Genotoxicity of cytokines at chemotherapy-induced 'storm' concentrations in a model of the human bone marrow

Abstract

Donor cell leukaemia (DCL) is a complication of haematopoietic stem cell transplantation where donated cells become malignant within the patient's bone marrow. As DCL predominates as acute myeloid leukaemia, we hypothesized that the cytokine storm following chemotherapy played a role in promoting and supporting leukaemogenesis. Cytokines have also been implicated in genotoxicity; thus, we explored a cell line model of the human bone marrow (BM) to secrete myeloid cytokines following drug treatment and their potential to induce micronuclei. HS-5 human stromal cells were exposed to mitoxantrone (MTX) and chlorambucil (CHL) and, for the first time, were profiled for 80 cytokines using an array. Fifty-four cytokines were detected in untreated cells, of which 24 were upregulated and 10 were downregulated by both drugs. FGF-7 was the lowest cytokine to be detected in both untreated and treated cells. Eleven cytokines not detected at baseline were detected following drug exposure. TNFα, IL6, GM-CSF, G-CSF, and TGFβ1 were selected for micronuclei induction. TK6 cells were exposed to these cytokines in isolation and in paired combinations. Only TNFα and TGFβ1 induced micronuclei at healthy concentrations, but all five cytokines induced micronuclei at storm levels, which was further increased when combined in pairs. Of particular concern was that some combinations induced micronuclei at levels above the mitomycin C positive control; however, most combinations were less than the sum of micronuclei induced following exposure to each cytokine in isolation. These data infer a possible role for cytokines through chemotherapy-induced cytokine storm, in the instigation and support of leukaemogenesis in the BM, and implicate the need to evaluate individuals for variability in cytokine secretion as a potential risk factor for complications such as DCL.

Keywords: chemotherapy; concentrations; cytokines; genotoxicity; induced; storm.

Figure 1.

Profile of cytokine secretion by untreated HS-5 cells. Out of 80 cytokines tested, 54 cytokines were secreted in untreated HS-5. Normalization was performed using positive control signals on each array. Cytokines were measured semi-quantitatively utilizing absorbance values of each cytokine spot on the membrane and corrected for background cytokines from the culture medium. Data show mean ± SD (n = 3).

Figure 2.

Cytokine secretion at 72 and 48 h from HS-5 cells following 1 h exposure to CHL and MTX, respectively. Cytokine secretion is arranged in the order of magnitude for CHL. (A) Cytokines which were upregulated by both drugs in comparison to untreated HS-5 cells. (B) Cytokines which were downregulated by both drugs relative to untreated HS-5 cells. (C) Cytokines which were upregulated by one drug, but down regulated by the other; ENA-78 and GRO-Alpha were upregulated by MTX, whereas the remaining cytokines were upregulated by CHL. Data show mean ± SD (n = 3). CHL, chlorambucil; MTX, mitoxantrone.

Figure 3.

Fold change of cytokine expression in HS-5 cells exposed to CHL and MTX for 1 h relative to untreated cells. Cytokines are arranged in order of magnitude of fold change for CHL. (A) Cytokines which have positive fold change by exposure to both CHL and MTX. (B) Cytokines which have negative fold change common to both drugs. (C) Cytokines which were positive for one drug, but negative fold change for the other. Data show mean ± SD (n = 3). CHL, chlorambucil; MTX, mitoxantrone.

Figure 4.

Detection of cytokines secreted only in response to CHL and MTX drug exposure. Cytokines were identified for being absent in untreated HS-5 (negative absorbance values on untreated membranes) but then detected following drug exposure. Out of 80 cytokines, 11 cytokines were not secreted from untreated HS-5 cells however, all were detected following drug exposure. FGF-7 as the lowest detected cytokine both in untreated and drug treated HS-5 cells is presented for comparison of relative secretion levels. Data are presented as mean ± SD (n = 3).

Figure 5.

The induction of micronuclei in TK6 cells due to direct treatment of cytokines at healthy and storm plasma concentrations. TK6 cells were cultured in the presence of each cytokine for 24 h. After a 24-h recovery period, cells were harvested and evaluated for relative population doubling (RPD) and chromosomal damage by scoring the number of MN present. Mitomycin C (MMC; 10,000 pg/ml) was the positive control, and PBS was the negative control. Controls were used for all experimental repeats. Data show the mean ± SD (n = 3). *P < 0.05, **P < 0.01, ***P < 0.001 (for MMC vs. PBS).

Figure 6.

The induction of micronuclei in TK6 cells following paired cytokine combination treatments. TK6 cells were cultured with two cytokines at different concentrations for 24 h. After a 24-h recovery period, cells were harvested and evaluated for the RPD and number of MN present. MMC (10,000 pg/ml) was the positive control and PBS was the negative control for all experimental repeats. Data are presented as mean ± SD (n = 3) for all the concentrations analysed and significant differences shown as *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001, and ****P ≤ 0.0001.