Altered Magnesium Environments Restrict Colorectal HT-29 Spheroid Growth by Disturbing Cellular Mg Homeostasis.

1. Introduction
Dysregulated magnesium (Mg2+) homeostasis is increasingly linked to colorectal cancer (CRC) initiation and progression [1,2], yet its mechanistic contribution remains unresolved and highly context-dependent. Epidemiological evidence reflects this complexity: while an insufficient dietary magnesium intake elevates the risk of CRC [3,4], other studies paradoxically associate excessive intake with a greater incidence [5]. At the tissue level, CRC specimens frequently exhibit heightened intracellular Mg2+ compared with adjacent normal mucosa [6,7], likely driven by increased uptake through Transient Receptor Potential Melastatin (TRPM)6, TRPM7, and Magnesium Transporter 1 (MagT1), coupled with reduced efflux via Cyclin M4 (CNNM4) [8,9,10,11]. Elevated intracellular Mg2+ correlates with enhanced doxorubicin (DXR) resistance, cancer stemness, spheroid (SP) stability, and invasive behavior in CRC cells [12,13]. However, in contrast to tumor profiles, experimental exposure to supraphysiologic Mg2+ (15–30 mM) induces cell-cycle arrest and apoptosis in CRC cells [14]. Despite suggesting therapeutic utility, such concentrations (>5 mM) constitute severe clinical hypermagnesemia in vivo, causing marked cardiovascular and neurological toxicity [15]. This discrepancy between the anti-tumor effects of extreme Mg2+ in vitro and their systemic hazards in vivo highlights the dualistic and poorly defined nature of Mg2+ in tumor biology. It underscores the need to identify clinically relevant thresholds.
Our previous work clarified this paradox by showing that HT-29 cells grown as SPs exhibit higher membrane TRPM6 and TRPM6/7 expression, increased Mg2+ influx, elevated intracellular Mg2+, and greater stemness and migratory capacity than their parental adherent counterparts under control (1.0 mM Mg2+) conditions [13]. Notably, pharmacological inhibition of TRPM6 and TRPM6/7 sharply reduced SP Mg2+ influx, intracellular Mg2+, cancer stemness, SP stability, and migration [13]. However, a central and unresolved question persists: how do sustained low and high Mg2+ environments directly modulate cellular Mg2+ homeostasis and CRC SP progression? Our previous study showed that pharmacological inhibition disrupts SP growth, but the effects of physiological extremes of extracellular Mg2+ and their underlying molecular mechanisms remain unexplored.
The tumor microenvironment (TME) comprises a complex network of cellular and non-cellular components that collectively drive tumor progression by supporting unchecked proliferation, evasion of apoptosis, and metastasis. Accordingly, targeting the TME has become a promising strategy for treating solid tumors [16]. Modulating Mg2+ within the TME has recently gained attention. Localized Mg2+-release systems using microsphere-encapsulated hydrogels can remodel the immune microenvironment and enhance the efficacy of immune checkpoint blockade in CRC [17]. Likewise, Mg-based biomaterials show potential for improving tumor treatment [18], highlighting the therapeutic value of precisely controlling tumoral Mg2+ fluxes. However, the direct mechanistic effects of Mg2+ on core cancer-cell hallmarks within the solid TME remain largely undefined.
To address this gap, we examined how extracellular Mg2+ availability shapes CRC SP progression and Mg2+ homeostasis. We used HT-29 SPs as a physiologically relevant 3D model that recapitulates the structural and microenvironmental features of solid tumors [19]. HT-29 cells, derived from colorectal adenocarcinoma (ATCC HTB-38), form robust SPs that mimic key aspects of CRC biology, including cell–cell interactions, nutrient gradients, and drug resistance mechanisms, providing a robust platform for probing ion dysregulation. We hypothesized that both low and high Mg2+ would disrupt Mg2+ homeostasis in CRC SPs, leading to structural disintegration and cell death through dysregulated TRPM6/7 signaling and mitochondrial dysfunction. Our objective was to determine how sustained deviations from physiological Mg2+ levels affect SP integrity, cell survival, Mg2+ transporter function, cellular Mg2+ homeostasis, and post-translational regulation of TRPM6/7 channels. We analyzed major Mg2+ transporters, including TRPM6, TRPM7, MagT1, CNNM4, and mitochondrial inner-membrane magnesium transporter 2 (Mrs2). Confocal and 4D microscopy enabled visualization of Mg2+ influx and channel localization, respectively, which we integrated with high-resolution global, phospho, and oxidoproteomic profiling to identify underlying signaling and stress pathways. This multi-modal strategy was designed to define how Mg2+ dysregulation affects CRC SPs and to evaluate its therapeutic relevance.

2. Results

2.1. Low and High Mg2+ Conditions Disrupted HT-29 SPs
To determine how extracellular Mg2+ affects SP stability, HT-29 SPs were generated and then maintained in control (1.0 mM Mg2+), low Mg2+ (0.6, 0.4, 0.2 mM), or high Mg2+ (1.5, 2.5, 5 mM) media, representing mild to severe deviations from physiological levels [15]. Under physiological conditions, SPs retained a compact spherical morphology with a dense core and defined borders (Figure 1a), and their area (Figure 1b) and structural integrity (Figure 1d) increased steadily over 48 h, confirming the robust growth of our CRC SP model. In contrast, both low and high Mg2+ induced rapid structural disruption (Figure 1a). By 24 h, peripheral cell shedding indicated reduced intercellular cohesion, and by 48 h, moderate and severe Mg2+ deviations caused extensive disaggregation and loss of spherical architecture (Figure 1a). Quantitative analyses demonstrated that both conditions significantly and dose-dependently reduced SP area (Figure 1b,c) and integrity scores (Figure 1d,e) in comparison to control conditions. These findings indicate that deviations from physiological extracellular Mg2+ in either direction compromise the structural integrity and growth of CRC SPs.

