Purging human ovarian cortex of contaminating leukaemic cells by targeting the mitotic catastrophe signalling pathway

Abstract

Purpose: Is it possible to eliminate metastasised chronic myeloid leukaemia (CML) and acute myeloid leukaemia (AML) cells from ovarian cortex fragments by inhibition of Aurora B/C kinases (AURKB/C) without compromising ovarian tissue or follicles?

Methods: Human ovarian cortex tissue with experimentally induced tumour foci of CML, AML and primary cells of AML patients were exposed to a 24h treatment with 1 μM GSK1070916, an AURKB/C inhibitor, to eliminate malignant cells by invoking mitotic catastrophe. After treatment, the inhibitor was removed, followed by an additional culture period of 6 days to allow any remaining tumour cells to form new foci. Ovarian tissue integrity after treatment was analysed by four different assays. Appropriate controls were included in all experiments.

Results: Foci of metastasised CML and AML cells in ovarian cortex tissue were severely affected by a 24h ex vivo treatment with an AURKB/C inhibitor, leading to the formation of multi-nuclear syncytia and large-scale apoptosis. Ovarian tissue morphology and viability was not compromised by the treatment, as no significant difference was observed regarding the percentage of morphologically normal follicles, follicular viability, glucose uptake or in vitro growth of small follicles between ovarian cortex treated with 1 μM GSK1070916 and the control.

Conclusion: Purging of CML/AML metastases in ovarian cortex is possible by targeting the Mitotic Catastrophe Signalling Pathway using GSK1070916 without affecting the ovarian tissue. This provides a therapeutic strategy to prevent reintroduction of leukaemia and enhances safety of autotransplantation in leukaemia patients currently considered at high risk for ovarian involvement.

Keywords: Aurora kinases; Cryopreservation; GSK1070916; Myeloid leukaemia; Ovarian cortex; Purging.

Fig. 1

Expression of AURKB and AURKC by Western blot. Expression of AURKB and AURKC examined by Western blot analysis of CML (JURL-MK1, K562 and MEG-01) and AML (KG-1a, NB4 and NOMO-1) cell lines cultured in suspension and human gonadal tissues (ovarian cortex and testicular tissue). GAPDH was used as a loading control. AURKB was detectable in all CML/AML cell lines with as a single protein band of approximately 35 kDa (NB4 and NOMO-1), or in combination with a slightly smaller protein band (JURL-MK1, K562, MEG-01 an KG-1a). AURKB was undetectable in ovarian cortex tissue or testicular tissue. AURKC was detected in JURL-MK1, K562, MEG-01, KG-1a, ovarian cortex tissue and testicular tissue at different expression levels but absent in NB4 and NOMO-1

Fig. 2

Expression of AURKB and AURKC by immunohistochemistry. Expression of AURKB was examined for all CML and AML cells in suspension and for ovarian cortex tissue by immunohistochemistry with cytokeratin AE1/AE3 as a control for small follicles (panel a). AURKB expression was not detectable in the oocytes or granulosa cells of either primordial, primary or secondary follicles (panel b and c). AURKB expression was detectable in the majority of cells of the CML/AML cell lines in suspension (only shown for the AML cell line KG-1a, panel d). Expression of AURKC was examined for CML and AML cells in suspension and ovarian cortex tissue separately. No AURKC expression was detectable in the oocytes or granulosa cell of primordial or primary follicles (panel e), but AURKC was present in oocytes of secondary follicles (panel f). AURKC expression was clearly detectable in the majority of CML/AML cell lines in suspension (only shown for the AML cell line KG-1a, panel g). Scale bars represent 20 μm

Fig. 3

CML and AML tumour foci in ovarian cortex are sensitive to GSK1070916. HE staining showing that CML tumour foci (JURL-MK, K562 and MEG-01) and AML tumour foci (KG-1a, NB4 and NOMO-1) were abundantly present in control human ovarian cortex treated with solvent-only (panel a–f, white arrowheads indicate tumour foci, the white arrow points at a primordial follicle). After treatment for 24 h with 1 μM GSK1070916 followed by an additional culture for 6 days, tumour foci of JURL-MK1, K562, KG-1a, NB4 and NOMO-1 cells could no longer be detected in ovarian cortex tissue. Large syncytia with sometimes large vacuoles (panel g–r, black arrowheads indicate syncytia) were present in cortex tissue harbouring JURL-MK1, K562 and KG1-a cells after treatment with 1 μm GSK1070916. Apoptotic bodies (grey arrow heads) were present in all cell lines in ovarian cortex tissue after purging except for MEG-01, whereas NB4 and NOMO-1 showed almost exclusively apoptotic bodies with few small syncytia (panel q–r). In contrast to the other five cell lines, MEG-01 showed mostly tumour foci harbouring morphologically normal tumour cells (white arrowheads in panel i and o) next to few small syncytia with up to four nuclei after treatment (black arrowhead in panel o). Scale bars represent 20 μm in panel a–l and 10 μm in panel m-r

