Crosstalk between Hedgehog pathway and the glucocorticoid receptor pathway as a basis for combination therapy in T-cell acute lymphoblastic leukemia
Abstract
Despite considerable advancements in therapeutic regimens and intensified treatment protocols, a significant proportion of patients afflicted with T-cell acute lymphoblastic leukemia (T-ALL) continue to face an alarmingly poor prognosis. This challenging clinical reality is primarily attributable to two critical factors: the occurrence of primary resistance to initial therapeutic interventions, meaning the leukemia cells do not respond adequately to standard treatments from the outset, and the disheartening phenomenon of disease relapse, where the cancer returns after a period of remission. These persistent challenges underscore an urgent and undeniable need for the identification and development of more efficient, more targeted, and ultimately, more effective therapeutic strategies to improve patient outcomes in T-ALL.
The Hedgehog (HH) signaling pathway is a fundamental and evolutionarily conserved cascade that plays an indispensable role during embryonic development, orchestrating crucial processes such as cell proliferation, differentiation, and tissue patterning. However, this powerful developmental pathway is frequently found to be aberrantly activated or deregulated in various forms of human cancer, contributing significantly to tumorigenesis, cancer progression, and therapeutic resistance. Mounting evidence from recent research has increasingly pointed towards an emerging and critical role for dysregulated HH signaling specifically within the context of T-ALL. What is particularly noteworthy and therapeutically challenging is the growing consensus that ligand-independent activation of the HH pathway commonly occurs in a range of cancers, including T-ALL. This means that the pathway can be activated without the presence of its usual extracellular signaling molecules, making it more difficult to target using conventional approaches that block ligand-receptor interactions. This intrinsic activation mechanism underscores the profound necessity of meticulously dissecting the intricate and complex interplay between the HH pathway and other pivotal signaling cascades that are involved in regulating its activation and overall cellular function.
In this comprehensive study, we delve into precisely such a complex interaction, revealing a novel and therapeutically relevant crosstalk between the Hedgehog signaling pathway and the glucocorticoid receptor (NR3C1) pathway. This critical interplay exerts its influence directly at the level of the GLI1 transcription factor, which serves as the primary downstream effector and key transcriptional activator of the Hedgehog pathway. Our investigations demonstrated that the combination of GANT61, a well-established small molecule inhibitor specifically targeting GLI activity, and dexamethasone, a synthetic glucocorticoid and widely used chemotherapeutic agent, exhibited a powerful synergistic anti-leukemic effect. This remarkable synergy was observed not only in controlled in vitro experiments using established T-ALL cell lines, where a combined reduction in leukemia cell viability was evident, but critically, it was also recapitulated in vivo within patient-derived xenografts. This latter finding highlights the direct translational potential of this combinatorial approach, as these xenograft models closely mirror the biological complexity and heterogeneity of human T-ALL.
Delving into the precise molecular mechanisms underlying this observed synergy, we elucidated how dexamethasone, by activating the nuclear glucocorticoid receptor (NR3C1), profoundly impairs the function of the GLI1 transcription factor. Mechanistically, activated NR3C1 dynamically modulates the recruitment of key epigenetic regulators to GLI1. Specifically, it influences the association of PCAF, an acetyltransferase that typically promotes gene expression by adding acetyl groups to proteins, and HDAC1, a deacetylase that removes acetyl groups, generally leading to gene repression. This dynamic modulation of recruitment orchestrated by NR3C1 leads to a significant increase in the acetylation state of GLI1. This heightened GLI1 acetylation, in turn, was found to be directly associated with two crucial consequences: a compromised transcriptional activity of GLI1, meaning its ability to activate its target genes is significantly reduced, and a concomitant reduction in GLI1 protein stability, leading to its more rapid degradation within the cell. The net result is a potent suppression of oncogenic HH signaling.