2.2. Low and High Mg2+ Conditions Induced Cell Death in HT-29 SPs
Structural disintegration of SPs under non-physiological Mg2+ was accompanied by peri-spheroidal vesicles, consistent with a cellular stress response. We first assessed the baseline metabolic activity and found that SP-derived cells exhibited significantly higher activity—a proxy for viability—than adherent parental HT-29 cells at comparable passage (Figure 2a,b), confirming their elevated proliferative and metabolic state. We subsequently investigated the functional consequence of Mg2+ dysregulation. Both low and high Mg2+ conditions markedly reduced the viability of parental and SP-derived cells in a dose-dependent manner (Figure 2a,b). Although SPs originate from a more robust population, their viability under Mg2+ stress fell below that of the corresponding treated parental groups, indicating that the SP phenotype, despite its resilience under physiological conditions, is more vulnerable to deviations in extracellular Mg2+.
To elucidate the mechanisms driving the observed viability loss, we examined key apoptotic regulators, including tumor suppressor protein p53 (p53), B-cell lymphoma 2 (Bcl-2), Bcl-2-associated X protein (Bax), and caspase-3 (Figure 3a) [20,21]. Both moderate low and high Mg2+ markedly increased p53 in parental cells (Figure 3b), whereas SP-derived cells showed a stronger induction of p53 under Mg2+ stress, exceeding both control SPs and stressed parental cells. This heightened responsiveness extended to Bcl-2 regulation. Under physiological conditions, SPs upregulated Bcl-2 relative to parental controls (Figure 3c), but low and high Mg2+ conditions significantly suppressed Bcl-2 in both models. As a result, SPs consistently exhibited lower Bcl-2 than treated parental cells, indicating an impaired anti-apoptotic response. A pro-apoptotic shift in SPs was further supported by a significant increase in Bax (Figure 3d), which remained unchanged in parental cells. Correspondingly, Mg2+ stress activated the apoptotic effector pathway, with total and cleaved caspase-3 significantly elevated in all treated groups (Figure 3e,f), and more strongly in SPs. Because apoptosis entails loss of mitochondrial membrane integrity, we assessed mitochondrial membrane potential (ΔΨm) using the JC-1 assay (Figure 3g). Control SPs displayed higher ΔΨm than parental cells, consistent with greater metabolic activity. However, both low and high Mg2+ significantly reduced ΔΨm in both models, with a more severe decline in SPs. Together, these findings demonstrate that although Mg2+ stress triggers apoptosis in both models, SPs exhibit a markedly heightened susceptibility, characterized by amplified p53 induction, reduced Bcl-2 expression, elevated Bax levels, enhanced caspase-3 activation, and a pronounced mitochondrial collapse.
Given that p53 can induce autophagy under cellular stress [22] and that it was markedly upregulated in Mg2+-stressed SPs, we next examined the autophagic response. We assessed key markers, including Unc-51-like kinase 1 phosphorylated at Serine 556 (pS556 ULK1), which initiates autophagy, and Sequestosome-1 (p62/SQSTM1), which is degraded during effective autophagic flux (Figure 4a). Both low and high Mg2+ conditions significantly elevated pS556 ULK1 (Figure 4b) and p62/SQSTM1 (Figure 4c) relative to their control conditions. This induction was substantially stronger in SP-derived cells than in parental cells exposed to the same Mg2+ stress.
Collectively, our findings show that the SP phenotype heightens sensitivity to Mg2+ dysregulation. This susceptibility is characterized by a dominant pro-apoptotic program, marked by reduced Bcl-2 and increased caspase activation, accompanied by a parallel, amplified autophagic response, ultimately resulting in extensive cell death.

2.3. Low and High Mg2+ Conditions Altered Mg2+ Homeostasis in HT-29 SPs
To assess early disturbances in cellular Mg2+ regulation before overt cell death, parental and SP HT-29 cells were exposed to moderate low and high Mg2+ conditions for 24 h. Intracellular Mg2+ dynamics were tracked using ΔF/F fluorescence, with baseline acquisition during the initial 30 s, followed by a 20 mmol/L MgSO4 challenge from 40 to 160 s. After MgSO4 addition, the parental cells showed a rapid rise in Mg2+ that plateaued (Figure 5a). As previously reported [13], SPs displayed a significantly greater Mg2+ influx rate under physiological conditions (Figure 5a,b). Remarkably, Mg2+ stress uncovered a core defect in the SP phenotype. Whereas parental cells exhibited only a mild influx reduction, SPs showed a severe suppression of Mg2+ influx under both low and high Mg2+ conditions (Figure 5b). We then assessed steady-state free intracellular Mg2+ concentrations. Reflecting their impaired influx, both Mg2+ stresses significantly reduced free Mg2+ in SPs (Figure 5c). In contrast, parental cells retained regulatory responses, with low Mg2+ lowering and high Mg2+ elevating free Mg2+. Total cellular Mg2+ content showed parallel trends: low Mg2+ decreased, and high Mg2+ increased total Mg2+ in parental cells (Figure 5d). Control SPs contained substantially more total Mg2+ than parental cells, consistent with their elevated basal influx, and this baseline and this elevation were similarly modulated by extracellular Mg2+. Overall, these findings reveal a fundamental divergence in Mg2+ handling: parental cells preserve influx regulation to buffer extracellular shifts, whereas SPs, despite higher baseline capacity, lose regulatory control and experience pronounced Mg2+ homeostatic disruption under stress.

