PD-L1+ CD49f+ CD133+ Circulating tumor cells predict outcome of patients with vulvar or cervical cancer after radio- and chemoradiotherapy

Abstract

Background: Monitoring individual therapy responses of patients with cancer represents a major clinical challenge providing the basis to early identify metastases and cancer relapse. We previously demonstrated that radio- or chemoradiotherapy affects the systemic cellular milieu of patients with vulvar or cervical cancer and creates individual post-therapeutic environments associated with cancer relapse. Circulating tumor cells (CTCs) in the systemic milieu are related to metastases and relapse; however, their quantitative and phenotypic characteristics during therapy of patients with vulvar and cervical cancer are still unknown.

Methods: In this prospective, longitudinal study, we verified the presence of CTCs via immunofluorescence and systemically characterized CTCs by flow cytometry from the blood of 40 patients with vulvar and 115 patients with cervical cancer receiving surgery, adjuvant radiotherapy (aRT), chemoradiotherapy (aCRT) or primary chemoradiotherapy (pCRT) and linked the presence of different CTC subpopulations with individual outcome of disease.

Results: Pre-therapeutic cytokeratin⁺ CD45^- CTC numbers significantly correlated with tumor FIGO stages, lymph node metastases and relapse. While surgery only did not significantly alter CTC occurrence, aRT and aCRT as well as pCRT differentially decreased or increased CTCs in patients with both tumor entities compared to baseline levels. Therapy-mediated increased CTC numbers were directly linked with subsequent cancer recurrence on follow-up. Phenotypic characterization of CTCs revealed enhanced expression of the stem cell marker CD133 as well as the integrin α6 (CD49f) after aRT, aCRT and pCRT. Furthermore, the aRT, aCRT and pCRT cohorts exhibited increased proportions of Programmed Cell Death Protein Ligand (PD-L1) expressing cells among post-therapeutic CTCs. Notably, post-therapeutic PD-L1⁺ CD49f⁺ CD133⁺ numbers ≥ 5/ml in patients with vulvar cancer and ≥ 2/ml in patients with cervical cancer were associated with reduced recurrence-free survival on follow-up.

Conclusion: Our study unravels individual therapy-induced changes in CTC phenotypic characteristics and occurrence in the patients' blood and their association with cancer relapse. Our results may help to explain differences in the individual courses of disease of patients with vulvar and cervical cancer and suggest PD-L1, CD49f and CD133 as targets for immunotherapy in vulvar and cervical cancer.

Keywords: Cervical cancer; Chemoradiotherapy; Circulating tumor cells; Radiotherapy; Vulvar cancer.

Fig. 1

Cytokeratin⁺ CD45⁻ CTCs in the blood of patients with vulvar or cervical cancer before therapy. (A, B) Cells from PBMCs layers of 40 patients with vulvar cancer, 40 patients with cervical cancer and 40 female age-matched healthy controls (HC), respectively were analyzed by IF for CD45 (green) and pan-cytokeratin (red) expression. (A) shows representative pictures of a probe from a patients with cervical cancer. (B) Numbers of cytokeratin⁺ CD45⁻ cells/ml were evaluated for patients with vulvar cancer (orange dots), patients with cervical cancer (green dots) and HC (grey dots), respectively. (C) Cells from PBMCs were evaluated by flow cytometry for CD45 and pan-cytokeratin expression. Numbers of cytokeratin⁺ CD45⁻ cells/ml were evaluated for 40 patients with vulvar cancer (orange dots), 115 patients with cervical cancer (green dots) and 40 or 115 HC (grey dots), respectively. (D) Correlation of cytokeratin⁺ CD45⁻ cell numbers/ml after IF and flow cytometry of patients with vulvar cancer (orange dots) and patients with cervical cancer (green dots). (E) Correlation of cytokeratin⁺ CD45⁻ cell numbers/ml after flow cytometry of patients with vulvar cancer (orange dots) and patients with cervical cancer (green dots) with tumor FIGO stages. (F, G) Numbers of cytokeratin⁺ CD45⁻ cells/ml after flow cytometry in 40 patients with vulvar cancer (orange dots) and 115 or 87 patients with cervical cancer, respectively (green dots) (F) with (blue background) and without (grey background) lymph node metastases, (G) with (purple background) and without (grey background) cancer recurrence. P value according to the nonparametric Mann-Whitney U-test (B, C, F, G), nonparametric Kruskal-Wallis test (E) or Spearman rank correlation with linear regression (D)

