Energy Blocker Lonidamine Reverses Nimustine Resistance in Human Glioblastoma Cells through Energy Blockade, Redox Homeostasis Disruption, and O⁶-Methylguanine-DNA Methyltransferase Downregulation: In Vitro and In Vivo Validation

Abstract

Tumor resistance seriously hinders the clinical application of chloroethylnitrosoureas (CENUs), such as O⁶-methylguanine-DNA methylguanine (MGMT), which can repair O⁶-alkyl lesions, thereby inhibiting the formation of cytotoxic DNA interstrand cross-links (ICLs). Metabolic differences between tumor and normal cells provide a biochemical basis for novel therapeutic strategies aimed at selectively inhibiting tumor energy metabolism. In this study, the energy blocker lonidamine (LND) was selected as a chemo-sensitizer of nimustine (ACNU) to explore its potential effects and underlying mechanisms in human glioblastoma in vitro and in vivo. A series of cell-level studies showed that LND significantly increased the cytotoxic effects of ACNU on glioblastoma cells. Furthermore, LND plus ACNU enhanced the energy deficiency by inhibiting glycolysis and mitochondrial function. Notably, LND almost completely downregulated MGMT expression by inducing intracellular acidification. The number of lethal DNA ICLs produced by ACNU increased after the LND pretreatment. The combination of LND and ACNU aggravated cellular oxidative stress. In resistant SF763 mouse tumor xenografts, LND plus ACNU significantly inhibited tumor growth with fewer side effects than ACNU alone. Finally, we proposed a new “HMAGOMR” chemo-sensitizing mechanism through which LND may act as a potential chemo-sensitizer to reverse ACNU resistance in glioblastoma: moderate inhibition of hexokinase (HK) activity (H); mitochondrial dysfunction (M); suppressing adenosine triphosphate (ATP)-dependent drug efflux (A); changing redox homeostasis to inhibit GSH-mediated drug inactivation (G) and increasing intracellular oxidative stress (O); downregulating MGMT expression through intracellular acidification (M); and partial inhibition of energy-dependent DNA repair (R).

The “Warburg effect” is the observation that tumor cells preferentially use glycolysis for glucose metabolism even in oxygen-rich conditions rather than the more efficient mitochondrial oxidative phosphorylation (OXPHOS).^1,2 Aberrant tumor proliferation and drug resistance are associated with this unique energy metabolism.³ Malignant glioblastoma cells exhibit high glycolytic rates that are associated with survival and drug resistance.^4,5 From a drug discovery perspective, this unique energy metabolism provides a promising therapeutic target for the specific killing of tumor cells.⁶

Chloroethylnitrosoureas (CENUs), such as nimustine (ACNU) and carmustine (BCNU), have been approved for the treatment of various tumors, especially glioblastoma, because of their ability to penetrate the blood–brain barrier.^7,8 The antitumor activity of CENUs is closely associated with the formation of DNA interstrand cross-links (ICLs) between guanine and cytosine, which are deemed biomarkers of their therapeutic efficacy.^9,10 However, tumor resistance leads to frequent failure of CENUs in clinical applications to a significant extent.¹¹ First, a unique DNA repair enzyme, O⁶-methylguanine-DNA methyltransferase (MGMT), transfers alkyl lesions from the guanine O⁶ position to the Cys145 residue at the active site of the protein.^12,13 In this case, the cross-linked precursors (O⁶-chloroethylguanine or N1,O⁶-ethanoguanine) produced by CENUs can be repaired, thereby avoiding the formation of lethal DNA ICLs, which result in resistance.¹² Second, the glutathione (GSH)/glutathione-S-transferase (GST)-mediated detoxification effect can weaken the antitumor activity through the quenching of reactive electrophiles produced by CENUs or direct reaction with parent CENUs.¹⁴ Third, multidrug resistance (MDR)-related proteins such as p-glycoprotein (ABCB1) and other ABC transporters can export intracellular drugs out of the cell (referred to as drug efflux), leading to drug resistance.¹⁵ In a previous study, reduced uptake and enhanced efflux of ACNU were considered the main causes of resistance in rat C6 and 9L glioma cells.¹⁶

Although several MGMT inhibitors have been developed to sensitize tumor cells to guanine O⁶-alkylating agents, such as ACNU, BCNU, and temozolomide (TMZ); in the past several decades, only two inhibitors, O⁶-benzylguanine (O⁶-BG) and O⁶-(4-bromothenyl)guanine (O⁶-4-BTG), which have been approved for clinical trials, did not exhibit desirable therapeutic effects when used in combination treatment.¹² Therefore, to reduce the adverse effects of CENUs on normal cells and overcome their resistance to tumors, it is necessary to identify new chemosensitizers and improve their clinical efficacy. Based on the Warburg effect, the use of glycolytic inhibitors as chemosensitizers may be a promising strategy for improving the antitumor activity of CENUs. Antiglycolytic agents such as 3-bromopyruvate (3-BrPA), 2-deoxyglucose (2-DG), and lonidamine (LND) have been widely explored for use in antitumor research in recent years.^2,17,18 Among these, LND has attracted wide attention because it exhibits great potential as an effective potentiator of other chemotherapeutic agents in vitro and in vivo.¹⁸ As a single agent, LND shows limited antitumor activity; however, it exhibits the unique potential to modulate the effects of conventional chemotherapeutic drugs: (a) tumor selectivity, (b) low toxicity to normal cells, and (c) multiple biochemical mechanisms, including inhibition of glycolysis, mitochondrial respiration, lactic acid efflux, and pyruvate uptake.^19,20 The sensitizing effects of LND in combination with other N-mustards,²¹ TMZ,²² and doxorubicin,^23,24 as well as in physical therapies,^25,26 have been shown. In these studies, LND has been proposed to inhibit GST and MGMT by inducing intracellular acidification, leading to decreased drug efflux and DNA repair.¹⁹⁻²¹ In addition, LND induces oxidative stress by acting on tumor mitochondria to enhance the cytotoxic effects triggered by chemotherapeutic agents.^27,28 Tumor de-energization is a typical feature of LND due to the inhibition of glycolysis and mitochondrial respiration.^19,20

Based on the above findings, LND, an energy metabolism blocker, is expected to overcome the resistance of human glioblastoma cells to CENUs by downregulating the intracellular energy status, disrupting redox homeostasis, and inducing intracellular acidification in tumor cells. In the present study, we investigated the effect of LND as a chemo-sensitizer on the sensitivity of human glioblastoma cells to ACNU in vitro and in vivo and further unraveled the underlying chemo-sensitizing mechanisms. This study provides a new perspective for the use of CENUs in the clinical treatment of glioblastoma.

