The anterior gradient homologue 2 (AGR2) co-localises with the glucose-regulated protein 78 (GRP78) in cancer stem cells, and is critical for the survival and drug resistance of recurrent glioblastoma: in situ and in vitro analyses

Abstract

Background: Glioblastomas (GBs) are characterised as one of the most aggressive primary central nervous system tumours (CNSTs). Single-cell sequencing analysis identified the presence of a highly heterogeneous population of cancer stem cells (CSCs). The proteins anterior gradient homologue 2 (AGR2) and glucose-regulated protein 78 (GRP78) are known to play critical roles in regulating unfolded protein response (UPR) machinery. The UPR machinery influences cell survival, migration, invasion and drug resistance. Hence, we investigated the role of AGR2 in drug-resistant recurrent glioblastoma cells.

Methods: Immunofluorescence, biological assessments and whole exome sequencing analyses were completed under in situ and in vitro conditions. Cells were treated with CNSTs clinical/preclinical drugs taxol, cisplatin, irinotecan, MCK8866, etoposide, and temozolomide, then resistant cells were analysed for the expression of AGR2. AGR2 was repressed using single and double siRNA transfections and combined with either temozolomide or irinotecan.

Results: Genomic and biological characterisations of the AGR2-expressed Jed66_GB and Jed41_GB recurrent glioblastoma tissues and cell lines showed features consistent with glioblastoma. Immunofluorescence data indicated that AGR2 co-localised with the UPR marker GRP78 in both the tissue and their corresponding primary cell lines. AGR2 and GRP78 were highly expressed in glioblastoma CSCs. Following treatment with the aforementioned drugs, all drug-surviving cells showed high expression of AGR2. Prolonged siRNA repression of a particular region in AGR2 exon 2 reduced AGR2 protein expression and led to lower cell densities in both cell lines. Co-treatments using AGR2 exon 2B siRNA in conjunction with temozolomide or irinotecan had partially synergistic effects. The slight reduction of AGR2 expression increased nuclear Caspase-3 activation in both cell lines and caused multinucleation in the Jed66_GB cell line.

Conclusions: AGR2 is highly expressed in UPR-active CSCs and drug-resistant GB cells, and its repression leads to apoptosis, via multiple pathways.

Keywords: Anterior gradient homologue 2 (AGR2); Cancer stem cells (CSCs); Drug resistance; Glioblastoma; Glucose-regulated protein 78 (GRP78).

Fig. 1

AGR2 localises with the UPR marker GRP78 in the glioblastoma tissues and corresponding cell lines Jed66_GB and Jed41_GB. Immunofluorescence images for AGR2 (red) co-localising with the GRP78 (green) in A) fresh frozen tissues and B) primary corresponding cell lines. DAPI is shown in blue. All images were taken at 20 × magnification

Fig. 2

UPR markers co-staining with CSC markers. The immunofluorescence images for AGR2 (red) co-localising with either Nestin (green), SSEA4 (green), or VIM (green); and GRP78 (green) co-localising with either SOX2 (red), FZD9 (red), or TUBB3 (red) in A) fresh frozen tissues, B) corresponding primary cell lines and C) Magnified images to show detail intracellular localization of the respective proteins in the cells. DAPI is shown in blue. All images were taken at 20 × magnification. D A barograph showing the percentages of double-positive cells for the aforementioned markers in the primary cell lines. Error bars indicate the data variability between counts for three independent experiments

Fig. 3

AGR2 expression in drug-exposed surviving cells. A A schematic diagram showing the timing of AGR2 detection following drug treatment. B Clonogenic growth evaluations of the primary cell lines treated with taxol, irinotecan, cisplatin, MKC8866 or etoposide. Error bars represent errors between counts for three independent experiments. Passage numbers for both cell lines ranged from 17 to 33. Three asterisks indicate significant difference at p < 0.001. C Immunofluorescence images for treated cells at IC₅₀ values for the tested drugs. The Jed66_GB cell line was treated with taxol (42ρM), irinotecan (81 nM), cisplatin (4 μM), MKC8866 (7.4 μM) or etoposide (14 μM). The Jed41_GB cell line was treated with taxol (2 nM), irinotecan (210 nM), cisplatin (12 μM), MKC8866 (6.7 μM) or etoposide (29 μM). AGR2 is shown in red and DAPI in blue. All images were taken at 20 × magnification

