Glycogen debranching enzyme (AGL) is a novel regulator of non-small cell lung cancer growth

Abstract

Glycogen debranching enzyme (AGL) and Glycogen phosphorylase (PYG) are responsible for glycogen breakdown. We have earlier shown that AGL is a regulator of bladder tumor growth. Here we investigate the role of AGL in non-small cell lung cancers (NSCLC). Short hairpin RNA (shRNA) driven knockdown of AGL resulted in increased anchorage independent and xenograft growth of NSCLC cells. We further establish that an increase in hyaluronic acid (HA) synthesis driven by Hyaluronic Acid Synthase 2 (HAS2) is critical for anchorage independent growth of NSCLC cells with AGL loss. Using gene knockdown approach against HAS2 and by using 4-methylumbelliferone (4MU), an inhibitor of HA synthesis, we show that HA synthesis is critical for growth of NSCLC cells that have lost AGL. We further show NSCLC cells without AGL expression are dependent on RHAMM for HA signaling and growth. Analysis of NSCLC patient datasets established that patients with low AGL/high HAS2 or low AGL/high RHAMM mRNA expression have poor overall survival compared to patients with high AGL/low HAS2 or high AGL/low RHAMM expression. We show for the first time that loss of AGL promotes anchorage independent growth of NSCLC cells. We further show that HAS2 driven HA synthesis and signaling via RHAMM is critical in regulating growth of these cancer cells with AGL loss. Further patients presenting with low AGL and HAS2 or RHAMM over expressing tumors might present the ideal cohort who would respond to inhibitors of HA synthesis and signaling.

Keywords: AGL; HAS2; RHAMM; hyaluronic acid; non-small cell lung cancer.

Figure 1. Glycogen Debranching Enzyme Loss and Tumor Growth

(A) AGL gene knockdown was validated by Western blot in the NSCLC cell lines. Cells transduced with control shRNA (shCTL) and cells transduced with AGL specific shRNA (shAGL). (B) Anchorage independent growth (n=3) of NSCLC cells with (shCTL) and without AGL (shAGL) expression. 15×10³ cells were plated in 6 well plate for agar growth. Results are shown as mean±SD, ^*p<0.05 by Student's t-test. (C) Xenograft growth of NSCLC cells with (shCTL) and without AGL (shAGL) expression. 7 mice per group were injected in the right and left flank with shCTL and shAGL i) H358 (4×10⁶ cells/site), ii) H2122 (1.0×10⁵ cells/site) or iii) A549 (2×10⁶ cells/site) cells. Tumors were measured as described in Material and Methods. Results are shown as mean±SEM, ^*p<0.05 by Student's t-test.

Figure 2. Glycogen Metabolism and Tumor Growth

(A) AGL expression in H2122 cells transduced with nontarget shRNA (shCTL) and cells transduced with shRNA against AGL specific to 3'UTR region (shAGL') stably overexpressing WT-AGL and enzymatic null AGL. (B) Anchorage independent growth (n=3) of H2122 cells with (shCTL) and without AGL (shAGL') expression stably overexpressing WT-AGL and enzymatic null AGL. 15×10³ cells per cell type were plated in 6 well plate for soft agar growth. (Ci-ii) qRT-PCR demonstrating efficacy of glycogen phosphorylase brain (shPYG-B) and liver (shPYG-L) isoform depletion in H2122 cells stably transduced with shRNA against glycogen phosphorylase brain and liver isoform. (D-E) Anchorage independent (n=3) and dependent (n=6) growth of H2122 cells transduced with nontargeted shRNA and shRNA against glycogen phosphorylase liver (shPYG-B) and brain (shPYG-L) isoform. 15×10³ and 10³ cells were plated in 6 well plates and 96 well plate for monolayer growth and agar growth. Results are shown as mean±SD, ^*p<0.05 by Student's t-test.

Figure 3. HAS2 Expression and HA synthesis with AGL Loss

(Ai-iii) AGL gene knockdown was validated by Western blot in the NSCLC cell lines H358, H2122 and A549 respectively. Cells transduced with control shRNA (shCTL) and cells transduced with AGL specific shRNA (shAGL). (Bi-iii) qRT-PCR demonstrating the expression of HAS2 in NSCLC cells with and without AGL expression (n=3). (Ci-iii) HA secreted into the media by NSCLC cells H358, H2122 and A549 cells with (shCTL) and without AGL (shAGL) expression detected by HA ELISA. Briefly, the cells were plated in 6 welled dish, next day fresh media was added to the cells and HA was measured in the media 24hrs later (n=3). Results are shown as mean±SD, ^*p<0.05 by Student's t-test.