Results

LND Increased the Cytotoxicity of ACNU to Two Human Glioblastoma Cell Lines

As shown in Figure 1A and B, LND alone produced a dose-dependent decrease in cell survival rate; the IC₅₀ values were 220 and 725 μM for SF126 and SF763 cells, respectively. We also calculated the IC₂₅ values for SF126 and SF763 cells at 100 and 480 μM, respectively. Likewise, ACNU alone also exhibited a dose-dependent manner in inhibiting cell survival; the IC₂₅ and IC₅₀ values were 140 and 480 μM for the SF126 cell line, and 605 and 1445 μM for SF763 cell line, respectively (Figure 1C and D).

Figure 1.

In vitro antitumor activity of SF126 and SF763 cells treated with LND, ACNU, and their combination, respectively. Survival rates of (A) SF126 and (B) SF763 cells (LND for 24 h: 0–1000 μM), (C) SF126 and (D) SF763 cells (ACNU for 24 h: 0–2000 μM), and (E) SF126 and (F) SF763 cells were pretreated with LND (0, IC₂₅, and IC₅₀) for 24 h, followed by 24 h exposure to ACNU (0–2000 μM). Colony formation abilities of SF126 and SF763 cells after exposure to combined LND and ACNU treatment: (G) and (H) indicate the colony images and efficiency of SF126 cells pretreated with LND for 24 h and then treated with 0 μM, IC₂₅, and IC₅₀ dosage of ACNU for 24 h, respectively; (I) and (J) indicate the cell colony images and efficiency of SF763 cells pretreated with LND for 24 h and then treated with 0 μM, IC₂₅, and IC₅₀ dosage of ACNU for 24 h, respectively. (K) Flow cytometric scatter plot and (L) quantitative analysis results of SF126 cells after drug treatment (44.5%, pretreatment with 220 μM LND for 24 h, followed by treatment with 480 μM ACNU for 24 h). (M) Flow cytometric scatter plot and (N) quantitative analysis result of SF763 cells after drug treatment (31.9%, pretreatment with 725 μM LND for 24 h, followed by treatment with 1445 μM ACNU for 24 h). The results are presented as mean ± SD (n = 5 per group). *p < 0.05 indicates statistically significant differences between the groups.

Compared with ACNU alone, LND plus ACNU showed significantly higher cytotoxicity (Figure 1E,F). By calculating the combination index (Q) of LND and ACNU (Table S1), we found that the combination had a synergistic effect on SF126 and SF763 cells in most cases. The Q values of SF126 and SF763 cells ranged 0.85–1.18 and 0.98–1.44, respectively. In general, SF763 showed a better drug combination effect and its Q values were superior to those of SF126 cells under the same treatment conditions.

Combination Effects of LND and ACNU on the Colony Formation, Cell Apoptosis, Migration, and Invasion of Human Glioblastoma Cells

In SF763 cells, the IC₂₅ and IC₅₀ values were 70 and 218 μM for LND and 56 and 114 μM for ACNU in the colony formation assay, respectively. The IC₂₅ and IC₅₀ values of LND against SF126 cells were 114 and 334 μM, and the IC₂₅ and IC₅₀ values of ACNU were 10 and 24 μM, respectively. The effect of the drug combination on the proliferation ability was verified using an orthogonal test. As shown in Figure 1G–J, the combination of LND and ACNU dramatically weakened the proliferative abilities of both glioblastoma cell lines compared with ACNU or LND alone. For SF126 cells, compared to ACNU alone, the combination with LND resulted in a significant decrease in colony formation rates (Figure 1G,H). Similarly, for SF763 cells, after LND pretreatment, ACNU significantly decreased the colony formation rates compared to ACNU alone (Figure 1I,J). These results showed that LND pretreatment enhanced the growth inhibitory effect of ACNU on glioblastoma cells.

The apoptosis-inducing effects were determined using an Annexin V-FITC/PI apoptosis assay kit. As shown in Figure 1K–N, compared to LND or ACNU alone, LND combined with ACNU significantly induced apoptosis in both human glioblastoma cell lines.

We investigated the migration and repair abilities of glioblastoma cells using a cell scratch assay. As shown in Figure 2A–D, scratches on SF126 and SF763 cells healed after 48 h without drug treatment and the effect of scratch healing was time-dependent. As shown in Figure 2A,C, the wound healing rate of SF126 cells exposed to 220 μM LND (24 h pretreatment) in combination with 140 or 480 μM ACNU for 48 h showed the most significant difference compared to that without LND pretreatment. Similarly, in SF763 cells, 725 μM LND (24 h of pretreatment) combined with 605 μM ACNU for 48 h had the most significant inhibitory effect on cell migration, which was approximately 75% lower than that without LND pretreatment (Figure 2B,D). The scratch healing assay showed that LND plus ACNU significantly inhibited the migration and repair of two glioblastoma cell lines.

Figure 2.