2.4. Low and High Mg2+ Conditions Altered Magnesiotropic Protein Expression
To elucidate the molecular basis of Mg2+ dysregulation in SPs, we assessed the expression and subcellular distribution of major Mg2+ transporters such as TRPM6, TRPM7, MagT1, and the efflux protein CNNM4 in total cell lysates, membranes, and cytosolic fractions (Figure 6a–c). Mirroring CRC tissue profiles [6,7], SPs under physiological conditions showed a marked shift toward Mg2+ acquisition. Relative to parental cells, SPs displayed substantial increases in total, membrane, and cytosolic TRPM6 (Figure 6d–f) and TRPM7 (Figure 6g–i). Only cytosolic MagT1 was elevated in control SPs, with no change in total or membrane pools (Figure 6j–l), indicating that enhanced membrane abundance of TRPM6/7 and likely their heterodimers primarily drives the heightened Mg2+ influx. Concurrently, SPs exhibited reduced total and membrane-bound CNNM4 but increased cytosolic CNNM4 (Figure 6m–o), a redistribution that aligns with their elevated Mg2+ influx and total Mg2+ content. Parental cells showed coherent, condition-specific adaptations. Under moderate low Mg2+, they upregulated TRPM6, TRPM7, and MagT1 while suppressing CNNM4, thereby enhancing net Mg2+ uptake (Figure 6d–o). Under moderate high Mg2+, they robustly induced total, membrane, and cytosolic CNNM4 to promote Mg2+ efflux (Figure 6m–o), reflecting intact epithelial homeostasis. In contrast, SPs failed to mount these selective responses. Both low and high Mg2+ conditions triggered downregulation of TRPM6, TRPM7, MagT1, and CNNM4 across all fractions (Figure 6d–o). This nonselective suppression of Mg2+ transport reveals a fundamental breakdown in regulatory control, highlighting the SP phenotype’s inability to maintain Mg2+ homeostasis under stress.

2.5. Mg2+ Dysregulation Disrupts TRPM6/7 Heterodimer Stability in HT-29 SPs
The channel permeability of TRPM6 and TRPM7 is suppressed by intracellular Mg2+ and Mg·ATP [23,24]. Critically, heterodimeric TRPM6/7 complexes are comparatively resistant to this inhibition [23,24], enabling sustained directional Mg2+ absorption. To determine whether impaired Mg2+ influx and reduced viability under Mg2+ stress were linked to reduced channel dimerization at the plasma membrane, we immunoprecipitated TRPM6 (IP-TRPM6) from membrane fractions and analyzed associated proteins by Western blot. Comparable TRPM6 levels in IP fractions confirmed uniform loading (Figure 7a). Subsequent reprobing for TRPM7 revealed a markedly higher TRPM6/7 association in control SPs compared to parental cells, indicating elevated basal heterodimerization (Figure 7b). Both low and high Mg2+ conditions significantly decreased co-precipitated TRPM7 in SPs, demonstrating loss of membrane-bound heterodimers. Under identical Mg2+ conditions, SPs consistently exhibited weaker TRPM7 signals than parental cells, confirming their greater susceptibility to Mg2+-dependent destabilization of TRPM6/7 complexes. We further illustrated this disruption using high-resolution imaging with the ZEISS LSM910 Lightfield 4D microscope (Figure 7c): control SPs displayed strong co-localization of TRPM6 (green) and TRPM7 (red). This peri-spheroidal expression of TRPM6, TRPM7, and their heterodimeric complexes was markedly reduced in both low and high Mg2+ conditions, visually corroborating the biochemical findings.

2.6. Mg2+ Dysregulation Preferentially Suppresses the Mitochondrial Mg2+ Importer Mrs2 in SPs
The mitochondrial Mg2+ channel Mrs2 is a central determinant of mitochondrial Mg2+ homeostasis and thereby of cellular bioenergetics and survival [25], with its expression inversely associated with stress-induced cell death [26]. We therefore examined Mrs2 regulation under Mg2+ stress in our CRC models. SPs, which display elevated metabolic activity and greater viability, exhibited significantly higher basal Mrs2 levels than parental cells (Figure 7d). Both low and high Mg2+ conditions significantly suppressed Mrs2 expression in each model. This suppression was consistently stronger in SPs compared to the same Mg2+ conditions, paralleling their enhanced susceptibility to cell death. These findings implicate impaired mitochondrial Mg2+ import via Mrs2 as a key driver of metabolic failure in Mg2+-stressed SPs. The amplified loss of Mrs2 provides a direct mitochondrial mechanism underlying the heightened vulnerability and extensive cell death induced by Mg2+ dysregulation in the SP phenotype.

