Deuterium Oxide (D2O) Induces Early Stress Response Gene Expression and Impairs Growth and Metastasis of Experimental Malignant Melanoma

Abstract

There are two stable isotopes of hydrogen, protium (¹H) and deuterium (²H; D). Cellular stress response dysregulation in cancer represents both a major pathological driving force and a promising therapeutic target, but the molecular consequences and potential therapeutic impact of deuterium (²H)-stress on cancer cells remain largely unexplored. We have examined the anti-proliferative and apoptogenic effects of deuterium oxide (D₂O; 'heavy water') together with stress response gene expression profiling in panels of malignant melanoma (A375^V600E, A375^NRAS, G361, LOX-IMVI), and pancreatic ductal adenocarcinoma (PANC-1, Capan-2, or MIA PaCa-2) cells with inclusion of human diploid Hs27 skin fibroblasts. Moreover, we have examined the efficacy of D₂O-based pharmacological intervention in murine models of human melanoma tumor growth and metastasis. D₂O-induction of apoptosis was substantiated by AV-PI flow cytometry, immunodetection of PARP-1, and pro-caspase 3 cleavage, and rescue by pan-caspase inhibition. Differential array analysis revealed early modulation of stress response gene expression in both A375 melanoma and PANC-1 adenocarcinoma cells elicited by D₂O (90%; ≤6 h) (upregulated: CDKN1A, DDIT3, EGR1, GADD45A, HMOX1, NFKBIA, or SOD2 (up to 9-fold; p < 0.01)) confirmed by independent RT-qPCR analysis. Immunoblot analysis revealed rapid onset of D₂O-induced stress response phospho-protein activation (p-ERK, p-JNK, p-eIF2α, or p-H2AX) or attenuation (p-AKT). Feasibility of D₂O-based chemotherapeutic intervention (drinking water (30% w/w)) was demonstrated in a severe combined immunodeficiency (SCID) mouse melanoma metastasis model using luciferase-expressing A375-Luc2 cells. Lung tumor burden (visualized by bioluminescence imaging) was attenuated by D₂O, and inhibition of invasiveness was also confirmed in an in vitro Matrigel transwell invasion assay. D₂O supplementation also suppressed tumor growth in a murine xenograft model of human melanoma, and median survival was significantly increased without causing adverse effects. These data demonstrate for the first time that systemic D₂O administration impairs growth and metastasis of malignant melanoma through the pharmacological induction of deuterium (²H)-stress.

Keywords: A375 melanoma xenograft model; A375-luciferase reporter cells; SCID mouse metastasis model; cellular stress response; deuterium oxide; heavy water; malignant melanoma; transwell invasion.

Figure 1

D₂O-induced apoptosis in a panel of human malignant melanoma cells (A375^V600E, A375^NRAS, LOX-IMVI, and G361). (A) Impairment of cellular viability in response to culture in D₂O (90%, 24 h) was monitored using flow cytometric analysis (annexin V-PI staining). Numbers in quadrants indicate percentage of viable cells (AV-negative, PI-negative) from a total of gated cells (mean ± SD, n = 3). Bar graph (right panel) indicates dose response of impaired cell viability (D₂O (≤90%, 24 h)) (n = 3). Human Hs27 dermal fibroblasts exposed to D₂O served as non-transformed, non-malignant controls. (B) Impairment of cellular viability in response to long term exposure to D₂O (27%, ≤6 days). Bar graph depicts dose response and time course ((≤6 days); (n = 3)). (C) Impairment of cellular proliferation (A375) in response to culture in D₂O (≤27%, 3 days). (D) M-phase depletion as a function of culture in D₂O (27%, 3 days) as assessed by phospho-histone H3 flow cytometry: individual histograms representative of three repeats (left panel); right panel: bar graph depiction of numerical analysis (n = 3; p *** < 0.001). (E) D₂O-induced (90%, 24 h) cell death (A375^V600E versus A375^NRAS) in the absence or presence of zVAD-fmk (40 µM). (F) D₂O-induced (90%, 24 h) induction of pro-caspase 3 cleavage as examined in A375^V600E cells by flow cytometry (left panel: representative histograms; right panel: bar graph depiction of numerical analysis (n = 3; p *** < 0.001). (G) Time course of PARP-1 cleavage in response to D₂O-exposure (90%, ≤24 h) in A375 cells (bottom panel: immunoblot; top panel: bar graph depiction of numerical analysis (n = 3).