Wound healing rates of SF126 and SF763 cells treated with LND (24 h pretreatment) plus ACNU (0–48 h): (A) wound healing at different time points and (C) the migration rates of SF126 cells after exposure to LND, ACNU, or LND plus ACNU at different time points; (B) wound healing at different time points and (D) the migration rates of SF763 cells after exposure to LND, ACNU, or LND plus ACNU at different time points. Invasion abilities of glioblastoma SF126 and SF763 cells after the treatment with LND and ACNU combination: (E) Invasion status and (G) invasion rates of SF126 cells after 24 h LND pretreatment and 24 h ACNU treatment; (F) Invasion status and (H) invasion rates of SF763 cells after 24 h LND pretreatment and 24 h ACNU treatment. (I) MMP2 protein expression in SF126 cells after exposure to indicated drugs and (J) quantitative analysis; (K) MMP2 protein expression in SF763 cells after exposure to indicated drugs and (L) quantitative analysis. For wound healing and invasion assays, the results are shown as mean ± SD (n = 5 per group), *p < 0.05 vs ACNU alone treated groups. For Western blotting analysis, the results are shown as mean ± SD of five independent experiments. *p < 0.05 indicates statistically significant differences between the groups (paired one-tailed t test).

The invasive or metastatic ability was further explored by using a transwell assay. By comparing the number of SF126 and SF763 cells crossing the membrane at the base of the chamber, we found that LND significantly enhanced the inhibitory effect of ACNU on glioblastoma cell invasion compared to that of ACNU alone (Figure 2E–H). For SF126 cells, the invasion rate in groups treated with 220 μM LND combined with 140 or 480 μM ACNU decreased by 76.67 and 62.67%, respectively, compared to ACNU alone (Figure 2E,G). For SF763 cells, the invasion rate of groups treated with 725 μM LND plus 1445 μM ACNU decreased to 4.33% of that of the control group (Figure 2F,H). These results suggest that LND combined with ACNU has a strong inhibitory effect on the invasion of the two glioblastoma cell lines. Therefore, their combination may inhibit the metastasis of malignant gliomas in clinical settings.

Western blotting was used to detect the expression of migration- and invasion-related protein matrix metalloproteinase-2 (MMP2). As shown in Figure 2I,J, SF126 cells were more sensitive to ACNU, and MMP2 protein expression decreased by nearly 71.8% compared with the control group after 480 μM ACNU treatment for 24 h. However, after pretreatment with LND, the level of MMP2 protein expression was further decreased by approximately 85.0%. SF763 cells were more resistant to ACNU alone than SF126 cells; MMP2 protein decreased by only 39.3% after treatment with ACNU alone for 24 h compared to the control group (Figure 2K,L). However, when ACNU was combined with LND, the level of MMP2 protein expression decreased by approximately 70%. Compared with SF126 cells, the combined effect in ACNU-resistant SF763 cells was more significant (Figure 2K,L).

LND Plus ACNU Reduced Intracellular ATP Levels through Glycolysis Inhibition and Mitochondrial Dysfunction

We found that LND alone inhibited hexokinase (HK) activity in SF126 and SF763 cells in a dose-dependent manner, although the decrease was not significant (Figure 3A,B). The decreasing degree of HK activity in SF126 and SF763 cells after exposure to 220 and 725 μM LND was 17 and 20%, respectively, in comparison with the control group. HK activity decreased only slightly after LND combined with the ACNU treatment. However, this decrease in HK activity was not sufficient to significantly downregulate intracellular energy levels.

Figure 3.

Effects of LND plus ACNU on HK activity, MMP, and intracellular energy status. HK activity in (A) SF126 and (B) SF763 cells after exposure to LND (24 h pretreatment) and ACNU (24 h treatment). Changes of MMP in human glioblastoma (C) SF126 and (D) SF763 cells after the combined treatment of LND (24 h pretreatment) and ACNU (24 h treatment). Cells were stained with JC-1 and Hoechst 33342 for 20 min and then imaged using a fluorescence microscope (n = 3 per group). Intracellular ATP levels in (E) SF126 and (F) SF763 cells after pretreatment with LND for 24 h, followed by exposure to different concentrations of ACNU (0–2000 μM) for 24 h. The results are shown as the mean ± SD (n = 5 per group). * p < 0.05 vs LND alone.

Notably, the mitochondrion is also a potential target of LND, which can affect mitochondrial metabolism by inhibiting complex II in the mitochondrial respiratory chain and inducing permeability transition pore opening in the mitochondrial membrane, leading to the dissipation of transmembrane potential.^18,20,27,28 As shown in Figure 3C,D, LND alone induced a dose-dependent decrease in the mitochondria membrane potential (MMP) in these two glioblastoma cell lines. Compared with the single treatment group, the red–green fluorescence ratio of JC-1 in the combined treatment group was further decreased, indicating that LND combined with ACNU could induce a more significant decrease in MMP.

As mentioned above, both glycolysis and mitochondrial function are inhibited by the combination treatment with LND and ACNU, leading to impaired energy metabolism. As shown in Figure 3E,F, LND combined with ACNU significantly decreased the intracellular adenosine triphosphate (ATP) level in SF763 and SF126 cells compared with the LND-alone group. Since intracellular energy status is associated with tumor resistance,^2,3,15 efficient de-energization by LND plus ACNU should overcome ATP-mediated MDR and DNA damage repair, improving the killing effects against glioblastoma cells. Indeed, based on the cellular uptake assay of daunorubicin, we observed that fluorescence intensity was enhanced in SF763 and SF126 cells after LND treatment (see Figure S1), indicating that LND reduced P-glycoprotein-mediated daunorubicin efflux, partly by decreasing ATP levels in SF763 and SF126 cells. In addition, using the human permeability P-glycoprotein (P-gp/ABCB1) ELISA Kit, we found that there was no significant change in the contents of P-glycoprotein after different administrations (Figure S2). These results further confirmed that the inhibition of drug efflux was induced by LND-mediated de-energization rather than by the downregulation of P-glycoprotein.

Effect of LND Plus ACNU on the Redox State of Glioblastoma Cells

The single use of LND or ACNU reduced the cellular GSH content in a dose-dependent manner (Figure 4A,B). The combination of LND and ACNU further reduced the GSH levels in both glioblastoma cell lines compared with LND or ACNU alone.

Figure 4.