2.7. Nano-LC-MS/MS Analysis of Membranous TRPM6 and TRPM7
To characterize the composition of Mg2+ transport complexes, we conducted nano-liquid chromatography tandem mass spectrometry (nano-LC-MS/MS) on membrane proteins isolated through TRPM6 immunoprecipitation [13,27]. Sequence analysis verified TRPM6 (UniProtKB: Q9BX84) and its partner TRPM7 (UniProtKB: Q96QT4) in all groups, and comparable exponentially modified protein abundance index (emPAI) values (TRPM6: parental 0.10, SP 0.12, SP + low-Mg 0.11, SP + high-Mg 0.10; TRPM7: parental 0.11, SP 0.11, SP + low-Mg 0.12, SP + high-Mg 0.10) confirmed equivalent loading.
We then examined functionally significant phosphorylation sites relevant to channel function. Particular attention was given to TRPM6 Ser141 (S141), a key determinant of heterodimer membrane localization [28]. S141 phosphorylation was detected in parental cells and control SPs but was lost entirely under both low and high Mg2+ conditions (Table 1), providing a mechanistic explanation for reduced TRPM6/7 heterodimer. Additionally, regulatory sites governing channel gating were also identified. Phosphorylation of S1252, a modification that increases permeability [29], occurred exclusively in control SPs, consistent with their elevated Mg2+ influx, and was absent in parental cells and Mg2+-stressed SPs. In contrast, phosphorylation of T1851, a known inhibitory site [30], was detected only in low and high Mg2+ SPs, confirming active suppression of Mg2+ influx under stress.
Globally, the total number of phosphorylated TRPM6 residues paralleled Mg2+ influx capacity [13]: 200 in control SPs (high influx) versus 171, 172, and 164 in parental, low Mg2+, and high Mg2+ groups (reduced influx), respectively. The relative abundance of these phosphorylated residues further confirmed distinct phosphoproteomic patterns across all experimental conditions (Figure 8a). These phosphoproteomic patterns reveal how non-physiological Mg2+ levels impair channel function by reshaping TRPM6’s phosphorylation landscape, altering its membrane localization and permeability.
We extended our phosphoproteomic analysis to TRPM7 and identified regulatory sites that closely mirrored the TRPM6 profile. Phosphorylation of S138, which governs TRPM7 membrane localization [28], and S1360, which stabilizes its plasma membrane presence [31], was detected in parental cells and control SPs but disappeared in Mg2+-stressed SPs (Table 2), consistent with reduced membranous TRPM7.
A global survey further showed that TRPM7 carried the highest number of phosphorylated residues in control SPs (153), followed by parental cells (138) and high Mg2+ SPs (136), with low Mg2+ SPs exhibiting the fewest (107). The relative abundance of these phosphorylated sites confirmed distinct condition-specific phosphoproteomic signatures (Figure 8b). This phosphorylation burden on TRPM7 correlated positively with Mg2+ influx across all groups [13].
Together, these phosphoproteomic findings define a unified mechanism in which non-physiological Mg2+ environments disrupt TRPM6/7 heterodimers by reconfiguring the phosphorylation states of both subunits, thereby modulating their membrane localization, stability, and channel permeability.
Our MS/MS analysis also revealed extensive methionine oxidation (MetO) across TRPM6 and TRPM7 (Table 3). We concentrated on the M1755 residue of TRPM6, as its oxidation is known to impede channel function [32]. Consistent with this mechanism, M1755 oxidation was strongly enriched in SPs exposed to low and high Mg2+, providing a direct post-translational basis for their reduced Mg2+ influx.
Global MetO was likewise elevated, with the low Mg2+ group exhibiting the highest number of oxidized residues (87) relative to control SPs (70), parental cells (57), and high Mg2+ SPs (64), indicating substantial Mg2+-deprivation-induced oxidative stress on channel proteins. This pronounced elevation under low Mg2+ condition indicates substantial oxidative stress on channel proteins induced by Mg2+ deprivation. The relative abundance of these oxidized residues confirmed distinct, condition-specific oxidation signatures for both TRPM6 and TRPM7 (Figure 8c,d).
Collectively, these phospho- and oxidoproteomic findings delineate a dual mechanism by which non-physiological Mg2+ conditions destabilize the TRPM6/7 heterodimer: perturbed phosphorylation reshapes membrane localization and permeability, and at the same time, oxidative modifications further diminish channel activity.