In summary, our study successfully identifies and characterizes a previously unrecognized and novel crosstalk between the GLI1 and NR3C1 signaling pathways. This groundbreaking discovery not only expands our fundamental understanding of signal transduction networks in cancer but also presents a compelling therapeutic opportunity. The synergistic anti-leukemic effects observed, coupled with the detailed mechanistic insights, strongly suggest that this identified crosstalk could be strategically exploited in the development of innovative combination therapies for Hedgehog-dependent malignancies, including T-ALL. By simultaneously targeting HH signaling with GLI inhibitors and leveraging the distinct modulatory effects of glucocorticoids via NR3C1, it may be possible to significantly increase the overall therapeutic efficacy and overcome treatment resistance in these challenging cancers.
Introduction
T-cell acute lymphoblastic leukemia (T-ALL) is a highly aggressive hematological malignancy characterized by the uncontrolled expansion of immature lymphocyte precursors that are aberrantly committed towards T-cell development within the bone marrow, thymus, and peripheral blood. Despite notable improvements in the overall outcome for T-ALL patients in recent years, achieved through intensified chemotherapy regimens and improved supportive care, a significant and concerning fraction of individuals still face a dismal prognosis. This unfavorable outlook is largely driven by two critical clinical challenges: primary resistance to initial treatment, where the leukemia cells fail to respond adequately to standard therapeutic protocols from the outset, and the disheartening phenomenon of disease relapse, which occurs in a substantial number of patients after a period of remission. These resistant and relapsed cases collectively account for a considerable proportion of the burden, affecting approximately 40% of adult T-ALL patients and around 20% of pediatric cases. Such statistics underscore a pressing and undeniable need for the discovery and development of novel, more effective, and precisely targeted therapeutic strategies to improve long-term survival and reduce morbidity in this challenging patient population.
The Hedgehog (HH) signaling pathway is an evolutionarily conserved and fundamental biochemical cascade that acts as a master regulator of various biological processes. During embryogenesis, it plays an indispensable role in orchestrating stem cell biology, guiding crucial events such as cell proliferation, differentiation, and the precise patterning of developing tissues. This developmental importance extends specifically to early T-cell development, where HH signaling is intricately involved in the proper formation and maturation of T lymphocytes. Beyond embryonic development, the HH pathway also maintains vital roles in adult life, contributing to tissue homeostasis and renewal, ensuring the proper maintenance and repair of various organs. Mammalian cells express three distinct HH ligands: Sonic Hedgehog (SHH), Desert Hedgehog (DHH), and Indian Hedgehog (IHH). These ligands initiate signaling by binding to the transmembrane receptor Patched-1 (PTCH1). In the absence of these ligands, PTCH1 constitutively exerts a repressive effect on Smoothened (SMO), another transmembrane protein, preventing its activation. However, upon the binding of an HH ligand to PTCH1, the repressive activity of PTCH1 on SMO is released, leading to the de-repression and subsequent activation of SMO. This activation of SMO then triggers a complex intracellular cascade of events that culminates in the activation of the GLI family of zinc finger transcription factors, which include GLI1, GLI2, and GLI3. Once activated, these GLI proteins translocate into the nucleus, where they exert their primary function by regulating the expression of a diverse array of target genes, which can be either pathway-specific or cell-specific. While GLI2 and GLI3 primarily function as transcriptional activators and repressors, respectively, GLI1 operates exclusively as a transcriptional activator, serving to amplify existing HH signaling in a powerful positive feedback loop, thereby enhancing the overall pathway activity.