Fig. 4

Repression of AGR2 in the primary cell lines. A Immunofluorescence images showing the expression of AGR2 following treatment using siRNA against AGR2 exons 2 and 5/6. AGR2 is shown in red and DAPI is shown in blue. All images were taken at 20x. B Barographs showing the mean cell densities within 3.9 mm² for the primary cells transfected with the negative control or siRNA AGR2 oligos. Asterisks indicate significant Z-test differences at p < 0.05. Data were collected from three independent experiments. C Inhibition of AGR2 using a double-hit approach; i) A schematic diagram displaying the double-hit approach. ii) Immunofluorescence images showing the repression of AGR2 following two consecutive siRNA treatments using siRNA against AGR2 exon 2B. AGR2 is shown in red and DAPI in blue, and images were taken at 20 × magnification. iii) A barograph showing the survival fraction of cells transfected with Cy3, the negative control, or with two treatments of siRNA against AGR2 exon 2B. Two asterisks indicate significant T-test difference at p < 0.01. The error bars represent errors between counts for three independent experiments. iv) A barograph showing the mean cell densities within 3.9 mm² for cells treated once or twice with negative control or siRNA against AGR2 exon 2B. The asterisk indicates a significant Z-test difference at p < 0.05. Data were collected from three independent experiments

Fig. 5

Inhibition of AGR2 exon 2B in conjunction with the clinically relevant drugs temozolomide (TMZ) and irinotecan (IR). A A schematic diagram showing the treatment protocols and the IC₅₀ values. B Immunofluorescence images for treated cells at day 5/6. AGR2 is shown in red and DAPI in blue, and images were taken at 20x. C Barographs displaying the survival fractions of the primary cell lines following individual and co-treatments using siRNA against AGR2 exon 2B with temozolomide or with irinotecan. The error bars represent errors between counts for three independent experiments. The connective lines indicate a significant T-test difference at p < 0.05

Fig. 6

The effects of early repression of AGR2 in both Jed66_GB and Jed41_GB. A i) Images showing minor repression of AGR2 in both cell lines following the short post-transfection of eight hours for Jed66_GB and 36 h for Jed41_GB. ii) The AGR2 mean intensities for negative control and AGR2 exon 2B siRNA-transfected cells. The asterisk indicates significant T-test difference at p < 0.05. B GRP78 expression and AGR2 expression in both negative control and AGR2 exon 2B siRNA-transfected cells for both cell lines. C Immunofluorescence images of nuclear Caspase-3 in negative control and AGR2 exon 2B siRNA-transfected cells. The arrow points to a Caspase-3-positive multinucleated cell. Caspase-3 is shown in red and DAPI in blue. D i) Images show multinucleation in Jed66_GB ii) The percentage of multinucleated cells in Jed66_GB following negative control and AGR2 exon 2B siRNA transfections. Asterisk indicate T-test significant difference at p < 0.05. E i) Immunofluorescence images displaying the expression of P53 in both cell lines following negative control and AGR2 exon 2B siRNA transfections. P53 is shown in green and DAPI in blue. ii) Magnified images showing P53 localisation. iii) The percentage of cells positive for nuclear P53 was significantly higher in Jed66_GB compared with Jed41_GB. Three asterisks indicate T-test significant difference at p < 0.001. F A suggested model that explains the role of AGR2 in glioblastoma. Both in situ and in vitro, AGR2 and GRP78 are highly expressed in CSCs and drug-resistant cells. Upon repression of AGR2, GRP78 was also reduced. Early cell fate seems to be dependent on the used model, and can lead either to multinucleation followed by cell death or to delayed growth followed by attenuated cell death