Analysis of tumor-associated cells in peripheral blood offers the advantage of a noninvasive approach to study changes in biomarkers associated with chemotherapy, tumor recurrence, or tumor treatment. However, many CTC technologies require the processing of fresh blood samples and are incompatible with cryopreserved samples, limiting the ability to perform retrospective studies on archived samples. This limitation may pose a challenge in multicenter clinical trials and require a central laboratory to process samples on the same day for the CTC assays. The assay reproducibility may be adversely affected by performing each fresh blood sample individually or by transportation of specimens between locations. Cryopreservation is one of the most commonly used methods for storage biologic specimens collected over time.
We optimized a cryopreservation work flow compatible with detecting tumor-associated cells in stored peripheral blood to overcome assay limitations that require processing of fresh blood. We showed the feasibility of CellSieve™ technology for isolation and identification of CTCs and CAMLs from cryopreserved samples. Our notable observations include (1) recovery of spiked tumor cells lines in cryopreserved peripheral blood using the CellSieve™ platform; (2) preservation of cellular morphologic characteristics and markers in cryopreserved samples from patients with metastatic RCC; and (3) preservation of CTC and CAML enumerations in samples cryopreserved up to 66 days at ?80 °C.
When evaluating our observations, we found limitations to the present investigation. First, we analyzed 1 CTC microfiltration platform, and the cryopreservation procedure may not be compatible with other antigen-dependent or antigen-independent cell capture methods. Second, although we evaluated breast, prostate, and kidney cancer cell lines, we analyzed CTCs and CAMLs only in samples from patients with kidney cancer. The DMSO concentration or Ficoll separation may not be compatible with all types of cancer diseases, and further study is required to determine whether the cryopreservation procedure is suitable for other cancers. Third, we did not extend the study beyond 60 days, and the impact of storage beyond this time point may or may not have effect on enumeration assays.
Cryopreservation of PBMCs allows phenotypical and genetic assays to be performed in batches at a central laboratory and minimize assay variation. We have demonstrated that the CellSieve™ system is compatible with Ficoll separation and cryopreservation for detection of CTCs and CAMLs. Investigators have previously reported that CTCs can be recovered from cryopreserved PBMC samples [7, 9, 16]; however, it was unclear whether CAMLs can be recovered from cryopreserved PBMC sample. To our knowledge, we are the first to report that CAMLs can be separated by Ficoll method, cryopreserved for future use, and retrieved after thawing samples stored up to 60 days at ?80 °C. In our Ficoll separation, CAMLs segregate to the interphase between Ficoll and plasma, where PBMCs and CTCs are located. CAMLs may have similar density to other mononuclear cells, such as lymphocytes and CTCs, allowing them to be enriched through Ficoll separation, and use of a microfiltration system can isolate cells on the basis of pore size. These enriched cells are suspended in cryoprotectant medium, which contains DMSO to reduce cell lysis associated with ice crystal formation. Preliminary analysis on the signal intensities of CKs between fresh and frozen CAMLs are similar, however, the CAML sample size is not large enough to draw definitive conclusions.
We observed from the matched samples that the freeze/thaw process has no adverse effect on morphologic characteristics and the biomarkers of RCC-associated cells. Although we did not observe statistically significant changes in enumeration of spiked breast and prostate cell lines, further study is required to evaluate whether our cryopreservation process is compatible with other tumor types. Furthermore, during the course of investigation, we also found denucleated CAMLs in frozen samples from a patient with RCC, consistent to fresh samples. The denucleated CAMLs resemble typical CAMLs closely in morphologic characteristics and staining patterns except for a lack of nuclei. However, the importance of these structures is undefined.