Beyond its physiological roles, aberrant activation of the HH signaling pathway has been extensively documented in various solid and hematological tumors. This aberrant initiation can occur through both ligand-dependent mechanisms, involving the overexpression of HH ligands, and, crucially, through ligand-independent mechanisms. These ligand-independent mechanisms often involve oncogenic mutations within pathway components or what are termed “non-canonical” GLI activation events, where GLI proteins are activated independently of SMO. Such aberrant HH signaling contributes significantly to tumor development, progression, and the maintenance of cancer stem cells, driving uncontrolled proliferation and therapeutic resistance. In the context of non-canonical activation, GLI1, the terminal effector, can be activated in a SMO-independent manner through intricate crosstalk with other oncogenic signaling pathways. Prominent examples include the PI3K/Akt, TGFβ/SMAD, and mTOR/S6K1 pathways, which can modulate GLI1 expression levels or induce critical post-translational modifications (PTMs) that directly impact GLI1 function, thereby bypassing upstream pathway components. A definitive role for HH signaling in T-ALL has only recently been established, with rare mutations within the pathway components reported in some T-ALL patients and active HH signaling observed in approximately 20% of T-ALL cases. This subset of patients might potentially derive therapeutic benefit from treatment with HH inhibitors. Preclinical studies have provided compelling evidence suggesting that the inhibition of the HH pathway at the level of its downstream effector, GLI1, is considerably more cytotoxic to T-ALL cells than inhibition at the level of the upstream SMO receptor. This observation further underscores the critical importance of non-canonical activation of the HH pathway in T-ALL. However, despite these promising initial findings, the specific characteristics of T-ALL cases that are genuinely sensitive to HH inhibitors, and consequently, the precise therapeutic relevance of targeting this pathway in a clinical setting, remain largely to be elucidated and require further in-depth investigation.
Herein, our research endeavors were specifically aimed at identifying and comprehensively characterizing the molecular pathways that exhibit functional crosstalk with Hedgehog signaling in T-ALL, with a particular focus on interactions that could hold therapeutic relevance. In this report, we describe a clinically significant and previously unappreciated crosstalk between the HH pathway and the glucocorticoid (GC) receptor pathway. This newfound interaction culminates in a profound and robust cytotoxic effect when glucocorticoids are combined with the GLI inhibitor GANT61, offering a promising avenue for novel combination therapies.
Results
HH Pathway is Active in T-ALL Cells and Crosstalks with NOTCH1 and GC Signaling Pathways
Given the accumulating evidence suggesting the importance of the Hedgehog (HH) pathway in T-ALL, and the specific proposition that GLI1, the terminal transcription factor of HH signaling, represents a more preferential therapeutic target in T-ALL compared to the upstream receptor Smoothened (SMO), our investigation commenced with a thorough characterization of HH pathway activity in a diverse cohort of T-ALL samples. This cohort included ten distinct T-ALL cell lines, eleven patient-derived xenografts (PDX), which offer a more physiologically relevant model of human leukemia, and two well-defined NOTCH1-induced murine T-ALL models, namely the HDΔPEST and ΔE models (three samples each), known for their homogeneity and well-characterized oncogenic alterations. We systematically assessed the expression levels of key HH pathway components: GLI1, SMO, and PTCH1.
Quantitative analysis of GLI1 transcript levels, which serve as a general and reliable readout of active HH signaling, consistently revealed significant upregulation in the vast majority of the tested T-ALL samples when compared to normal thymocytes. This finding unequivocally confirmed a frequent and widespread activation of the HH pathway within T-ALL. Furthermore, Western blot analysis corroborated these transcriptional data, demonstrating elevated GLI1 protein expression in numerous T-ALL cell lines and PDX samples. Interestingly, GLI2 protein, another member of the GLI family, was also frequently detected, particularly in T-ALL samples that exhibited lower GLI1 expression, suggesting potential compensatory or distinct roles for different GLI family members in these contexts.
Previous studies had indicated that while HH inhibitors could partially impede the growth of T-ALL cells, the GLI inhibitor GANT61 consistently displayed substantially greater cytotoxicity compared to SMO inhibitors. This observation further propelled our focus on GLI1-mediated non-canonical HH activation. Consequently, we sought to systematically identify other signaling pathways that might modulate the response to GANT61, with the strategic aim of uncovering potential combinatorial treatment approaches that hold therapeutic value for future patient applications. For this initial comprehensive screening, we utilized HDΔPEST murine T-ALL cells. These cells were particularly suitable for the screening due to their strong NOTCH1 dependency and limited number of other oncogenic alterations, which collectively result in the expansion of a highly homogeneous and well-characterized tumor cell population, minimizing confounding factors. HDΔPEST leukemia cells were treated ex vivo with GANT61 in combination with a broad panel of inhibitors, each targeting diverse oncogenic signaling pathways that are frequently found to be deregulated in T-ALL.
