomas but Not in Healthy Subjects or Patients with Nonmalignant Diseases. Clin. Cancer Res. 2004, 10, 6897–6904, doi:10.1158/1078-0432.CCR-04-0378.
2. Fischer, J.C.; Niederacher, D.; Topp, S.A.; Honisch, E.; Schumacher, S.; Schmitz, N.; Föhrding, L.Z.; Vay, C.; Hoffmann, I.; Kasprowicz, N.S.; et al. Diagnostic Leukapheresis Enables Reliable Detection of Circulating Tumor Cells of Nonmetastatic Cancer Patients. Proc. Natl. Acad. Sci. U. S. A. 2013, 110, 16580–16585, doi:10.1073/pnas.1313594110.