Fig. 2

Impact of surgery, radiotherapy and chemoradiotherapy on CTCs in patients with vulvar or cervical cancer. (A) Cells from PBMCs of 40 patients with vulvar cancer (n = 22 receiving surgery, grey dots or n = 18 receiving aRT, orange dots; n = 3 patients with concurrent chemotherapy; black edged dots) and (B) cells from PBMCs of 115 patients with cervical cancer (n = 57 receiving surgery, grey dots, n = 36 receiving aCRT, blue dots or n = 22 receiving pCRT, red dots) were analyzed for CD45 and pan-cytokeratin expression by flow cytometry. (C, D, E, F) Numbers of post-therapeutic cytokeratin⁺ CD45⁻ cells/ml in patients with vulvar (orange dots) or cervical cancer (green dots) (C, E) with (purple background) and without (grey background) cancer recurrence, (D) before and after surgery (circles) or aRT (triangles) and (F) before and after aCRT (squares) and pCRT (diamond) with (purple lines) and without cancer recurrence (grey lines). P value according to the nonparametric matched pairs Wilcoxon test or paired Friedeman test (A, B, D, F) or Mann-Whitney U-test (C, E). (G) Recurrence-free survival (RFS) of 40 patients with vulvar cancer with pre-therapeutic numbers of cytokeratin⁺ CD45⁻ cells < 7/ml (dotted grey line) or ≥ 7/ml (grey line) as well as with post-therapeutic numbers < 7/ml (dotted orange line) or ≥ 7/ml (orange line). Median RFS was 35 months for pre-therapeutic numbers < 7/ml or 27 months for post-therapeutic numbers ≥ 7/ml. Comparison of survival analysis was performed using log-rank (Mantel-Cox) test; chi-square: 17.18 or 22.65, respectively. P < 0.0001. (H) RFS of 50 patients with cervical cancer with pre-therapeutic numbers of cytokeratin⁺ CD45⁻ cells < 6/ml (dotted grey line) or ≥ 6/ml (grey line) as well as with post-therapeutic numbers < 3/ml (dotted green line) or ≥ 3/ml (green line). Median RFS was 12 months for pre-therapeutic numbers ≥ 6/ml or 10 months for post-therapeutic numbers ≥ 3/ml. Comparison of survival analysis was performed using log-rank (Mantel-Cox) test; chi-square: 13.00 or 28.60, respectively, P = 0.0002 or P < 0.0001

Fig. 3

Phenotypic characterization of cytokeratin⁺ CD45⁻ cells during surgery, radiotherapy and chemoradiotherapy. Cells from PBMCs of 40 patients with vulvar cancer who received surgery only (n = 22; grey dots) or aRT (n = 18, orange dots; n = 3 patients with concurrent chemotherapy; black edged dots) and 115 patients with cervical cancer who received surgery only (n = 57; grey dots), aCRT (n = 36; blue dots) or pCRT (n = 22; red dots) were analyzed by flow cytometry for (A, B) frequencies of CD133⁺ cytokeratin⁺ CD45⁻ cells, (C, D) frequencies of SOX2⁺ CD133⁺ cytokeratin⁺ CD45⁻ cells, (E, F) frequencies of CD49f⁺ CD133⁺ cytokeratin⁺ CD45⁻ cells, (G, H) frequencies of PD-L1⁺ CD49f⁺ CD133⁺ cytokeratin⁺ CD45⁻ cells. P value according to the nonparametric Wilcoxon test or nonparametric Mann-Whitney U-test