The results of this combinatorial screening were highly revealing. The highest degree of synergism, quantitatively inferred from combination index (CI) values—where a CI value less than 1 indicates a synergistic interaction—was strikingly observed between GANT61 and the gamma-secretase inhibitor dibenzazepine (DBZ), which targets the NOTCH signaling pathway. Equally significant was the profound synergism detected between GANT61 and the synthetic steroid dexamethasone (dexa), a potent activator of the glucocorticoid receptor (GC) pathway. These findings strongly suggested a previously unappreciated functional crosstalk between the HH pathway and both the NOTCH signaling pathway and the GC receptor pathway. While a direct link between NOTCH1 signaling and the HH pathway has already been described in various contexts, the potential connection with the GC receptor pathway was particularly intriguing and warranted further in-depth investigation. Glucocorticoids, such as dexamethasone, represent a cornerstone of treatment for lymphoid tumors, including T-ALL. Moreover, a poor response to prednisone, a precursor to dexamethasone, is recognized as an early and significant marker of an unfavorable prognosis in T-ALL patients. Therefore, given the critical role of GCs in T-ALL therapy, we prioritized further exploration of the potential connection between the HH pathway and the GC receptor pathway due to its direct clinical relevance.
Dexamethasone Synergizes with the GLI Inhibitor GANT61 to Induce a Strong Cytotoxic Effect in Human T-ALL Cell Lines and PDX Samples
Following the promising results from our initial screening, we then rigorously addressed the therapeutic significance of jointly targeting the Hedgehog (HH) pathway and the glucocorticoid (GC) receptor pathway in human T-ALL. We conducted extensive in vitro experiments using a panel of diverse human T-ALL cell lines, subjecting them to a combination treatment of dexamethasone (dexa) and the GLI inhibitor GANT61 for 48 hours. The results consistently demonstrated a significantly enhanced cytotoxicity with the combined treatment compared to either drug administered alone, irrespective of the cell lines’ intrinsic sensitivity to GCs. This profound synergistic effect was quantitatively confirmed by CI values consistently below 1 for most drug concentrations across all tested T-ALL cell lines.
To gain deeper molecular insights, we investigated the expression levels of HH target genes (specifically *GLI1* and *PTCH1*) and GC-responsive genes (*NR3C1*, *BIM*, and *GILZ*) in DND41 T-ALL cells, a cell line known to be sensitive to GCs. Our analysis revealed that the combination treatment (GANT61 + dexa) was highly effective in repressing the HH target gene *PTCH1* and notably increased GC-sensitivity, as evidenced by enhanced expression of *NR3C1* and *GILZ*, despite an unexpected increase in *GLI1* transcript levels. This increased GC-sensitivity, indicated by *GILZ* expression, was also observed in additional T-ALL cell lines treated with the combination, reinforcing the robustness of this interaction.
Further investigation into the cellular mechanisms underlying this pronounced cytotoxic effect indicated that it was predominantly due to a significant increase in apoptosis, rather than merely cell-cycle arrest. Similar compelling results were obtained using patient-derived xenograft (PDX) cells, which offer a more clinically relevant model. The combination treatment (GANT61 + dexa) proved to be highly cytotoxic in these PDX-derived cells, notably affecting both GC-sensitive and GC-resistant samples. Western blot analysis of key proteins implicated in apoptotic cell death, such as cleaved PARP-1 and XIAP, further confirmed the superior efficacy of the combination treatment compared to single agents. Overall, these comprehensive data collectively highlight the substantial therapeutic value of combining GANT61 and dexamethasone for the treatment of T-ALL cells, providing strong preclinical rationale for this novel dual targeting strategy.
Dexamethasone Acts by Inhibiting GLI1 Transcriptional Activity
To precisely unravel the mechanistic underpinnings of how glucocorticoids (GCs) influence the Hedgehog (HH) pathway and specifically impact GLI1 function, and recognizing the inherent challenges in manipulating T-ALL cells for such detailed mechanistic studies, we transitioned our investigation to HEK-293T cells. We utilized a stable HEK-293T cell line engineered to express the glucocorticoid receptor NR3C1 (designated 293T HA-NR3C1), which had previously demonstrated an enhanced cytotoxic response to the GANT61 + dexa combination, closely mimicking the effects observed in T-ALL cells.