Effect of LND plus ACNU on the redox state of two glioblastoma cell lines. GSH contents in (A) SF126 and (B) SF763 cells after pretreatment with LND for 24 h and treatment with ACNU for 24 h. Intracellular ROS levels in (C) SF126 and (D) SF763 cells pretreated by LND for 24 h and then treated with ACNU for 24 h using 2′,7′-dichloro fluorescent yellow diacetate (DCFH-DA) as a fluorescence probe. The results are shown as mean ± SD (n = 5 per group). *p < 0.05 indicates statistically significant differences between the groups.

As shown in Figure 4C,D, LND and ACNU alone increased reactive oxygen species (ROS) levels in a dose-dependent manner in both glioblastoma cell lines. The change trend in intracellular ROS levels was analogous to that of the GSH content after treatment with the indicated drugs, showing an inverse relationship. Compared to the single-drug group, the combined treatment markedly enhanced ROS levels, manifested by a pronounced increase in green fluorescence intensity.

GSH is involved in the metabolic detoxification of active pharmaceutical ingredients through reactions with reactive intermediates or parent drugs, leading to tumor resistance.²⁹ Intracellular ROS also contribute to a decrease in cellular GSH content and are associated with apoptosis by causing DNA oxidation and damage.³⁰ Therefore, the LND-induced decrease in intracellular GSH content and the increase in ROS levels could enhance the sensitivity of glioblastoma cells to ACNU.

LND Plus ACNU-Induced Intracellular Acidification and Downregulated MGMT Expression

As illustrated in Figure 5A,B, ACNU alone had little effect on the intracellular lactic acid levels in both cell lines. In contrast, LND alone produced a notable dose-dependent increase in intracellular lactic acid content. In particular, when LND was combined with the IC₅₀ dose of ACNU, the level of intracellular lactic acid increased.

Figure 5.

Effect of LND, ACNU, and their combination on the contents of intracellular lactic acid, the levels of MGMT expression and DNA interstrand cross-links in SF126 and SF763 cells. (A) SF126 and (B) SF763 cells were exposed to LND (0 μM, IC₂₅, and IC₅₀ dosage) for 24 h pretreatment and ACNU (0 μM, IC₂₅, and IC₅₀ dosage) for an additional 24 h treatment; MGMT expression levels in (C) SF126 and (D) SF763 cells treated with LND, ACNU, or their combination, and the quantitative analysis. Evaluation of DNA interstrand cross-links in SF126 and SF763 cells using the alkaline comet assay (head: circle, tail: box): representative electrophoretograms of (E) SF126 and (F) SF763 cells and corresponding quantitative analysis of cross-linking potential under different drug treatments. Five groups were set in the comet assay, namely, control, H₂O₂, LND, ACNU, and LND + ACNU groups. The results are shown as mean ± SD (n = 5 per group). *p < 0.05 indicates statistically significant differences between the groups (paired one-tailed t test for Western blotting analysis).

As shown in Figure 5C, almost no MGMT expression was detected in SF126 cells, consistent with the results of our previous study.³¹ In contrast, SF763 cells showed a high MGMT expression (Figure 5D). ACNU alone had a weak effect on MGMT expression, whereas LND alone exhibited a strong inhibitory effect (decrease to 7.85% of the control group). After the combination treatment with LND and ACNU, MGMT expression was almost undetectable (Figure 5D). These results further confirmed that LND inhibited MGMT expression by inducing intracellular acidification,¹⁹ thereby reducing the repair of cross-linked precursors by tumor cells and improving the antitumor activity of ACNU.

LND Increased the Formation of DNA Interstrand Cross-Links Produced by ACNU in Glioblastoma Cells

Because of LND-mediated de-energization, downregulation of MGMT expression in an acidic microenvironment, and GSH content reduction, we speculated that LND may increase the formation of DNA ICLs induced by ACNU. As shown in Figure 5E,F, compared with the H₂O₂-alone group, LND alone had almost no effect on DNA cross-linking. However, when cells were treated with ACNU alone, the comets generally had larger, brighter heads and shorter tails than the H₂O₂ control, suggesting that ACNU induced the formation of DNA ICLs. In SF126 cells (Figure 5E), the mean comet tail moment was shortened by approximately 59% after ACNU treatment. After exposure to the combination of LND and ACNU, the comet tail was further shortened and the relative tail moment was reduced by approximately 74.33%. Compared with ACNU alone, the tail moment of the combined treatment group was further reduced by approximately 15%. For SF763 cells, as shown in Figure 5F, the tail moment was reduced by approximately 63% when exposed to ACNU alone compared to the H₂O₂ group. When LND was combined with ACNU, the tail moment was further reduced by 18% compared with that of ACNU alone. These results suggested that LND improved the formation of ACNU-induced DNA ICLs.

Antitumor Effect of LND Plus ACNU on SF763 Tumor Mice Xenografts and the Safety Evaluation

To further verify whether LND plays a role in the antitumor effects of ACNU in vivo, we constructed human glioblastoma SF763 mouse xenografts. The schedule of the experiment is shown in Figure 6A. After treatment with the indicated drugs, we analyzed several indices including tumor volume, tumor weight, body weight, and tumor size. Figure 6B shows representative images of mice bearing SF763 tumor xenografts or control mice. Figure 6C shows the tumor sizes in different groups at the end of the experiment. The mean tumor volumes of saline, ACNU (15 mg/kg, low dosage), ACNU (30 mg/kg, high dosage), LND (50 mg/kg), 50 mg/kg LND + 15 mg/kg ACNU, and 50 mg/kg LND + 30 mg/kg ACNU groups were 266 ± 29, 221 ± 28, 177 ± 15, 196 ± 25, 145 ± 12, and 128 ± 13 mm³, respectively (Figure 6D). According to the statistical analysis, we found that LND, ACNU (high dose), and LND plus ACNU (low and high doses) produced significant tumor growth inhibition compared to the saline group. In particular, a low dose of ACNU alone did not efficiently suppress tumor growth. The combination of LND and ACNU showed better antitumor effects than LND or ACNU alone (Figure 6D). We also measured the tumor weight and found that the combination treatment produced the lowest mean tumor weight. To evaluate the in vivo safety profile of the combination treatment, we monitored body weight changes in mice during the experiment. As shown in Figure 6E, the average body weight of the control, saline, ACNU (15 mg/kg, low dosage), ACNU (30 mg/kg, high dosage), LND (50 mg/kg), 50 mg/kg LND + 15 mg/kg ACNU, and 50 mg/kg LND + 30 mg/kg ACNU groups was 24 ± 0.8, 21 ± 2.9, 22.4 ± 1.8, 19.9 ± 2.9, 24.7 ± 1.7, 20.4 ± 2.3, and 15.2 ± 1.2 mm³, respectively. We observed notable body weight loss in mice treated with LND plus a high dose of ACNU compared to the saline group, indicating signs of toxicity under this regimen. None of the other drug-treated mice exhibited any significant differences in body weight compared to the saline group.