3. Discussion
Dysregulated cellular Mg2+ homeostasis in CRC is well-established [6,7,8,9,10,11,12,13]. Our study advances this understanding by clarifying the mechanisms involved using a three-dimensional SP model. Consistent with CRC tissue observations [6], SPs displayed markedly elevated intracellular Mg2+ relative to parental cells. This increase stemmed from enhanced Mg2+ influx driven by upregulated total and membrane-bound TRPM6, TRPM7, and TRPM6/7, together with reduced expression of the efflux transporter CNNM4. The increased Mg2+ influx in SPs seems to be related to the higher number of membrane-localized TRPM6/7 heterodimers. These complexes exhibit diminished sensitivity to inhibition by intracellular Mg2+ and Mg·ATP, allowing for persistent Mg2+ uptake [23,24]. Our proteomic data offer direct molecular support for this activated state. In control SP, we identified TRPM6 phosphorylation at Ser141 (promoting heterodimerization [28]) and Ser1252 (increasing channel permeability [29]), along with TRPM7 phosphorylation at Ser138 (regulating membrane localization [28]) and Ser1360 (stabilizing plasma membrane residency [31]). This distinct phospho-pattern likely defines the constitutive Mg2+-acquisitive phenotype of CRC SPs.
The heightened viability of the SP phenotype reflects a multifaceted pro-survival program defined by increased Mg2+ content, elevated expression of the mitochondrial Mg2+ channel Mrs2, and high levels of the anti-apoptotic protein Bcl-2. Mrs2-dependent mitochondrial Mg2+ uptake is crucial for efficient ATP synthesis and serves as a physiological inhibitor of the mitochondrial calcium uniporter (MCU) [33]. By restricting MCU activity, elevated intramitochondrial Mg2+ prevents Ca2+ overload, thereby blocking activation of the mitochondrial permeability transition pore (MPTP) [34], a process that induces ΔΨm collapse, halts ATP synthesis, and triggers apoptosis [34]. Consequently, the high basal Mg2+ and Mrs2 levels in SPs establish a protective mitochondrial environment. This is supported by their markedly higher ΔΨm, a direct measure of the proton gradient magnitude and coupled ATP-generating capacity [26]. This enhanced ΔΨm, together with the reported ability of increased Mg2+/Mrs2 to suppress caspase activity [35], renders the SP phenotype intrinsically apoptosis-resistant and confers protection against inducers such as doxorubicin (DXR) and staurosporine (STS) [12,26]. Complementing this mitochondrial defense, SPs displayed elevated basal Bcl-2 expression. In cancer cells, Bcl-2 overexpression blocks mitochondrial translocation and activation of pro-apoptotic Bax, thereby directly inhibiting the intrinsic apoptotic cascade and promoting resistance to agents like STS [36]. Thus, the integrated profile of elevated Mg2+ influx, increased Mrs2, enhanced ΔΨm, and upregulated Bcl-2 generates a potent state of augmented energy metabolism and apoptotic resistance under physiological conditions. This strong basal framework not only reinforces the validity of our model but also highlights the profound cytotoxic consequences that arise when Mg2+ dysregulation perturbs this finely balanced system.
Our investigation indicates that the SP phenotype, despite its strong baseline, displays heightened sensitivity to extracellular Mg2+ dysregulation compared with parental cells. This vulnerability was reflected in the marked loss of viability and the amplified activation of apoptotic and autophagic markers under low and high Mg2+ conditions. We propose that this hypersensitivity arises from the convergence of metabolic and oxidative stresses. First, an Mg2+ imbalance is known to trigger cellular and mitochondrial oxidative damage capable of directly initiating cell-death pathways [37]. Although we did not directly assess reactive oxygen species, our oxidoproteomic data provide persuasive indirect evidence: SPs exposed to Mg2+ stress showed a substantially higher number of oxidatively modified methionine residues (Table 3), indicating increased protein oxidation. Second, the pronounced loss of ΔΨm in stressed SPs indicates a profound failure of energy metabolism, likely impairing ATP production. Thus, the SP’s sensitivity appears rooted in an inability to withstand the dual pressures of Mg2+ imbalance: (1) an elevated oxidative load demonstrated by protein damage and (2) a collapse of mitochondrial bioenergetics. This combination likely produces an irreversible metabolic breakdown, ultimately driving the extensive cell death observed.
Our previous work showed that pharmacological inhibition of TRPM6 and TRPM6/7 suppresses Mg2+ content, thereby impairing the stemness, progression, and metastatic capacity of HT-29 SPs [13]. The present study expands on these findings by evaluating the impact of both physiological extremes—Mg2+ deprivation (low Mg2+) and Mg2+ excess (high Mg2+). This represents a novel contribution, as previous studies focused on pharmacological inhibition rather than the physiological consequences of Mg2+ imbalance. Our results reveal a nuanced and context-dependent toxicity profile. Although low Mg2+ condition significantly impaired SP progression, high Mg2+ exerted an even more disruptive influence across several parameters, including a sharper reduction in SP area and density, lower viability, and greater induction of p53, cleaved caspase-3, pS556 ULK1, and p62/SQSTM1. These findings indicate that supra-physiological Mg2+ levels are not merely inhibitory but can provoke a more severe stress response, leading to accelerated metabolic and structural collapse. Thus, both removal of the essential ion and its pathological excess effectively compromise CRC SP integrity, with high Mg2+ condition demonstrating superior potency in eliciting specific hallmarks of programmed cell death.
Low Mg2+ conditions, paralleling the effects of TRPM6/7 inhibition [13], promoted CRC SP degradation and death through intracellular Mg2+ depletion. Reduced extracellular Mg2+ lowered the driving force for Mg2+ influx through TRPM6, TRPM7, and their heterodimers, as directly evidenced by our Mg2+ influx assays. This process led to a loss of both free and total cellular Mg2+. Coupled with decreased Mrs2 expression, this deficit resulted in mitochondrial Mg2+ depletion [37]. The ensuing loss of Mg2+-mediated the MCU inhibition, which activated the MCU [33,37], triggering Bcl-2 downregulation, ΔΨm collapse, and the induction of apoptotic and autophagic pathways [34,37], ultimately leading to the extensive cell death observed (Figure 9). Translating this mechanism into therapy remains challenging; inducing systemic hypomagnesemia is clinically unfeasible due to severe neuromuscular and cardiac risks. A more practical approach would involve localized Mg2+ channel blockade within the TME. This could be achieved by delivering TRPM6/7 inhibitors with complementary pro-apoptotic agents through controlled-release systems, such as microsphere-encapsulated formulations targeted to the colon or rectum. Such an approach would exploit the Mg2+-dependence of CRC SPs while avoiding the systemic liabilities associated with whole-body Mg2+ depletion.
High Mg2+ condition produces a more potent toxic effect on CRC SPs than low Mg2+, as shown by sharper reductions in SP area, structural integrity, and viability, along with stronger induction of p53, cleaved caspase-3, p62, and ULK1, and a more pronounced suppression of Mrs2 expression and ΔΨm (Figure 8). We propose a distinct mechanism underlying high Mg2+ toxicity. Although total cellular Mg2+ increases, the free intracellular Mg2+ pool in SPs is paradoxically reduced (Figure 5c), likely reflecting Mg2+ buffering and sequestration. The permeability of membrane TRPM6/7 channels is significantly inhibited when intracellular Mg2+ and Mg·ATP levels increase [24]. The elevated total Mg2+ under high Mg2 may therefore directly suppress TRPM6/7 activity, accounting for the markedly impaired Mg2+ influx (Figure 5b). This process initiates a detrimental cycle: (1) inhibited influx prevents restoration of the free cytosolic Mg2+ pool; and (2) reduced Mrs2 expression (Figure 8) restricts mitochondrial Mg2+ import. As a result, despite elevated total Mg2+, the SP experiences a functional deficit in the bioavailable Mg2+ needed for metabolic support and anti-apoptotic signaling. This “functional Mg2+ depletion” at key regulatory sites (cytosol, mitochondria) disturbs bioenergetics (low ΔΨm) and activates cell-death pathways, driving the extensive SP disintegration observed.