Our initial experiments focused on assessing the effect of dexamethasone (dexa) on GLI1-driven luciferase activity in these 293T HA-NR3C1 cells. We observed a clear and dose-dependent inhibition of GLI reporter activation by dexa, indicating that the glucocorticoid could effectively suppress GLI1 transcriptional activity. Crucially, this inhibitory effect was entirely dependent on the presence of the GC receptor, as no such effect was observed in control HEK-293T cells that did not express NR3C1 (293T pMSCV-puro). Surprisingly, despite this clear reduction in GLI1 activity, Western blot analysis revealed that it was not associated with a decrease in GLI1 protein expression; instead, an upregulation of GLI1 protein was consistently observed in dexa-treated cells. Further characterization of the 293T HA-NR3C1 cells confirmed the functional integrity of the GC receptor: dexa treatment for 24 hours induced the expression of several classical GC-responsive genes, including *NR3C1*, *BIM*, and *GILZ*, and also led to a significant G1 cell-cycle arrest. Interestingly, NR3C1 protein expression itself was found to be reduced in response to dexa, a phenomenon previously described in various non-lymphoid cells. Our data suggest that even with this reduction in NR3C1 protein following dexa treatment, the residual NR3C1 levels are sufficient to mediate a robust functional response, including cell-cycle arrest and apoptosis, thus validating this model for further mechanistic exploration. We further confirmed a similar inhibitory action of GCs on GLI1 function in endogenous Jurkat T-ALL cells and in HeLa cells, and also with another synthetic steroid, fluticasone, broadening the applicability of our observations.
To definitively ascertain the specific requirement for NR3C1 in mediating the dexamethasone-induced repression of GLI1, we treated 293T HA-NR3C1 cells with mifepristone (RU486), a well-known GC antagonist, either alone or in combination with dexa. RU486 effectively abrogated the negative modulation of GLI1 by dexa and, significantly, reversed the dexa-induced upregulation of GLI1 expression. Consistent with this, RU486 also reverted the induction of the NR3C1 target gene *GILZ* and the downregulation of NR3C1 protein that typically followed dexa treatment. Gene expression analysis further confirmed the functional impairment of GLI1 by dexa, as a downregulation of HH target genes *PTCH1* and *HHIP* was consistently observed, despite the higher expression levels of *GLI1* itself. Collectively, these data strongly suggest that the regulation of GLI1 activity by dexa occurs through a post-translational mechanism, rather than at the transcriptional level.
Given that GLI transcription factors must translocate and accumulate in the nucleus to exert their activated function, we initially evaluated the subcellular distribution of GLI1 upon dexa treatment. Cell fractionation analysis of 293T HA-NR3C1 cells indicated a predominant nuclear expression of GLI1 protein under basal conditions, consistent with active pathway signaling. However, no significant changes in its subcellular distribution were observed in response to dexa treatment. Nevertheless, GLI1 protein levels in these cells were indeed upregulated in dexa-treated cells compared to controls, in agreement with our previous findings. In contrast, dexa-activated NR3C1 was observed to be downregulated and simultaneously shuttled from the cytoplasm to the nucleus. Extending our analysis to T-ALL cells, we found that GLI1 was almost exclusively nuclear under basal conditions in cell lines such as DND41 and CUTLL1, and dexa treatment similarly did not significantly modify its subcellular distribution in these lines. We observed a different pattern, however, in PDX cells, which displayed a more mixed expression of GLI1 distributed between cytoplasmic and nuclear fractions. In these PDX cells, dexa treatment led to an increased expression of GLI1 protein, which was notably associated with an increased cytoplasmic accumulation. Interestingly, dexa treatment also increased GLI1 transcript levels in some T-ALL cells (HSB2, DND41), but its effects on total GLI1 protein levels were variable, suggesting important differential post-translational alterations across various T-ALL cell lines.