Figure 6.

In vivo antitumor effect of LND in combination with ACNU in human glioblastoma SF763 mouse tumor xenografts. (A) Schedule for the entire experiment. (B) Images of control and tumor-bearing mice (red circle indicates the tumor). (C) Tumor tissue images of different formulations. (D) Tumor volume. (E) Tumor weight. (F) Mice body weight. During the treatment, tumor volume, and mouse body weight were measured twice weekly; after a 4-week treatment, the mice were sacrificed and the tumors were removed and weighted. Values in graphs were shown as mean ± SD (n = 5 per group). *p < 0.05 vs saline or control group.

In addition, the tumor mass and major organs were harvested and subjected to hematoxylin and eosin (H&E) staining to explore pathological profiles. As shown in Figure 7, clearly visible necrosis was observed in the tumor tissues after a combination treatment. No significant injury was observed in these major organs in any of the cases, except in the liver. In the liver, we observed moderate or severe vacuolization and approximately 50% necrosis in mice treated with ACNU alone or with LND plus 30 mg/kg of ACNU (Table S2). Minimal vacuolization and single-cell necrosis were observed in mice treated with LND plus 15 mg/kg ACNU.

Figure 7.

H&E staining of tumor tissues and major organs including heart, liver, spleen, lungs, and kidneys.

Overall, the combination of 50 mg/kg LND and low-dose ACNU is more appropriate for the treatment of glioblastoma in vivo. In this study, LND pretreatment did not exhibit any signs of toxicity in animals, suggesting that the combination of LND and ACNU may be a potential strategy for glioblastoma treatment in clinical settings.

Discussion

CENUs are often used to treat brain tumors because of their lipophilicity, which enables them to cross the blood–brain barrier.⁷ However, their clinical application is associated with adverse reactions such as myelosuppression and hepatotoxicity.⁷ In addition, the development of tumor resistance limits their therapeutic effects, as evidenced by MGMT-mediated repair of critical DNA ICL precursors, ATP-dependent drug efflux, and GSH/GST-mediated drug inactivation.^12,14,16 Therefore, MGMT has been frequently investigated as a resistance protein that regulates the sensitivity of tumor cells to chemotherapy drugs such as BCNU, ACNU, and TMZ.³² The prognosis of patients treated with these alkylating agents is inversely correlated with the MGMT expression level.^32,33 Therefore, it is important to effectively downregulate the activity and expression of MGMT in tumor cells. To reduce MGMT-mediated tumor resistance, two benchmark MGMT inhibitors, O⁶-4-BTG and O⁶-BG, have been approved in combination with CENUs in clinical trials.¹² However, most clinical studies have shown that the combination of MGMT inhibitors with CENUs produces strong myelosuppressive toxicity because these two MGMT inhibitors exhibit no selectivity for tumor cells, leading to enhanced sensitivity of normal tissue to chemotherapy.^12,32 Moreover, overcoming resistance in MGMT-negative (MGMT^—) tumor cells remains a challenge.³⁴

Thus, there is an urgent need to overcome CENU resistance in both MGMT⁺ and MGMT^— tumor cells while avoiding serious side effects as much as possible, which requires the discovery of new chemotherapeutic sensitizers.

Tumor cells preferentially utilize glycolysis to provide energy (approximately 60%) and blocking materials for growth and survival rather than mitochondrial OXPHOS (approximately 40%), a phenomenon known as the “Warburg effect”.^1−3,6 This reprogrammed energy metabolism is closely associated with chemo- or radio-resistance, such as the generation of a chemical reduction milieu (radio-resistance), extracellular acid microenvironment (immunosuppression), activation of DNA damage repair, and triggering of exosome release (expression of resistance proteins).³ Therefore, this unique metabolic characteristic provides a mechanism-based basis for targeting tumor energy metabolism to enhance the effects of chemotherapy. In such cases, the strategy of combining energy metabolism blockers with chemotherapy is more rational. Recently, we explored the chemo-sensitization effects of two typical glycolytic inhibitors, 3-BrPA and 2-deoxy-glucose (2-DG), on human hepatocellular carcinoma and glioblastoma cells in vitro and in vivo.³⁵⁻³⁷ These promising results prompted us to investigate the sensitizing effects of other glycolytic inhibitors, because MGMT-mediated resistance still exists in the case of 3-BrPA and 2-DG.

In this study, we selected LND as a glycolytic inhibitor to enhance the sensitivity of two glioblastoma cell lines to ACNU. LND was chosen as a sensitizer considering the following reasons: (1) LND was originally used as a class of antispermatogenic drug without the common side effects of traditional antitumor drugs;¹⁸ (2) tumor cells are the one susceptible to LND-induced de-energization as they highly rely on the glycolysis pathway, thus LND can induce corresponding biochemical effects in tumor cells, including de-energization, increasing cellular oxidative stress and inducing intracellular acidification that may inhibit MGMT;²⁰ (3) LND has shown potency as a sensitizer when in combination with other chemotherapeutic agents (e.g., nitrogen mustards, TMZ and anthracyclines) or radiotherapy for tumor killing;¹⁸ (4) LND can penetrate the blood–brain barrier, which is an indispensable premise in the clinical treatment of glioblastoma.³⁸ Furthermore, in order to extend the “HLAGR”³⁶ or “HLAGMOR”³⁷ chemo-sensitization mechanisms proposed in our previous studies, we investigated the chemo-sensitization effect of LND on ACNU in human glioblastoma cells in vitro and in vivo.