Figure 2

D₂O-induced apoptosis in a panel of human pancreatic ductal adenocarcinoma cells (PANC-1, MIA PaCa-2, and Capan-2). (A) Loss of cellular viability in response to D₂O (90%, 24 h) as analyzed in Figure 1A. (B) D₂O-induced (90%, 24 h) induction of pro-caspase 3 cleavage examined in PANC-1 cells by flow cytometry (left panel: representative histograms; right panel: bar graph depiction of numerical analysis (n = 3; p *** < 0.001). (C) D₂O-induced (90%, 24 h) cell death examined in the absence or presence of zVAD-fmk (40 µM). (D) PARP-1 cleavage in response to D₂O-exposure (90%, 24 h) observable in pancreatic ductal adenocarcinoma (PDAC) cells by immunoblot analysis; bar graph depiction summarizes densitometric analysis (n = 3; p* < 0.05). (E) Impairment of cellular proliferation (PANC-1) by culture in D₂O (≤27%, 3 days). (F) M-phase depletion as a function of culture in D₂O (27%, 3 days) as assessed and analyzed in Figure 1D (n = 3; p ** < 0.01).

Figure 3

D₂O-induced early stress response gene expression comparing melanoma (A375) and pancreatic ductal adenocarcinoma (PANC-1) cells. (A) Time-course analysis of cell viability impairment assessed by flow cytometry (performed as in Figure 1A) in A375 and PANC-1 cells cultured in 90% D₂O (≤24 h; n = 3; p * < 0.05; p ** < 0.01). (B) Volcano plot depicting differential gene expression (untreated versus D₂O-exposed (90%, 6 h)) as identified by the Human Stress and Toxicity PathwayFinder^TM PCR Array technology (cut off criteria: expression differential >2; p-value ≤ 0.05; n = 3; A375 (black diamond); PANC-1 (empty diamond). (C) Comparative gene expression array analysis in Venn diagram depiction; in the overlapping region, single arrow indicates congruent up- or downregulation, and double arrows indicate opposing expression changes between cell lines. (D) Comparative gene expression array analysis with total number of genes per group as summarized numerically (A375 vs. PANC-1; D₂O exposed as in panel B).

Figure 4

D₂O-induced stress response gene expression and rapid onset of modulated phospho-protein signaling in A375 melanoma cells. (A) RT-qPCR assessment of gene expression; left row, dose response (≤90 % D₂O); right row, time course (≤6 h). (B) Stress response protein phosphorylation in response to acute D₂O exposure as profiled by immunoblot analysis: time course (90% D₂O; ≤24 h). Bar graphs summarize quantitative analysis by densitometry (mean ± SD). (C) For γH2AX detection, flow cytometry was performed; bar graphs summarize quantitative analysis (mean ± SD; n = 3; p *** < 0.001).

Figure 5

Systemic administration of D₂O attenuates A375 melanoma cell invasiveness in vitro, while impairing metastasis and tumor growth in SCID mouse models of human malignant melanoma. (A,B) A375-Luc2 melanoma cells were tail vein injected (n = 8 per group) followed by bioluminescent image analysis of lung metastasis (14 and 28 days later). Starting at time of cell injection until day 14, mice received H₂O-based or D₂O-supplemented (30% v/v in H₂O) drinking water, followed by 14 days H₂O in both groups. A, injection scheme; B, bioluminescent imaging (day 28) with bar graph depicting numeric image analysis of bioluminescent signal (p *** < 0.001). (For bioluminescent imaging on day 14, see Figure S4). (C) Invasion through Matrigel-coated Boyden chamber (H₂O-based versus D₂O-supplemented (27%) medium). Bar graphs with representative images (10× magnification) after crystal violet staining of inserts (n = 3; p * < 0.05). (D–G) A375 melanoma cells were injected subcutaneously (n = 10 per group); after pair-matching, tumor growth was monitored over a 24-day period; starting at time of pair matching (day 0) until end of experiment (day 24), mice received H₂O-based or D₂O-supplemented (30% v/v in H₂O) drinking water. (D) Experimental scheme. (E) Kaplan–Meier analysis of mouse survival as a function of treatment groups; numbers indicate survivors per group. (F) Tumor burden in treatment groups as a function of time; graph (right panel), individual tumor size at termination. (G) At the end of the experiment, tumors were processed for IHC (left panels; 20× magnification); bar graph: tissue H-scores per antigen (right panel; n = 3; p * < 0.05).