Fig. 4

Frequencies and numbers of CTC subpopulations after therapy and their association with cancer relapse. Cells from PBMCs of 40 patients with vulvar cancer (orange dots) who received surgery only (n = 22; circles) or aRT (n = 18, triangles; n = 3 patients with concurrent chemotherapy; black edged triangles) and 50 patients with cervical cancer (green dots) who received aCRT (n = 36; squares) or pCRT (n = 22; diamonds) were analyzed for frequencies and numbers of (A, B) CD133⁺cytokeratin⁺ CD45⁻ cells, (C, D) frequencies and numbers of CD49f⁺ CD133⁺ cytokeratin⁺ CD45⁻ cells, (E, F, G, H) frequencies and numbers of PD-L1⁺ CD49f⁺ CD133⁺ cytokeratin⁺ CD45⁻ cells by flow cytometry. Frequencies and numbers were depicted for patients with (purple background) and without (grey background) relapse. P value according to the nonparametric Mann-Whitney U-test (A, B, C, D, E left, F left, G left, H left) or nonparametric Wilcoxon matched-pairs test (E right, F right, G right, H right)

Fig. 5

Association of numbers of CTC subpopulations with RFS of patients with vulvar or cervical cancer. (A) Recurrence-free survival (RFS) of 40 patients with vulvar cancer who received surgery (n = 22) or aRT (n = 18) was evaluated for a cohort with post-therapeutic numbers of CD133⁺ CTCs < 14/ml (light blue line) or ≥ 14/ml (blue line). Median RFS was 27 months for CTCs ≥ 14/ml. Comparison of survival analysis was performed using log-rank (Mantel-Cox) test; chi-square: 14.21, P = 0.0002. (B) RFS was evaluated for post-therapeutic numbers of CD49f⁺ CD133⁺ CTCs < 5/ml (light red line) or ≥ 5/ml (red line). Median RFS was 27 months for CTCs ≥ 5/ml. Comparison of survival analysis was performed using log-rank (Mantel-Cox) test; chi-square: 24.60, P < 0.0001. (C) RFS was evaluated for post-therapeutic numbers of PD-L1⁺ CD49f⁺ CD133⁺ CTCs < 5/ml (light purple line) or ≥ 5/ml (purple line). Median RFS was 25.5 months for CTCs ≥ 5/ml. Comparison of survival analysis was performed using log-rank (Mantel-Cox) test; chi-square: 41.26, P < 0.0001. (D) RFS for 50 patients with cervical cancer who received aCRT (n = 36) or pCRT (n = 22) was evaluated for post-therapeutic numbers of CD133⁺ CTCs < 2/ml (light blue line) or ≥ 2/ml (blue line). Median RFS was 10 months for CTCs ≥ 2/ml. Comparison of survival analysis was performed using log-rank (Mantel-Cox) test; chi-square: 23.60, P < 0.0001. (E) RFS was evaluated for post-therapeutic numbers of CD49f⁺ CD133⁺ CTCs < 2/ml (light red line) or ≥ 2/ml (red line). Median RFS was 10 months for CTCs ≥ 2/ml. Comparison of survival analysis was performed using log-rank (Mantel-Cox) test; chi-square: 17.83, P < 0.0001. (F) RFS was evaluated for post-therapeutic numbers of PD-L1⁺ CD49f⁺ CD133⁺ CTCs < 2/ml (light purple line) or ≥ 2/ml (purple line). Median RFS was 10 months for CTCs ≥ 2/ml. Comparison of survival analysis was performed using log-rank (Mantel-Cox) test; chi-square: 34.33, P < 0.0001

Fig. 6

Schematic presentation of CTC subpopulations during therapy and vulvar cancer relapse. Adjuvant radio- or chemoradiotherapy increased occurrence of PD-L1⁺-CD49f⁺-CD133⁺ CTCs. Increased post-therapeutic numbers of PD-L1⁺-CD49f⁺-CD133⁺ CTCs were associated with vulvar or cervical cancer recurrence (scheme was generated with the BioRender software)