As shown in Figure 1A–F and Table S1, we observed that LND and ACNU exhibited significant synergistic toxicity in the two human glioblastoma cell lines. The combination showed the best effect on SF763 cells, whose Q values were superior to those of SF126 cells under the same treatment conditions. The underlying reason is that SF763 cells are MGMT high-expressing cells (MGMT⁺) compared to MGMT-deficient SF126 cells (MGMT^—); therefore, LND-induced MGMT downregulation (Figure 5E,F) through intracellular acidification should act on MGMT⁺ SF763 cells rather than on MGMT^— SF126 cells. For tumor cell growth inhibition, the combination of LND and ACNU also significantly decreased colony formation compared with ACNU alone (Figure 1G– J). Furthermore, LND plus ACNU also produced more apoptotic cells (Figure 1K–N) and exhibited remarkable inhibition of migration (Figure 2A–D) and invasion (Figure 2E–H) in the glioblastoma SF126 and SF763 cell lines. The suppression of migration and metastasis was also verified by Western blot analysis of MMP2 expression levels (Figure 2I–L). According to our hypothesized mechanism, the LND-sensitized tumor cells to the killing effect of ACNU could be attributed to energy inhibition; thus, we determined HK activity (Figure 3A,B) and changes in the MMP (Figure 3C,D). We found that HK activity and MMP decreased upon combination treatment; thus, de-energization (Figure 3E,F) depends on the combined inhibition of glycolysis and mitochondrial OXPHOS. Based on these results, ATP-dependent MDR, biomolecule synthesis, and DNA damage repair were partially inhibited.

GSH is a typical antioxidant that can quench intracellular ROS or react with parent drugs and their active intermediates, thereby reducing the tumor-killing effects of antitumor alkylating agents.²⁹ GSH is overexpressed in numerous tumor cells and confers resistance to chemotherapeutic agents. High GSH content is also linked to GST activity.²⁹ In another study, LND was shown to induce ROS generation by inhibiting mitochondrial complex II and decreasing GSH content in tumor cells.²⁷ Thus, the LND pretreatment should partially suppress GSH/GST-related resistance. Indeed, as shown in Figure 4, LND or ACNU alone produced a dose-dependent increase or decrease in cellular ROS/GSH content, and their combination induced a more pronounced effect. The changes in the ROS/GSH content in the combination treatment group could be attributed to mitochondrial dysfunction (MMP depolarization, Figure 4C,D), quenching of GSH by ACNU or its active intermediates, and inhibition of the pentose phosphate pathway.

Previous studies have suggested that LND inhibits the functions of monocarboxylate transporter (MCT) and mitochondrial pyruvate carrier (MPC), thereby increasing intracellular acidity, which may lead to the inhibition of GST and MGMT activities.¹⁹⁻²¹ Especially, for MGMT, no experimental evidence supports this hypothesis so far. In this study, we explored the effect of LND on the intracellular lactate content and MGMT expression levels in glioblastoma SF126 and SF763 cells. As shown in Figure 5A,B, LND alone induced intracellular acidification and its combination with ACNU had a more remarkable acidic effect. Interestingly, MGMT was almost completely downregulated by LND alone or in combination with ACNU in MGMT⁺ SF763 cells (Figure 5C,D). It is a novel and important finding that MGMT-mediated resistance can be inhibited by LND, because other glycolytic inhibitors, 3-BrPA and 2-DG, cannot inhibit or downregulate MGMT.³⁵⁻³⁷ Based on the downregulation of intracellular ATP, GSH, and MGMT, the number of lethal DNA ICLs induced by ACNU significantly increased after LND pretreatment (Figure 5E,F). Finally, the in vivo antitumor effects and safety of LND in combination with ACNU were evaluated in MGMT⁺ SF763 tumor mouse xenografts (Figures 6 and 7). The significant delay in tumor growth and minimal adverse effects justified the feasibility of combining LND and ACNU (low dose). It should be noted that the animal model (mouse subcutaneous tumor xenografts) used in this study is not the orthotopic transplantation tumor model. Tumors formed in mouse subcutaneous tumor models may not fully mimic the biological features and complex microenvironments of human brain tumors. Particularly, the presence of the blood–brain barrier may affect the penetration, effective concentration, and bioavailability of administrated drugs. Therefore, using the mouse subcutaneous tumor model for antitumor research on brain tumors may not accurately assess the efficacy of treatment methods. Thus, further in vivo studies on the mouse orthotopic transplantation tumor model are necessary.

Similar to “HLAGR” or “HLAGMOR” mechanisms,³⁵⁻³⁷ we proposed a new chemo-sensitizing mechanism of LND, namely, “HMAGOMR”, that includes (Figure 8) (1) moderate inhibition of HK activity (H); (2) mitochondrial dysfunction (M); (3) suppressing ATP-dependent drug efflux (A); (4) changing redox homeostasis to inhibit GSH-mediated drug inactivation (G) and increasing intracellular oxidative stress (O); (5) downregulating MGMT expression through intracellular acidification (M); and (6) partial inhibition of energy-dependent DNA repair (R).

Figure 8.

Molecular mechanism of LND increasing cytotoxicity of ACNU to tumor cells and overcoming drug resistance in tumor cells.

Conclusions

In conclusion, LND has great potential as an efficient potentiator of ACNU in vitro and in vivo, especially in MGMT⁺ human glioblastoma cells, as we verified for the first time that LND downregulates MGMT expression by inducing intracellular acidification. The tumor resistance should mainly be inhibited at three key points: DNA damage repair, drug efflux, and drug inactivation. Thus, LND reverses ACNU resistance in human glioblastoma cells and mouse xenografts through the combined effects of energy blockade, redox homeostasis disruption, and MGMT downregulation. It should be noted that the dose used in cell culture and in vivo mouse models cannot be directly transformed into a clinical dose for human use. The most important finding was the feasibility of a therapeutic strategy combining LND with ACNU or other alkylating agents, which warrants further exploration in preclinical (primary intracranial glioblastoma in animals) and clinical settings.