NR3C1 Interacts with GLI1 and Promotes its Acetylation
The delicate balance of GLI1 function and its activation is meticulously controlled by various post-translational modifications (PTMs), including phosphorylation, acetylation, and ubiquitination. These modifications are critical because they not only regulate the protein’s half-life and its precise subcellular trafficking but also directly influence its inherent transactivation potential. Acetylation, in particular, has been demonstrated to inhibit the transcriptional activity of both GLI1 and its homolog GLI2 by preventing their efficient occupancy of target gene promoters. Given that activated NR3C1 is known to interact with histone acetyltransferases (HATs) and histone deacetylases (HDACs) at gene promoter sites, where these proteins act as transcriptional co-activators and co-repressors, respectively, we therefore sought to determine whether dexa treatment could alter the acetylation status of GLI1.
Our experiments revealed that in 293T HA-NR3C1 cells, GLI1 exhibited basal acetylation. Crucially, the degree of GLI1 acetylation was significantly enhanced following treatment with either dexa or Trichostatin A (TSA), a pan-HDAC inhibitor known to increase global protein acetylation, serving as a positive control. To further dissect the functional consequences of acetylation and deacetylation on HH signaling, we tested the effects of known modulators of GLI1 acetylation. This included the HATs p300 and p300/CBP-associated factor (PCAF), and the histone deacetylase HDAC1, on GLI1-driven reporter activity. Indeed, overexpression of both p300 and PCAF demonstrably repressed GLI1-reporter activity, mirroring the inhibitory effect of dexa. Conversely, overexpression of HDAC1 led to an enhancement of GLI1-reporter activity, an effect that was effectively reverted by dexa treatment. Furthermore, when HDAC1 was co-expressed in the presence of p300 or PCAF, it counteracted the repressive effects observed with p300 or PCAF alone. Collectively, these data strongly indicate a pivotal role for GLI1 acetylation in mediating the repressive function of dexa on GLI1 transcriptional activity.
To ascertain whether dexa could specifically alter the recruitment of these critical acetylation modulators to GLI1, we conducted immunoprecipitation (IP) experiments in transfected 293T HA-NR3C1 cells. Consistent with previous studies, we were able to detect both HDAC1 and PCAF in anti-HA immunoprecipitates, indicating their constitutive association with HA-GLI1 or HA-NR3C1. Significantly, after dexa treatment, a reproducible reduction in immunoprecipitated HDAC1 was observed, alongside a notable enrichment in immunoprecipitated PCAF. These findings are highly consistent with our hypothesis that activated NR3C1 promotes GLI1 hyper-acetylation by dynamically influencing the association of these key modifying enzymes. Interestingly, p300, another well-described HAT known to target GLI1, was not found to be recruited after dexa treatment, suggesting that PCAF may be the predominant HAT responsible for acetylating GLI1 following glucocorticoid exposure.
Dexamethasone Promotes GLI1 Degradation
Given that the acetylation of GLI1 homologs in *Drosophila* (Ci) and GLI proteins in mammalian cells has been associated with proteasome-dependent cleavage and degradation, we meticulously investigated whether dexamethasone (dexa) could influence the protein stability of GLI1. To address this, we treated 293T HA-NR3C1 cells with dexa and concurrently inhibited *de novo* protein synthesis using cycloheximide (CHX). Our experiments revealed a remarkable reduction in GLI1 protein stability in 293T HA-NR3C1 cells following dexa treatment, indicating accelerated degradation. Similar compelling results were consistently obtained in two distinct T-ALL cell lines, HSB2 and DND41, further supporting the generalizability of this effect.
Since GLI1 protein stability is intrinsically linked to its rate of ubiquitination and subsequent proteasomal degradation, we further investigated whether GLI1 ubiquitination was altered upon dexa treatment. To accomplish this, 293T HA-NR3C1 cells were co-transfected with constructs expressing GLI1 and ubiquitin, and then treated with dexa. Proteasomal degradation was meticulously inhibited using MG132 to allow for the accumulation of ubiquitinated forms, and GLI1 was subsequently immunoprecipitated to evaluate its degree of ubiquitination. Dexa treatment unequivocally resulted in a dramatic increase in GLI1 ubiquitination, a finding that is entirely coherent with our hypothesis that dexa reduces GLI1 stability by enhancing its proteasomal degradation through increased ubiquitination. In T-ALL cells, this observed reduction in GLI1 protein stability may manifest with faster kinetics in GC-sensitive cells that express high levels of NR3C1 (and which may exhibit auto-upregulation of NR3C1), such as DND41. In contrast, GC-resistant cells expressing low levels of NR3C1 (which may or may not undergo auto-upregulation, like CUTLL1 and HSB2, respectively) might exhibit different kinetics of GLI1 degradation.