Figure 4. HAS2 loss and growth of NSCLC cells with AGL loss

(A, B) qRT-PCR demonstrating efficacy of HAS2 depletion in A549 and H2122 control (shCTL) and AGL knockdown (shAGL) cells. Cells were plated and 24hrs later transfected with scrambled (siCTL) or directed siRNA against HAS2 (siHAS2). Details of siRNA used are in Materials and Methods. Cells were harvested at 48hrs for mRNA followed by qRT-PCR analysis (n=3). (C, D) HA secreted by A549 and H2122 control (shCTL) and AGL knockdown (shAGL) cells after depletion of HAS2. Cells were plated and 24hrs later transfected with scrambled (siCTL) or directed siRNA against HAS2 (siHAS2). 48 hrs after transfection media was changed on the cells. 24hrs later media was collected for HA ELISA (n=3). (E, F) Proliferation of A549 and H2122 control (shCTL) and AGL knockdown (shAGL) cells after depletion of HAS2. Cells were plated and 24hrs later transfected with scrambled (siCTL) or directed siRNA against HAS2 (siHAS2). 48 hrs after transfection cells were plated in 96 welled dish (10³ cells/well) (n=6) for proliferation over 5 days. Cell proliferation was measured by CyQUANT assay. Results are shown as mean±SD, ^*p<0.05 by Student's t-test.

Figure 5. Effect of 4MU on NSCLC cells with AGL loss

(A, B) A549 and H2122 cells with AGL knockdown (shAGL) were plated in 6 well plates (n=3). 24hrs later media was changed and 4MU was added at different concentrations. HA in the media was measured 24hrs later by ELISA. (C, D) Proliferation ofA549 and H2122 AGL knockdown (shAGL) cells following treatment with 4MU and 4MU+HA. Cells were plated in 96 well plate (10³ cells/well) (n=6). Next day cells were treated with vehicle control (PBS) of 4MU (600μM) or 4MU(600μM)+HA (20μg/ml). Proliferation was measured over a 5 day period by CyQUANT assay. ^*p<0.05 by Student's t-test.

Figure 6. HA receptors and AGL loss in NSCLC

(A) Expression of CD44 and RHAMM in NSCLC cells with (shCTL) or without (shAGL) AGL expression detected by Western blot (n=3). (B) Western blot demonstrating efficacy of RHAMM depletion in A549 and H2122 control (shCTL) and AGL knockdown (shAGL) cells. Cells were plated and 24hrs later transfected with scrambled (siCTL) or directed siRNA against RHAMM (siRHAMM). Details of siRNA used are in Materials and Methods. Cells were harvested at 72hrs for protein followed by Western blot (n=3). (Ci-ii) Proliferation of A549 and H2122 control (shCTL) and AGL knockdown (shAGL) cells after depletion of RHAMM. Cells were plated and 24hrs later transfected with scrambled (siCTL) or directed siRNA against RHAMM (siRHAMM). 48 hrs after transfection cells were plated in 96 well dish (10³ cells/well) (n=6) for proliferation over 5 days. Cell proliferation was measured by CyQUANT assay. Results are shown as mean±SD, ^*p<0.05 by Student's t-test.

Figure 7. Relationship of AGL, HAS2 and RHAMM mRNA to predict NSCLC patient outcome

(Ai-iv) Kaplan–Meier analysis of categorized median mRNA levels of AGL and overall survival in four independent NSCLC patient datasets. (Bi-iv) Kaplan–Meier analysis of categorized median mRNA levels of AGL and HAS2 (High AGL/Low HAS2 vs Low AGL/High HAS2) and overall survival in four independent NSCLC datasets. (Ci-iv) Kaplan–Meier analysis of categorized median mRNA levels of AGL and RHAMM (High AGL/Low RHAMM vs Low AGL/High RHAMM) and overall survival in four independent NSCLC datasets. Details of datasets are in Materials and Methods. Hazard Ratios (HR) and logrank P values are shown.