Fig. 4

Immunohistochemical analysis of ovarian cortex tissue harbouring CML and AML tumour foci after purging for 24 h with 1 μM GSK1070916. Large-scale formation of syncytia and apoptotic bodies in human ovarian tissue containing CML (JURL-MK1 and K562) and AML (KG-1a, NB4 and NOMO-1) tumour foci after 24-h treatment with 1 μM GSK1070916 followed by 6 days of culture. Nuclei of intact syncytia were positive for proliferation marker Ki-67 (panel g–j) whereas the degraded syncytia and surrounding apoptotic bodies were positive for the apoptotic marker anti-active caspase 3 (AC3) (panel m–r). Note that MEG-01 cells did not show extensive formation of syncytia or apoptotic bodies (panel c) with cells being positive for Ki-67 (panel i) but negative for AC3 (panel o). Scale bars represent 10 μm except in panel b, h and n where it represents 25 μm

Fig. 5

Primary AML tumour foci in ovarian cortex are sensitive to GSK1070916. Experimentally induced tumour foci of primary AML cells in human ovarian cortex tissue treated with solvent only showed normal cell morphology (panel a and c; white arrowheads indicate tumour foci) with a substantial part of the AML cells positive for the proliferation marker Ki67 (panel e and g) and very few cells positive for the apoptotic marker active caspase-3 (AC3, panel i and k). After treatment for 24 h with 1 μM GSK1070916 foci of normal tumour cells could no longer be detected and AML cells gave rise to large syncytia containing multiple nuclei and sometimes large vacuoles (panel b and d, black arrowheads indicate syncytia). Most nuclei of intact syncytia were positive for the proliferation marker Ki-67 (panel f and h) whereas degrading syncytia and apoptotic bodies were positive for the apoptotic marker active caspase-3 (panel j and l). Scale bars represent 30 μm. In supplemental Fig. 1, the results of purging primary cells from ovarian cortex tissue of two additional AML patients are shown

Fig. 6

Morphology and viability of follicles. Ovarian cortex tissue was exposed to a 24-h ex vivo treatment of solvent-only or 1 μM GSK1070916, washed and cultured for an additional 24 h, to allow tissue damage to become apparent. At least 100 follicles were evaluated per patient per condition. a Graphical depiction of the percentage of intact and damaged follicles after treatment. Morphology was examined by HE staining. No statistically significant difference (P = 0.83) was observed between follicles from the control or the tissue treated with the inhibitor. b Graphical depiction of the percentage of viable Neutral red positive follicles and non-viable Neutral red negative follicles. Ovarian cortex tissue was examined by Neutral red staining. No statistically significant difference (P = 0.92) was observed between follicles from the control or the tissue treated with the inhibitor

Fig. 7

In vitro growth of follicles after GSK1090716 exposure. Ovarian cortex tissue of two subjects was exposed to 1 μM GSK1070916 or solvent-only for 24 h, washed and subsequently cultured in IVG culture medium for 8 days. a Graphical depiction of percentage of primordial, primary and secondary follicles in ovarian cortex tissue of two subjects in uncultured tissue (control) and tissue subjected to IVG for 8 days after a 24 h exposure to 1 μM GSK1070916 or solvent only. At least 100 follicles were evaluated per sample. Standard HE staining showed mainly primordial follicles in uncultured ovarian cortex tissue (panel b and f) and a substantial increase in secondary follicles after IVG for 8 days irrespective of previous GSK1070916 treatment (panel c and g). Immunohistochemistry with AC3 indicated apoptosis was absent in oocytes or granulosa cells in GSK1070916 treated tissue (panel d and h) whereas Ki-67 positive granulosa cells showed ongoing folliculogenesis in secondary follicles (panel e and i, black arrowheads point to Ki-67 positive granulosa cells). Scale bars represent 20 μm