Materials and Methods

Chemicals, Reagents, and Drugs

Please see details in the Supporting Information.

Cell Culture Studies

Two human glioblastoma cell lines, SF763 (RRID: CVCL_6949) and SF126 (RRID: CVCL_1688), were purchased from Peking Union Medical College (Beijing, China) and cultured in MEM-EBSS medium supplemented with 10% FBS, 1% penicillin (100 U/mL), and streptomycin (100 μg/mL). The cells were cultured under a humidified atmosphere of 37 °C and 5% CO₂/95% air. For the drug treatment, cells in the exponential growth phase were inoculated at an adjusted density into different culture flasks or well plates. When the cells reached confluence, they were washed twice with phosphate-buffered saline (PBS) and treated with fresh medium containing LND or ACNU.

Animal Studies

All animal studies were approved by the Science & Technology Ethics Committee of Beijing University of Technology (Approval No.: HS202106002) and conducted in accordance with the U.K. Animals (Scientific Procedures) Act 1986 and associated guidelines (EU Directive 2010/63/EU for animal experiments). We obtained male BALB/c nude mice at 8 weeks of age (18–20 g) from Beijing Vital River Laboratory Animal Technology Co., Ltd., to evaluate their antitumor activity. Mice were maintained under the condition of 25 °C/50% relative humidity, natural light/dark, and with free access to food and water. Appropriate measures were taken to minimize animal discomfort, and tumor implantation and drug administration were performed using sterile surgical techniques. Every effort was made to reduce the number and suffering of experimental animals to comply with the 3Rs principle.

Cytotoxicity Assay

The cytotoxicity was assessed in 96-well plates using a methyl thiazolyl tetrazolium (MTT) assay.³⁹ The absorbance was measured at 560 nm by using a Multiskan FC microplate reader (Thermo Scientific, Waltham, Massachusetts, USA). Whether LND and ACNU have synergistic effects on the cytotoxicity (1 – survival rate) of glioblastoma cells were determined using the formula⁴⁰Q = E_(A+B)/(E_A + E_BE_A × E_B), in which Q = 0.85–1.15 means additive, Q < 0.85 means antagonism, and Q > 1.15 means synergism, where E_(A+B) represents the combined inhibitory rate and and E_A and E_B represent the individual inhibitory rates of drugs A and B, respectively.

Colony Formation, Wound Scratch, and Cell Invasion Assays

Please see details in the Supporting Information.

HK Activity Assay

The HK activity in cell supernatant was determined using the HK assay kit and expressed as nmol per minute per mg of protein. Absorbance was measured at 340 nm using an ultraviolet spectrophotometer (U-3010, Hitachi, Japan).

Determination of Intracellular ATP Levels

Intracellular ATP levels were determined using a CellTiter-Lumi assay kit. The luminescence signal was measured by a Multimode Plate Reader (PerkinElmer, Waltham, Massachusetts, USA).

Measurement of Cellular GSH Content and ROS Level

Cellular GSH was detected by using a reduced GSH assay kit. Absorbance was recorded at 405 nm, and the protein concentration was determined using a bicinchoninic acid assay (BCA) assay kit.⁴¹ Cellular GSH content was expressed as μmol per gram of protein (μmol/g of protein).

Intracellular ROS generation was monitored by staining cells with 2′,7′-dichloro fluorescent yellow diacetate (DCFH-DA). DCFH-DA is a nonfluorescent dye that can be hydrolyzed by cellular esterases to generate DCFH, which can then be oxidized by ROS to yield fluorescent DCF.⁴² The fluorescence intensity of DCF inside the cells was detected using a fluorescence microscope (excitation wavelength, 488 nm; emission wavelength, 525 nm; Olympus IX51, Tokyo, Japan).

Intracellular Lactic Acid Assay

Intracellular lactic acid levels were determined using a lactic acid assay kit. A 20 μL aliquot of lysate was taken from each sample for intracellular lactic acid quantitation, according to the manufacturer’s protocol.

Cell Apoptosis Analysis

The apoptosis-inducing capabilities of LND or ACNU, alone or in combination, were assessed using an Annexin V-FITC/PI apoptosis detection kit. Please see details in the Supporting Information.

MMP Using JC-1 Staining

JC-1 dye is an ideal fluorescent probe for the detection of MMP (represented as Δψm).⁴³ JC-1 exhibits an aggregated form that emits a red fluorescence in healthy cells. When the mitochondrial membrane is destroyed, JC-1 exists as a monomer and emits a green fluorescence, indicating mitochondrial depolarization. SF126 and SF763 cells were plated in six-well plates at a density of 2 × 10⁵ cells/well for 24 h and then treated with LND for 24 h and ACNU for another 24 h. After discarding the old medium, we added 2 mL of fresh medium solution containing JC-1 and Hoechst 33342 to each well and allowed incubation at 37 °C for 20 min. Next, 1 mL of fresh culture medium was added to wash the cells. After three washes, the stained cells were observed under a fluorescence microscope.

Western Blotting Analysis

Two human glioblastoma cell lines were inoculated into T25 culture flasks at an adequate density and cultured until the cell density reached approximately 90% confluence, followed by treatment with the indicated concentrations of LND and ACNU for 24 h. After lysis, centrifugation, and denaturation, cell protein extracts (60 μg per sample) were electrophoresed on a 12% sodium dodecyl-sulfate polyacrylamide gel electrophoresis (SDS-PAGE) gel and transferred to a polyvinylidene fluoride (PVDF) membrane (Millipore Inc., Billerica, Massachusetts, USA). The membrane was incubated for 12 h at 4 °C with a primary antibody at 1:1000 (rabbit monoclonal anti-MGMT/anti-MMP2 antibodies) or 1:10,000 (mouse monoclonal anti-β-actin antibody). After washing in PBST and PBS (15 min/time) for three times, the membrane was incubated with preadsorbed goat antimouse or antirabbit IgG second antibody (IRDye 800CW) at room temperature for 1 h. The washing was repeated using PBST and PBS; finally, the Odyssey infrared imaging system (LI-COR Biosciences, Lincoln, Nebraska, USA; RRID:SCR_013430) was used to detect the protein band. β-Actin was regarded as a reference control protein.

