Ovarian cancer (OC) is one of the leading causes of death among gynecological tumors and molecular characterization of single circulating tumor cells (sCTCs) in blood is one way to gain insight into CTC genotypes and their impact on disease progression. Here, we combined immunomagnetic enrichment and image-based sCTC sorting for phenotypic and genotypic analysis of sCTCs in OC follow-up to evaluate sCTC heterogeneity.

Blood from patients with high-grade serous OC (HGSOC) before treatment (b.tre.), after chemotherapy (a.CTX) and after Avastin treatment (a.Av.) was enriched for CTCs by density gradient centrifugation for further application of MACS^® (Miltenyi Biotec, Germany) with CD45 and CD235a antibodies coupled to MACS^® MicroBeads. Image-based sCTC sorting was performed on the DEPArray^TM NxT system using two immunofluorescence protocols, the ERCC1-protocol, staining for Hoechst, cytokeratin (CK), ERCC1, CD11b, CD45 and the FRα-protocol, staining for Hoechst, CK, folate receptor alpha (FRα), vimentin (Vim) and CD45. Whole Genome Amplification (WGA) was performed in recovered single cells (SCs), including cell lysis, DNA digestion, adaptor ligation and amplification PCR for subsequent sequencing. Copy number alterations (CNAs) were analyzed by Ampli1^TM LowPass (Menarini Silicon Biosystems; MSB) and single nucleotide variants (SNVs) by the Ampli 1^TM OncoSeek panel (MSB). In addition, 21 matched primary tumors (PT) were also investigated for whole genome CNAs.

Examination of 13 HGSOC patients b.tre. with the ERCC1 protocol, detected Type A-cells (Hoechst^pos, ERCC1^pos, CD45^neg, CD11b^pos, CK^pos) with epithelial characteristics, Type B-cells (Hoechst^pos, ERCC1^pos, CD45^neg, CD11b^pos, CK^neg) with epithelial-like characteristics and Type C-cells (Hoechst^pos, ERCC1^pos, CD45^neg, CD11b^neg, CK^neg) with potentially mesenchymal characteristics. In total, five (38.5%) patients were identified with sCTCs, mainly Type A-cells cells (80%). Interestingly, only 30.8% of the analyzed Type A-cells had their aberrant character confirmed by genome-wide CNAs. A large number of Type B- and C-cells were identified in each patient but their aberrant character was only confirmed in 6.25% and 4.76% of cases, respectively. Therefore, further establishment of sCTC identification markers were essential to improve detection and allow insight into tumor evolution on a sCTC resolution.

Establishment of the FRα-protocol in two phenotypically distinct OC cell lines and healthy donor samples allowed phenotypic assessment of different cells over the treatment period. Phenotypic evaluation of 42 HGSOC patient samples b.tre., 23 a.CTX and 11 a.Av. treatment showed a significant (p= 0.0205) reduction of FRα-positive cells after CTX. Interestingly, cells with high nuclear staining but no target antigen expression were found in 88% (37/42) of patients and expanded significantly (p=0.002) during treatment.

Genotype assessment of sCTCs in all 50 HGSOC patients revealed that only 6.58% (19/289) of all analyzed SCs, from 26% (13/50) of patients, were sCTCs. Whole genome CNA analysis in all sCTCs b.tre. revealed enriched CNAs in chromosomes 2, 7 and 12. In sCTCs a.CTX and a.Av. treatment, CNA dynamics highlighted the adaptive potential of the homologous recombination (HR) pathway and suggested further investigation of MAPK signaling in treatment resistant HGSOC. TP53 variants were most frequently detected and persistent CNAs in the CDK4 oncogene as well as emerging CNAs in the ALK oncogene were found in sCTCs from two patients.

LowPass CNA comparison between sCTCs and the PT was feasible in nine patients and revealed persistent PT features in sCTCs at different sampling time points. In addition, two patients with sCTCs at different time points showed a similar persistent PT feature in sCTCs in chromosome 8, covering the MYC gene.

In conclusion, sCTCs were identified with both protocols and demonstrated inter- and intrapatient heterogeneity of sCTCs. Matched sCTCs and PT data showed a certain amount of similar CNAs, however, our low percentage of sCTC detection rate suggests that higher patient numbers are needed to confirm our findings. Nevertheless, the results of this study have highlighted several genomic features for further analysis in understanding the role of HGSOC CTCs in disease progression.