3.1. Interacting Pathways and Broader Biological Context
While our study focused on the TRPM6/7-Mg2+ axis, it is important to acknowledge that Mg2+ homeostasis interacts with multiple signaling pathways that could contribute to the observed effects. Reactive oxygen species (ROS) generation is known to be modulated by Mg2+ status [37], and oxidative stress could exacerbate channel dysfunction and mitochondrial damage. Additionally, calcium signaling through the mitochondrial calcium uniporter (MCU) is directly inhibited by Mg2+ [33], creating cross-talk between these two essential divalent cations. The mTOR pathway, which regulates autophagy and cell growth, is also sensitive to Mg2+ availability [38]. Future studies should explore these interactions to provide a more comprehensive understanding of Mg2+’s role in CRC progression.
Our findings in CRC SPs contrast with some reports in other cell types, highlighting the context-dependent nature of Mg2+ biology. The intestinal-specific expression of TRPM6 and functional TRPM6/7 heterodimers [15] provides a unique molecular basis for the distinct Mg2+ handling we observed in HT-29 SPs. For instance, in bone marrow mesenchymal stem cell (BM-MSC) SPs, which likely rely on a different repertoire of Mg2+ channels, elevated Mg2+ was shown to promote proliferation and upregulate MagT1 expression [39]. Conversely, in our HT-29 SPs, elevated Mg2+ downregulated MagT1 and inhibited growth. This fundamental discrepancy may stem not only from differences between normal stem cell and transformed cancer cell metabolism but also from the cell-type-specific Mg2+ influx machinery, where the intestinal TRPM6/7 system in CRC is co-opted to support distinct, potentially more rigid oncogenic programs that are uniquely vulnerable to dysregulation.
Similarly, the relationship between Mg2+ and oxidative stress appears to be finely tuned by cellular context. While Mg2+ deficiency consistently increases ROS production across models, as demonstrated in chick embryo hepatocytes [40], the effect of high Mg2+ is variable. In human keratinocytes, Mg2+ supplementation up to 5 mM was shown to attenuate H2O2-induced mitochondrial damage and cell death, likely through a protective release of Mg2+ from ATP [41]. In stark contrast, our model suggests that similarly elevated Mg2+ induces a state of functional Mg2+ deficiency and metabolic collapse in CRC SPs, associated with increased protein oxidation. This opposing outcome underscores that the cellular response to Mg2+ is not linear and is critically shaped by the underlying metabolic landscape and, importantly, the specific molecular pathways (such as TRPM6/7) that govern cellular Mg2+ homeostasis.
Unlike systemic hypomagnesemia, inducing a localized high Mg2+ TME through controlled-release systems is a feasible and promising therapeutic strategy. Mg-based biomaterials or microsphere-encapsulated hydrogels can deliver sustained, spatially confined Mg2+ release [17,18]. Such an approach could exploit the SP’s susceptibility to Mg2+ overload, triggering the toxic cycle described above, while leveraging the immunomodulatory properties of Mg2+ to enhance treatments such as immune checkpoint blockade [17]. The biodegradability and biocompatibility of Mg-based implants make them especially suitable for this purpose [18], highlighting the therapeutic potential of precisely modulating tumoral Mg2+ fluxes.

3.2. Limitations
This study has several limitations that should be acknowledged. First, we used a single colorectal cancer cell line (HT-29), which may not fully represent the heterogeneity of CRC. Future studies should include additional CRC cell lines and patient-derived organoids. Second, our findings are based entirely on in vitro 3D SP models, which, while more physiologically relevant than 2D cultures, still lack the complexity of in vivo TME. The absence of immune cells, stromal components, and vascularization limits direct translation to clinical settings. Third, we did not measure ROS directly, although our oxidoproteomic data suggest increased oxidative stress. Fourth, the study design does not establish causality but rather demonstrates associations between Mg2+ dysregulation and cellular outcomes. Finally, we focused primarily on the TRPM6/7-Mrs2-Bcl-2 axis without fully exploring other potential mediators of Mg2+-induced cell death. These limitations highlight the need for future in vivo validation and more comprehensive mechanistic studies.

4. Materials and Methods

4.1. Cell Culture
HT-29 cells (ATCC HTB-38) were cultured in Dulbecco’s modified Eagle medium (DMEM; Gibco, Waltham, MA, USA) supplemented with 15% fetal bovine serum, 1% L-glutamine, 1% nonessential amino acids, and 1% penicillin-streptomycin (all from Gibco). All experiments used early-passage cells (passages 5–15) that were not exposed to prior treatments, ensuring consistent metabolic status. Cells were maintained under standard conditions at 37 °C in a humidified atmosphere containing 5% CO2. Subculturing was performed in 75 cm2 tissue culture flasks (Corning Inc., Corning, NY, USA) following the ATCC guidelines.

4.2. SP Formation and Magnesium Modulation
SP formation was triggered by culturing HT-29 cells under nonadherent, serum-free conditions in a 96-well hanging-drop plate (Perfecta3D®, Sigma-Aldrich, Saint Louis, MO, USA). A total of 5 × 104 cells/well were seeded in DMEM/F-12 medium containing 1 × B27 supplement, 20 ng/mL EGF, and 20 ng/mL bFGF (all from PeproTech, Cranbury, NJ, USA). The reservoir was filled with 5 mL PBS to prevent evaporation. Cultures were maintained for up to 14 days at 37 °C and 5% CO2, during which cells aggregated into compact, uniform SPs. Morphological analysis was performed using AnaSP version 3.0 as previously described [13].
To generate controlled Mg2+ environments, SPs were washed with Ca2+/Mg2+-free Dulbecco’s Phosphate-Buffered Saline (DPBS; HiMedia Laboratories, Mumbai, India) and subsequently transferred to a custom HyClone™ DMEM base lacking Ca2+, Mg2+, L-glutamine, and sodium pyruvate (Cytiva, Marlborough, MA, USA). This base was supplemented with 15% FBS, 1% L-glutamine, 1% non-essential amino acids, and 1.25 mM CaCl2. Sterile MgCl2 was added to achieve the following final concentrations:Control (physiological): 1.0 mM Mg2+ (based on normal serum Mg2+ levels [15])

Low Mg2+ conditions: 0.2 mM (severe), 0.4 mM (moderate), 0.6 mM (mild) Mg2+ (representing clinically relevant deficiency levels [15])

High Mg2+ conditions: 1.5 mM (mild, mimicking moderate clinical hypermagnesemia), 2.5 mM (moderate, mimicking symptomatic clinical hypermagnesemia), 5.0 mM (severe, mimicking severe clinical hypermagnesemia) Mg2+ (supraphysiological levels used in experimental models, based on clinical thresholds [15])

All media were verified to fall within physiological pH and osmolarity ranges.

4.3. Cell Viability Assay (WST-1)
SP viability was quantified using the WST-1 cell proliferation reagent (Sigma-Aldrich, Cat. No. 05015944001), which reports mitochondrial dehydrogenase activity. After treatment, SPs were exposed to WST-1 for 4 h at 37 °C to allow enzymatic conversion of the tetrazolium substrate to formazan. Absorbance was measured using an EnSight™ multimode plate reader (PerkinElmer, Waltham, MA, USA) at 450 nm with 620 nm as reference. Each experiment included blank controls (medium without cells) and negative controls (untreated cells). Data were normalized to control conditions (set as 100% viability).