Determination of DNA Interstrand Cross-Links Using the Alkaline Comet Assay

Single-cell gel electrophoresis⁴⁴ was used to detect the levels of DNA ICLs in SF126 and SF763 cells after the drug combination treatment. After drug treatment, an aliquot of 100 μL of H₂O₂ solution (final concentration, 0.3% in PBS) was added to the cells for 20 min on ice to induce DNA strand breaks. Control groups were treated with PBS. The reaction was quenched by adding 1% (v/v) dimethyl sulfoxide (DMSO). Subsequent steps were performed using a DNA damage detection kit. The cells were embedded in agarose and electrophoresed under alkaline conditions, followed by PI staining. Images were captured using a fluorescence microscope, and single cells revealed a comet-like form. Here, cross-linked DNA migrates slowly as the round comet “head” whereas the small pieces of DNA generated by strand breaks yield the “tail.” Analyses were performed using Comet Assay Software Project. The cross-linking rate was calculated as the decrease in the percentage of olive tail moment (OTM) in cells treated with the indicated drugs and H₂O₂ relative to cells treated with only H₂O₂.^44,45 The percentage decrease representing the level of ICLs in treated cells was calculated as follows:

where OTM_control, OTM_H2O2, OTM_drug, and OTM_drug+H2O2 represent the OTMs of cells treated with neither drug nor H₂O₂, with only H₂O₂, with only drug, and with both drug and H₂O₂, respectively.

In Vivo Xenograft Studies

The in vivo antitumor effect of LND in combination with ACNU was evaluated in BALB/c nude mice (male, 8 weeks old) with resistant SF763 tumors. Approximately 3 × 10⁶ SF763 cells in the logarithmic phase were subcutaneously inoculated into the right armpit of each mouse. When the tumor volume was approximately 100 mm³, nude mice bearing uniform tumor size were divided randomly into six groups (n = 5, in each group) and ip administrated with indicated drugs: group A (PBS); group B (ACNU, 15 mg/kg); group C (ACNU, 30 mg/kg); group D (LND, 50 mg/kg); group E (50 mg/kg LND + 15 mg/kg ACNU); group F (50 mg/kg LND + 30 mg/kg ACNU); and normal group (n = 3) was normal feeding, without tumor inoculation and drug treatment. In the combination treatment groups, animals were pretreated with LND for 24 h prior to ACNU administration. The mice were administered the drug twice weekly. Body weight was measured twice weekly, and tumor volume (V) was calculated as V = (length × width²)/2. Nude mice were sacrificed after 4-week treatment. Primary tumors were excised and weighed. Subsequently, the tumors and the main organs, including the heart, liver, spleen, lung, and kidney, were fixed with 4% paraformaldehyde for further analysis.

Histopathology Analysis

Tumor tissues and major organs were fixed in 4% paraformaldehyde in phosphate buffer (pH7.4). The paraffin-embedded tissues were sectioned, deparaffinized, and hydrated. All slides were stained with hematoxylin and eosin (H&E) and imaged under a light microscope. The Suzuki score⁴⁶ was used to evaluate the degree of liver injury.

Glossary

Abbreviations

CENUs

chloroethylnitrosoureas

MGMT

O⁶-methylguanine-DNA methylguanine

ICLs

interstrand cross-links

LND

lonidamine

ACNU

nimustine

GSH

glutathione

ROS

reactive oxygen species

BCNU

carmustine

GST

glutathione-S-transferases

MDR

multidrug resistance

ABCB1

p-glycoprotein

TMZ

temozolomide

O⁶-BG

O⁶-benzylguanine

O⁶-4-BTG

O⁶-(4-bromothenyl)guanine

3-BrPA

3-bromopyruvate

2-DG

2-deoxyglucose

SDS-PAGE

sodium dodecyl-sulfate polyacrylamide gel electrophoresis

HK

hexokinase

MMP

mitochondrial membrane potential

MTT

methyl thiazolyl tetrazolium

ATP

adenosine triphosphate

DCFH-DA

2′,7′-dichloro fluorescent yellow diacetate

PBS

phosphate-buffered saline

OTM

olive tail moment

SD.

standard deviation

BCA

bicinchoninic acid assay

PVDF

polyvinylidene fluoride

DMSO

dimethyl sulfoxide

Q

combination index

MMP2

matrix metalloproteinase-2

PPP

pentose phosphate pathway

MCT

monocarboxylate transporter

MPC

mitochondrial pyruvate carrier

Supporting Information Available

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsptsci.4c00085.

• Chemicals, reagents, and drugs; colony formation assay; wound scratch assay; cell invasion assay; cell apoptosis analysis; data and statistical analyses; combination index (Q) of LND and ACNU in SF763 and SF126 cells; histological assessment of liver damage using the Suzuki score; and SF763/SF126 cells were treated with LND (control) or daunorubicin or LND + daunorubicinfor 24 h, and cell fluorescence was detected by the live cell imaging system; and the amounts of P-glycoprotein in SF763 and SF126 cells after treatment with LND alone, ACNU alone, or LND plus ACNU (PDF)

Author Contributions

^# Y.H. and P.W. contributed equally to this work.

Author Contributions

Y.H.: methodology, formal analysis, investigation, writing—original draft; P.W.: formal analysis, investigation, resources, writing—review and editing; T.F.: formal analysis, writing—original draft; N.Z.: visualization, software; L.Z.: resources, supervision; R.Z.: writing—review and editing, supervision; and G.S.: conceptualization, writing—review and editing, supervision, funding acquisition.

The authors declare no competing financial interest.