FANCD2 re-expression is associated with glioma grade and chemical inhibition of the Fanconi Anaemia pathway sensitises gliomas to chemotherapeutic agents

Abstract

Brain tumours kill more children and adults under 40 than any other cancer. Around half of primary brain tumours are glioblastoma multiforme (GBMs) where treatment remains a significant challenge, where survival rates have improved little over the last 40 years, thus highlighting an unmet need for the identification/development of novel therapeutic targets and agents to improve GBM treatment. Using archived and fresh glioma tissue, we show that in contrast to normal brain or benign schwannomas GBMs exhibit re-expression of FANCD2, a key protein of the Fanconi Anaemia (FA) DNA repair pathway, and possess an active FA pathway. Importantly, FANCD2 expression levels are strongly associated with tumour grade, revealing a potential exploitable therapeutic window to allow inhibition of the FA pathway in tumour cells, whilst sparing normal brain tissue. Using several small molecule inhibitors of the FA pathway in combination with isogenic FA-proficient/deficient glioma cell lines as well as primary GBM cultures, we demonstrate that inhibition of the FA pathway sensitises gliomas to the chemotherapeutic agents Temozolomide and Carmustine. Our findings therefore provide a strong rationale for the development of novel and potent inhibitors of the FA pathway to improve the treatment of GBMs, which may ultimately impact on patient outcome.

Figure 1. The FA pathway is highly expressed and active in high-grade gliomas, with FANCD2 expression associated with tumour grade

A: Representative images of FANCD2 staining in archived FFPE sections of schwannoma (negative control) and gliomas of various tumour grades as indicated. FANCD2 is a nuclear protein, which can be seen in the grade III and grade IV tumours shown (brown DAB staining). B: Quantification of FANCD2 expression in 131 schwannoma and gliomas of various tumour grades. A chi-squared test was used to test for association with tumour grade. C: Representative images of FANCD2 staining in normal brain and gliomas of various tumour grade taken from commercial tissue microarray. Insert at bottom left of each image shows a magnified field from the section to show nuclear FANCD2 staining. The table underneath shows quantification of FANCD2 staining in the 36 samples present on the tissue microarray, indicating number of each sample exhibiting FANCD2 staining and the percentage of each tumour grade positive for FANCD2 expression. The percentage of FANCD2 positive tumours that also stained positive for Ki67 (supplementary Figure S1B) is also shown in the table. A chi-squared test was used to test for association with tumour grade. D: Western blots showing FANCD2 expression and activation (mono-ubiquitinated FANCD2; Ub-FANCD2) following Temozolomide (TMZ) treatment in two independently derived primary glioma cultures. To determine FA pathway activation, the normalised ratio of Ub-FANCD2 (L) to FANCD2 (S) was calculated for TMZ vs DMSO treated cells as previously described [30], and is shown under the blots as an L:S ratio. Also shown is MGMT expression and tubulin, which serves as a loading control.

Figure 2. The FAPi curcumin, EF-24 and DDN inhibit FA pathway activation in immortalised and primary glioma cell cultures

A: Representative western blots showing FANCD2 and actin (loading control) expression in U87 (left panel) and U138 cells (right panel). Treatment with Temozolomide (TMZ) activates the FA pathway as evidenced by the appearance of a slower migrating form of FANCD2 with represents the mono-ubiquitinated form (Ub-FANCD2). Pre-treatment with the indicated that FAPi severely inhibits FA pathway activation as evidenced by significantly reduced Ub-FANCD2 after TMZ treatment. DMSO treatments were used as negative controls. To determine FA pathway activation, the normalised ratio of Ub-FANCD2 (L) to FANCD2 (S) (normalised to DMSO+TMZ treated cells) was calculated for each treatment as previously described [30] and shown under the blots for U87 cells. Note that all values are zero for U138 cells). As described previously, U138 cells express low levels of FANCD2 and exhibit a defective FA pathway (no Ub-FANCD2 after TMZ treatment). The FANCD2 antibody recognises a non-specific band in U138 cell line as indicated (n.s. band), which is most likely due to reduced epitope expression in this cell line. B: Immunofluorescence detection and quantification of cells exhibiting 10 or more nuclear FANCD2 foci (a marker of an active FA pathway) in U87 and U138 cells treated as described above. Upper panel shows representative images for treated cells and the graph below show quantified data from at least three independent experiments. Error bars represent standard deviation of the means. C: Western blot showing FANCD2 expression and activation in a primary glioma culture. As with immortalised glioma cell lines, pre-treatment of the primary culture with the indicated FAPi compromises FA pathway activation as calculated by determining the normalised ratio of Ub-FANCD2 (L) to FANCD2 (S) as described in panel A. D: Quantification of FANCD2 nuclear foci formation in 2 separate primary glioma cultures. Cells were either treated with DMSO (negative control) or TMZ, with or without pre-treatment with the indicated FAPi as described in the material and methods. Representative images for this data are shown in supplementary Figure S1D. Due to the difficulty in the long-term propagation of such cultures and their weak adherence on glass coverslips, error bars represent standard deviation of the means from 2 replica plates.

Figure 3. FAPi sensitise immortalised and primary glioma cultures to chemotherapeutic agents irrespective of MGMT status

A: MTT Temozolomide cytotoxicity assays for U87 and U138 cells treated with the FAPi curcumin, EF-24 or DDN. Solid lines represent cells pre-treated with DMSO and dashed lines represent cells pre-treated with the indicated FAPi. Data shown is the mean from at least 3 independent experiments with error bars representing the standard deviation of the means. B: MTT Temozolomide cytotoxicity assays for two independently derived primary glioma cultures from at least three independent experiments. An additional example is shown in supplementary Figure S2B. Solid, dashed lines and error bars are as described in panel A. C: MTT Temozolomide cytotoxicity assays for T98G cells pre-treated with either DMSO or the FAPi curcumin, EF-24 or DDN. Solid, dashed lines and error bars are as described in panel A.

Figure 4. Generation of isogenic FANCD2 proficient and deficient U87 cell lines and their use to determine FAPi specificity

A: Western blots showing FANCD2 expression in a number of stable shRNA expressing clones. Control cells represent those stably transfected with a non-targeting shRNA plasmid, while FANCD2 are clones derived from cells transfected with FANCD2-targeting shRNA plasmids. Actin is shown as a loading control and red asterisks highlight clones selected for use in further studies based on their residual FANCD2 levels. B: Western blot showing FANCD2 expression and activation (Ub-FANCD2) following Temozolomide treatment in the selected stable U87 clones outlined in panel A. C: MTT Temozolomide cytotoxicity assays for the selected stable U87 clones outlined in panel A pre-treated with either DMSO or the FAPi curcumin, EF-24 or DDN. Data shown was derived from at least three independent experiments and solid, dashed lines and error bars are as described in Figure 3A.