4.4. Mg2+ Influx Measurements
For Mg2+ influx analysis, subconfluent cells grown in μ-Dishes (Ibidi, Gräfelfing, Germany) were preincubated in DMEM for 3 days to equilibrate baseline transport activity. On the assay day, cells were rinsed with Ca2+/Mg2+-free DPBS (HiMedia Laboratories, Thane, India) and incubated for 1 h in a custom HyClone™ DMEM base devoid of Ca2+, Mg2+, L-glutamine, and sodium pyruvate (Cytiva, Marlborough, MA, USA), supplemented with 15% FBS, 1% L-glutamine, and 1% non-essential amino acids. Cells were then loaded with 5 μmol/L Mag-Fura-2-AM (Invitrogen, Waltham, MA, USA) in N-methyl-D-glucamine buffer for 1 h, followed by a 30 min de-esterification in dye-free buffer. Real-time fluorescence imaging was performed on an Olympus FV3000 confocal microscope (Olympus Corporation, Tokyo, Japan). Mg2+ influx was calculated as ΔF/F = [F(t) − F(0)]/[F(0) − F(b)], where F(t) is the mean fluorescence intensity at a given time, F(0) is the basal fluorescence intensity, and F(b) is the background fluorescence. Image analysis was conducted using ImageJ (https://imagej.net/ij/download.html accessed on 11 December 2025) [13].

4.5. Analysis of Mitochondrial Membrane Potential (ΔΨm)
ΔΨm was evaluated in parental and SP-derived cells using the JC-1 Mitochondrial Membrane Potential Assay Kit (Abcam, Cambridge, UK, cat. no. ab113850). The assay was conducted according to the manufacturer’s instructions, with modifications based on the work of Anaya-Eugenio et al. [42]. Briefly, cells were seeded in 96-well plates and allowed to adhere before incubation with JC-1 working solution for 20 min at 37 °C in the dark. After staining, cells were gently rinsed with warm PBS to remove excess dye. Fluorescence was recorded using an EnSight™ multimode plate reader (PerkinElmer). JC-1 exhibits potential-dependent mitochondrial accumulation, as reflected by an emission shift from green (~530 nm; monomers, low ΔΨm) to red (~590 nm; aggregates, high ΔΨm). The red-to-green fluorescence ratio was calculated to quantify mitochondrial polarization. Control wells with FCCP (carbonyl cyanide-p-trifluoromethoxyphenylhydrazone) were included to confirm depolarization.

4.6. Measurement of Intracellular Mg2+ Levels
Free Mg2+: Subconfluent cells in 6-well plates were preincubated for 3 days, then loaded with 5 μmol/L Mag-Fura-2-AM in standard bathing solution [13] for 30 min at 37 °C, followed by a 30 min de-esterification period. Fluorescence was acquired using an Olympus FV3000 confocal microscope to quantify cytosolic free Mg2+.
Total Mg2+: Total cellular Mg2+ content was measured by atomic absorption spectrophotometry (Shimadzu, Kyoto, Japan). Cells were lysed by sonication, and Mg2+ concentrations were determined relative to a standard curve, as per the standard protocols [13].
For both measurements, blank samples (no cells) and standard solutions were included for calibration.

4.7. Immunoprecipitation
TRPM6 immunoprecipitation (IP) was carried out according to established procedures [13,27]. Cells were lysed in cold RIPA buffer containing protease and phosphatase inhibitors, and membrane fractions were isolated with the Mem-PER™ Plus kit (Thermo Fisher Scientific, Waltham, MA, USA). Lysates were incubated overnight at 4 °C with an anti-TRPM6 polyclonal antibody (Thermo Fisher Scientific). Afterward, immune complexes were captured using Protein A/G Sepharose® beads (Abcam). Following rigorous washing, bound proteins were eluted with glycine–Tris buffer (Sigma-Aldrich) and concentrated using Vivaspin® 20 centrifugal filters (Sartorius AG, Göttingen, Germany) for subsequent analysis. Control IPs with normal rabbit IgG were performed to assess non-specific binding.

4.8. Western Blot Analysis
Western blotting was performed according to standard protocols [13,27]. Protein samples were resolved by SDS–PAGE and electrotransferred onto nitrocellulose membranes. Precision Plus Protein™ Dual Xtra Prestained Standards (Cat. No. 1610377; Bio-Rad Laboratories, Hercules, CA, USA) served as molecular weight markers. Membranes were probed with specific primary antibodies against p53 (PB9008; Bosterbio, Pleasanton, CA, USA), BCL-2 (A00040; Bosterbio), Bax (A00183; Bosterbio), caspase-3 (PB9188; Bosterbio), cleaved caspase-3 (ab32042; Abcam), pS556 ULK1 (ab203207; Abcam), p62/SQSTM1 (ab109012; Abcam), TRPM6 (0ST00108W; Thermo Fisher Scientific Inc.), TRPM7 (ab729; Abcam), MagT1 (ab244490; Abcam), CNNM4 (ab191207; Abcam), Mrs2 (ab246915; Abcam), β-actin (ab6276; Abcam), and Na+/K+-ATpase (ab76020; Abcam). After incubation with HRP-conjugated secondary antibodies, protein bands were detected using SuperSignal® West Pico chemiluminescent substrate (Thermo Fisher Scientific) and imaged with a ChemiDoc™ Touch system (Bio-Rad Laboratories). Densitometric quantification was performed with ImageJ software. Protein expression levels were normalized to the parental control group. β-actin served as the loading control for total cell lysates and cytosolic fractions, while Na+/K+-ATPase was used for membrane fractions.