In conclusion, our collective findings strongly support a refined model of action. In this model, active NR3C1, stimulated by glucocorticoids such as dexamethasone, initially promotes the accumulation of an acetylated form of GLI1. This occurs through a dynamic and crucial modulation of the recruitment of specific acetyltransferases, notably PCAF, and the concurrent dissociation of deacetylases, such as HDAC1, from GLI1 protein complexes. This enhanced acetylation directly inhibits GLI1’s transcriptional activity, effectively turning off its oncogenic gene expression program. Subsequently, this hyper-acetylated GLI1 undergoes further ubiquitination, marking it for rapid proteasomal degradation, leading to a reduction in its overall protein levels. This dual mechanism—initial inactivation followed by degradation—provides a robust explanation for the observed therapeutic synergy.
Discussion
The inhibition of the Hedgehog (HH) pathway represents an increasingly attractive and highly promising therapeutic option, not merely confined to a limited spectrum of tumors definitively categorized as HH-dependent, but potentially extending its utility to a broader range of hematological malignancies. While all currently approved clinical HH inhibitors primarily exert their action by targeting the SMO receptor, a significant challenge arises from the acquisition of drug resistance and the inherent, *a priori* insensitivity observed in some tumors to SMO antagonists. These limitations underscore the critical and often overlooked role of SMO-independent activation mechanisms within the HH pathway in driving tumor progression and resistance. Previous studies have indeed highlighted the importance of such non-canonical activation of HH signaling in T-cell acute lymphoblastic leukemia (T-ALL), consistently identifying the GLI1 transcription factor as a preferred therapeutic target over the upstream SMO receptor in this specific leukemia subtype.
In this comprehensive work, we rigorously demonstrate a profound and consistent synergism between the GLI1 inhibitor GANT61 and the glucocorticoid (GC) dexamethasone across various T-ALL experimental models in vitro. These compelling findings lead us to propose that the combined targeting of the HH pathway and the GC receptor pathway represents a highly effective and rational therapeutic strategy for T-ALL. We have meticulously shown that synthetic GCs like dexamethasone are capable of significantly impairing GLI1 transcriptional activity in both HEK-293T and Jurkat T-ALL cells, and critically, this repressive effect is unequivocally dependent on the presence and activity of the GC receptor, NR3C1. The mechanistic link between HH and GCs has recently garnered increased scientific interest, with some studies identifying specific GC compounds as SMO agonists that promote its accumulation in the primary cilium, while others (such as Budesonide and Ciclesonide) have been reported to impair ciliary localization and, consequently, inhibit the HH pathway. However, it is important to emphasize that in many of these studies, HH pathway modulation was often presented as an off-target effect of GCs on the SMO receptor, typically occurring at much higher concentrations than those physiologically relevant for NR3C1 activation. Notwithstanding this, other research has demonstrated that dexamethasone, despite failing to compete for known SMO-binding sites, was still able to effectively inhibit the HH pathway in cell-based assays, which aligns well with our own findings. More recently, dexamethasone has been shown to inhibit the mitogenic effect of SHH, thereby impeding the proliferation of cerebellar neuronal precursors, although the exact underlying mechanism for this observation was not fully elucidated in those contexts.
Mechanistically, we propose a refined model of action wherein glucocorticoid-activated NR3C1 plays a pivotal role. This active NR3C1 initially promotes the accumulation of an acetylated form of GLI1. This is achieved through a dynamic and intricate interaction that favors the recruitment of PCAF acetyltransferase to GLI1 protein complexes, while simultaneously facilitating the dissociation of HDAC1 deacetylase from these complexes. This shift in the balance of acetylation and deacetylation profoundly affects GLI1 function, impacting both its transcriptional activity and its overall protein stability. Ultimately, this leads to a reduction in the aberrant oncogenic signaling driven by GLI1.
