Metabolic reprogramming in macrophages: A glutamine rescue in β-catenin mutant hepatocellular carcinoma

Liver cancer ranks as the sixth most common cancer worldwide, with hepatocellular carcinoma (HCC) representing its predominant form. Despite the regulatory approval by the FDA of multiple immune checkpoint inhibitors (ICIs) for the treatment of advanced HCC, their response rates remain below 20%.1 The situation highlights the urgent need to understand the molecular mechanisms underlying resistance to ICI treatment and to develop reliable predictive biomarkers. The heterogeneity of treatment responses is likely rooted in various distinct oncogenic drivers.2 A key example is the CTNNB1 mutation, present in ~30% of HCC cases, which is distinctly correlated with the immune-excluded class of “cold” tumors and inferior ICI outcomes, forming a clear contrast with the profiles of TP53, PTEN, and KEAP1.3 While foundational research by Ruiz de Galarreta et al.4 established a clear mechanism by which β-catenin activation drives immune evasion and resistance toward anti-PD-1 treatment, clinically, it unveils a more intricate reality. In contrast to the previously thought uniform resistance, a subset of patients with CTNNB1-mutated HCC exhibits positive responses to ICIs, mounting effective T cell responses and benefiting from ICIs, as evidenced by an inflamed tumor microenvironment (TME) and elevated markers such as GZMA and CXCL9.5 This divergence suggests that the CTNNB1-mutant classification itself is insufficient, further warranting a characterization of the immune landscape within this genetic subgroup to unlock and better shape predictive biomarkers and deliver personalized ICI treatments, thereby enhancing patient survival outcomes. In this issue of HEPATOLOGY, Delgado et al. reported that heterogeneity exists within CTNNB1-mutated HCCs. The authors delineate a distinct and novel immune subclass characterized by low glutamine synthetase (GS) expression, stemming from reduced GLUL gene activity.6 This finding is mechanistically significant, given the crucial role of GS in detoxifying ammonia as a fundamental liver function by converting it to glutamate (Glu) and glutamine (Gln) consecutively, hence preventing it from entering the systemic circulation. This GS-low subclass is correlated with enhanced immune suppression and poorer patient survival, a finding that is particularly striking as it defies the established oncogenic pathway, where GLUL is a known downstream target of the Wnt/β-catenin signaling and robust GS expression is a hallmark of CTNNB1-mutated HCC. This discovery helps reframe the controversial prognostic role of GS in HCC. Meanwhile, Liu et al.7 demonstrated that overexpression of GS in HCC correlates with poor patient survival in moderately differentiated HCC cases and unfavorable prognosis in HCC patients undergoing sorafenib treatment.8 Conversely, GS has also been associated with improved survival outcomes following treatment with atezolizumab plus bevacizumab.9 This indicates that the influence of GS on patient survival in HCC may be contingent upon histological and molecular subtypes, as well as the specific treatment regimen a patient receives. In an effort to replicate the clinical finding that low GLUL expression is linked to poorer patient survival, Delgado et al. employed a method involving hydrodynamic tail vein injection of mutant CTNNB1 (T41A) and mutant NFE2L2 (G31A), alongside Sleeping Beauty transposase in Glulf/f mice. Their investigation revealed that tumor areas with lower GS expression exhibited increased cell proliferation in GLUL-competent HCCs, which is consistent with earlier studies demonstrating that GLUL ablation leads to increased cell proliferation.10 Utilizing a Molecular Cartography approach, they determined that the intrinsic molecular metabolic phenotype and Wnt target genes were comparable between the Glul-knockout (KO) and wild-type counterparts. This suggests that the tumor-promoting role of GS loss is not attributable to alterations in these intrinsic pathways. Consistent with this observation, the mTORC1-selective inhibitor (RMC-6272) suppressed tumor growth equally in both Glul-KO and control models, a finding consistent with their similar levels of mTOR pathway activity, as indicated by the expression level of S6 phosphorylation. A comparable anti-tumor effect was also achieved by directly targeting the oncogenic driver using a CTNNB1-specific siRNA (LNP-CTNNB1), confirming that the therapeutic strategy remains effective regardless of the tumor’s Glul expression. However, Glul-KO hepatic cells showed a significant increase in Glu, which was associated with a higher tumor burden and poorer survival. This suggests an alternative compensatory mechanism for Gln loss to support HCC growth. Furthermore, by adopting the Molecular Cartography approach, Delgado et al. discovered that the aggressive Glul-KO HCC cells were enriched with specific endothelial and myeloid cells, particularly marked by genes such as Lyz2, Cd68, Itgam, Adgre1, Pecam1, Fabp4, and Tek. By spatially resolving intra-tumoral and peri-tumoral regions, macrophages specifically aggregate within the tumor mass, further confirmed by mapping Cd68+ macrophages, showing their presence intra-tumorally or at the tumor boundary of Glul-KO HCCs. To investigate the effects of hepatic Gln loss on immune cells, the team Glul in non-tumor bearing mice and performed single-cell RNA sequencing, where results confirmed a distinct increase in macrophages in Glul-KO HCCs. Subsequent differential gene expression analysis revealed a comprehensive rewiring of macrophage gene expression in the Gln-deficient setting, characterized by immune exhaustion, enhanced expression of core macrophage identity genes, robust pro-inflammatory or antigen presentation signature, and increased expression of complement components and histocompatibility genes. Strikingly, these macrophages showed signs of metabolic