4.9. Immunofluorescence
For TRPM6 and TRPM7 localization, SPs were fixed in 4% paraformaldehyde for 10 min at room temperature and then permeabilized in pre-chilled methanol (−20 °C) for 10 min. Blocking was performed sequentially with 0.1% glycine for 10 min, followed by a 30 min incubation in a protein-free blocking buffer (Visual Protein, Neihu Dist., Taipei, Taiwan). Primary antibodies were diluted 1:100 in blocking buffer and applied overnight at 4 °C. After washing, SPs were treated with fluorophore-conjugated secondary antibodies (1:500; FITC, ab6717; or Alexa Fluor® 568, ab175474; Abcam) for 1 h at room temperature. Nuclei were counterstained with Hoechst (Molecular Probes, Eugene, OR, USA) according to the manufacturer’s guidelines. Fluorescent images were acquired using the ZEISS LSM910 Lightfield 4D microscope (Carl Zeiss Microscopy GmbH, Jena, Germany). Negative controls (no primary antibody) were included to assess non-specific staining.

4.10. Nano-Liquid Chromatography Tandem Mass Spectrometry Analysis (nanoLC-MS/MS)
Proteomic profiling of membranous TRPM6 and TRPM7 was performed by nanoLC-MS/MS as previously described [27]. To minimize artifactual methionine oxidation (MetO), all steps were conducted rapidly using fresh reagents. Proteins were purified (Clean-up kit, GE Healthcare, Chicago, IL, USA), solubilized in 8 M urea, and quantified via Bradford assay (Bio-Rad Laboratories, Hercules, CA, USA). For in-solution digestion, 20 µg of protein was reduced (100 mM DTT in 100 mM TEAB, 30 min, RT), alkylated (100 mM iodoacetamide, 30 min, dark), and digested with trypsin (Trypsin Gold, Promega, Madison, WI, USA; 1:50 w/w, 16 h, 37 °C). Peptides were desalted (C18 Zip-tips, MilliporeSigma), dried, and stored at −80 °C.
Prior to LC-MS/MS, peptides were reconstituted in 0.1% formic acid. Separations were performed on a Dionex Ultimate 3000 RSLCnano system coupled to a Q-ToF Compact II mass spectrometer (Bruker Daltonics, Billerica, MA, USA). Peptides (1 µg) were loaded onto a PepMap100 C18 column (75 µm × 500 mm, 3 µm) and eluted over a 90 min linear gradient (2–95% solvent B: 80% acetonitrile, 0.08% FA) at 300 nL/min. Data were acquired in positive-ion, data-dependent acquisition mode (m/z 150–2200) with collision-induced dissociation (CID).
Raw files were processed using DataAnalysis software (Bruker Daltonics, version 5.3) and searched against the UniProtKB human reference proteome database (release 2023_01) using the MASCOT search engine (v2.3; Matrix Science, London, UK). Search parameters specified: trypsin/P digestion with up to two missed cleavages; precursor and fragment mass tolerances of 20 ppm and 0.05 Da, respectively; fixed modification of carbamidomethylation (C); variable modifications of MetO, N-terminal acetylation, and phosphorylation (Ser, Thr, Tyr). Protein and peptide identifications were validated at a false discovery rate (FDR) of <1%. The exponentially modified protein abundance index (emPAI) was used for semi-quantitative protein abundance comparison.
To specifically address changes in post-translational modifications (PTMs), the relative abundance of MetO for key residues and phosphorylation for key regulatory sites was determined as previously described [43]. The extracted ion chromatogram (XIC) peak area for the modified peptide form was compared to the peak area of its non-modified counterpart identified in the same LC-MS/MS run. The peptide spectrum match (PSM) ratios were calculated for each sample, providing a relative measure of phosphorylation and oxidation status across experimental groups. All samples were analyzed in triplicate (technical replicates, n = 3).

4.11. Limitations
Our mass spectrometry analysis, while providing novel insights into TRPM6/7 modifications, has inherent constraints. Phosphopeptide enrichment was not employed; therefore, our identification of phosphorylation sites, though validated, may not be comprehensive for low-abundance phosphopeptides. The quantification of post-translational modifications was focused on MetO via XIC-based ratios; phosphorylation sites were identified but not similarly quantified due to the lack of enrichment and lower spectral counts. Although stringent precautions were taken, the reported MetO levels represent relative changes between groups and could include minor artifactual oxidation.

4.12. Statistical Analysis
Data are expressed as mean ± standard deviation (SD). Data normality was confirmed using the Shapiro–Wilk test. Differences between two groups were analyzed using an unpaired Student’s t-test (used for comparisons between parental and SP groups under the same Mg2+ condition), whereas comparisons across multiple groups employed one-way analysis of variance with Dunnett’s post hoc test for multiple comparisons versus control groups. A p-value < 0.05 was considered statistically significant. All statistical analyses were performed using GraphPad Prism 8.0.1 (GraphPad Software Inc., San Diego, CA, USA).

5. Conclusions
Our in vitro study demonstrates that CRC spheroids, while displaying a Mg2+-acquisitive phenotype that supports growth under physiological conditions, harbor a critical vulnerability: their dependence on tightly regulated Mg2+ homeostasis makes them acutely sensitive to both depletion and overload of this essential ion. We further clarify the mechanistic basis of this sensitivity, showing that departures from physiological Mg2+ disrupt the core regulatory apparatus—chiefly the TRPM6/7 heterodimer—through post-translational modifications that diminish membrane localization and channel conductance. This results in a progressive collapse of Mg2+ handling, mitochondrial impairment, and the activation of coordinated apoptotic and autophagic cell-death programs. Notably, high Mg2+ stress proved especially potent, inducing a form of “functional Mg2+ depletion” that drives profound metabolic failure. These findings shift the paradigm from viewing Mg2+ merely as a pro-tumorigenic factor to recognizing its dysregulation as a compelling therapeutic target. However, the in vitro nature of this study and use of a single cell line necessitate cautious interpretation and future in vivo validation. They offer strong justification for developing localized strategies, such as controlled-release Mg2+ formulations or selective channel inhibitors, that intentionally disrupt intratumoral Mg2+ homeostasis to hinder CRC progression while limiting systemic toxicity. Our results suggest that both deficiency and excess of Mg2+ can be exploited therapeutically, with high Mg2+ showing particular promise for localized intervention. The “dividing line” between beneficial and harmful Mg2+ levels appears to depend on cellular adaptive capacity rather than a fixed concentration, highlighting the importance of context-specific modulation.