Our results strongly suggest that the powerful synergistic cytotoxic effect observed with the combination treatment (GC + GLI inhibitor) arises from a dual mechanism. Firstly, GLI inhibitors such as GANT61 appear to increase the sensitivity of leukemia cells to GCs, although the precise molecular mechanism underpinning this initial sensitizing effect warrants further detailed investigation in future studies. Secondly, and equally importantly, GCs themselves possess the distinct capacity to further attenuate the residual transcriptional activity of GLI1. This attenuation is likely mediated through the promotion of increased GLI1 protein degradation and a reduction in its ability to bind effectively to DNA. In conclusion, our study provides compelling evidence of a novel and clinically significant crosstalk between the glucocorticoid receptor signaling pathway and the Hedgehog pathway. This work highlights a previously undescribed role for NR3C1 as a negative regulator of GLI1 function and establishes a strong therapeutic rationale for combining GLI1 inhibitors with glucocorticoids in the treatment of T-ALL. While direct GLI1 inhibitors like GANT61 or arsenic trioxide (ATO) may not be ideal for broad clinical use due to unfavorable pharmacokinetic profiles or limited specificity, our findings suggest that indirect inhibition of GLI1 function, particularly by affecting its acetylation status through the use of HDAC inhibitors, could represent an intriguing and clinically viable alternative approach when combined with glucocorticoids. This warrants further rigorous clinical evaluation to translate these promising preclinical insights into improved outcomes for patients with Hedgehog-dependent malignancies.
Materials and Methods
Cell Lines and In Vitro Treatments
A comprehensive panel of T-cell acute lymphoblastic leukemia (T-ALL) cell lines, including CUTLL1, DND41, HSB2, MOLT4, CCRF-CEM, JURKAT E6, RPMI 8402, ALL-SIL, KOPTK1, and HPB-ALL, were routinely maintained in complete RPMI-1640 medium. This medium was specifically supplemented with 10% fetal bovine serum (FBS) and incubated at 37 °C under a 5% CO2 atmosphere to ensure optimal growth conditions. Additionally, UP-ALL13 cells were cultured in complete RPMI-1640 medium, with an elevated supplementation of 20% FBS to meet their specific growth requirements. HEK-293T cells were cultured in complete DMEM medium, similarly supplemented with 10% FBS. For functional studies that necessitated glucocorticoid (GC) treatment, with the exception of cell viability and apoptotic assays, both DMEM and RPMI-1640 media were meticulously supplemented with Charcoal-stripped FBS. This crucial step removes endogenous steroids, preventing confounding effects from hormones present in regular FBS.
For cell viability and apoptosis assays, cells were treated with increasing doses of dexamethasone, ranging from 1 nM to 1 μM, administered either as a single agent or in combination with GANT61 at concentrations between 10 and 30 μM. For the extensive drug combination screening, a diverse panel of inhibitors was used in conjunction with 2.5 μM GANT61. These inhibitors targeted various commonly deregulated oncogenic signaling pathways in T-ALL and included: DBZ (gamma-secretase inhibitor, targeting NOTCH1, at 2.5–20 nM), BEZ235 (mTOR inhibitor, at 2.5–20 nM), MK2206 (AKT inhibitor, at 5–40 nM), PF4708671 (S6K inhibitor, at 1–10 μM), H89 (PKA inhibitor, at 1–5 μM), Vitamin D3 (at 5–10 μM), PD98059 (MAPK inhibitor, at 1–10 μM), and GSK650394 (SGK inhibitor, at 1–5 μM). For detailed functional analysis of the HH pathway, cells were treated with increasing concentrations of dexamethasone (10 nM–1 μM), RU486 (a GC antagonist, at 1–10 μM), or trichostatin A (TSA, a pan-HDAC inhibitor, at 1 μM).
Quantitative Real-Time RT-PCR
The expression levels of specific genes were quantitatively assessed using quantitative real-time RT-PCR. The primer sequences meticulously designed for these reactions are listed in a supplementary table.