stress, including reduced expression of genes for Glu biosynthesis and the urea cycle. This was accompanied by a broad downregulation of core energy pathways, including oxidative phosphorylation, fatty acid metabolism, and glycolysis, suggesting their exhausted state is underpinned by a severe metabolic deficit. Single-cell RNA sequencing of isolated myeloid cells demonstrated that Glul-KO HCCs exhibit a dramatic shift in population dynamics of tumor-associated macrophages (TAMs), compared with tissue-resident and peripherally recruited macrophages. This suggests that the loss of hepatic Gln synthesis acts as a driver for the recruitment and/or local expansion of specific macrophage subsets. To investigate the role of TAMs, the authors depleted CSF1R+ macrophages, noting that it is highly expressed in both tissue-resident and peripherally recruited macrophages. Depletion of CSF1R+ macrophages drastically retarded tumor growth in Glul-KO HCCs, where, importantly, the anti-tumor effect was more pronounced in the Glul-deficient tumors than in their Glul-competent counterparts. This specific vulnerability was further confirmed when broader deletion of all myeloid cells using clodronate treatment replicated the growth-inhibitory effect. Faced with Gln deprivation in Glul-KO HCCs, increased Glul expression was observed in macrophages, particularly a subset of tissue-resident cells marked by high Adgre1 expression, which responded by significantly increasing their own expression of GS to compensate for the acute reduction of Glu. To confirm that these macrophages were fueling tumor growth via induced GS expression, Glul was conditionally deleted from myeloid cells expressing Lyz2, resulting in a reduction in tumor growth. They revealed that compensatory glutamine production from HCC-associated macrophages significantly contributed to the maintenance and growth of HCC in Glul-KO β-catenin-mutated mice. Targeting GS in this specific population of macrophages achieved a dual benefit, where it directly reduced the overall tumor burden by cutting off this compensatory nutrient supply, and it also showed potential to enhance the efficacy of ICIs, likely by altering the immunosuppressive TME. Overall, this study offers new mechanistic insights with significant therapeutic implications for β-catenin-mutated HCC. First, the work highlights the clinical heterogeneity of β-catenin mutations, demonstrating that low expression of GS is associated with poorer patient survival. Second, increased tumor burden was observed in Glul-KO β-catenin-mutated mice, which demonstrates that the loss of GS creates a tumor-promoting effect, independent of their intrinsic molecular metabolic phenotype and the canonical Wnt signaling pathway, driven by changes in the TME. Third, despite the metabolic heterogeneity, these β-catenin mutation-driven HCC tumors remained sensitive to direct molecular targeting. Both mTORC1 inhibition and CTNNB1-specific siRNA therapy were effective in suppressing tumor growth, regardless of GS expression levels in the tumor. Fourth, the Glul-KO β-catenin-mutated mice displayed a profoundly remodeled microenvironment, marked by an enrichment of myeloid cells (especially macrophages) and endothelial cells. Fifth, in response to hepatic Glu loss, macrophages in these mice underwent metabolic reprogramming, shifting from elevated fatty acid oxidative phosphorylation to pro-inflammatory phenotypes accompanied by increased GS production to compensate for the nutrient deficit. This reprogramming contributes to immune T cell exhaustion (Figure 1), fostering “immune-cold” tumors. Crucially, depleting these macrophages reversed this immunosuppression, effectively converting “cold” tumor phenotype to “hot,” and could potentially sensitize them to immunotherapy.FIGURE 1: Schematic representation of the metabolic programming mechanisms in macrophages within β-catenin mutant HCC characterized by low GS expression. Abbreviations: Cd68, cluster of differentiation 68; Cd86, cluster of differentiation 68; Cd274, cluster of differentiation 274; GS, glutamine synthetase.The current mouse and clinical data presented by Delgado and colleagues are both impressive and encouraging, warranting further investigation. It would be interesting to explore the mechanisms underlying low GS expression in β-catenin-mutated HCCs, given that GS is a direct transcriptional target of the Wnt signaling pathway. While this study emphasizes the crucial role of macrophages in Glul-KO β-catenin-mutated mice, the potential involvement of enhanced endothelial cells in this subtype may also contribute to the increased tumor load. This enrichment could help elucidate prior findings indicating that low GS expression in HCC is associated with a favorable response to atezolizumab combined with bevacizumab.9 It would also be intriguing to examine whether metabolic adaptation in macrophages applies to HCC across various etiologies. Further exploration is warranted to determine if this holds true in MAFLD-induced HCC, which is an emerging risk factor and frequently associated with β-catenin mutations that are refractory to ICI therapy. Depleting these macrophages may offer valuable therapeutic insights for treating this specific type of HCC. Before pursuing this approach, it is crucial to examine whether the metabolic reprogramming of macrophages is indeed influenced by HCC cells exhibiting a glutamine metabolic deficiency. Additional studies involving the deletion of Glul in macrophages within non-CTNNB1-mutated HCC are necessary to explore this relationship. With a more in-depth analysis of Glul expression and macrophage activity in relation to outcomes of ICIs, along with proof-of-concept confirmation of PD-1 combination therapy with macrophage depletion, their findings have the potential to translate to personalized ICI therapies and serve as reliable biomarkers for predicting treatment responses. In summary, this study reveals a novel mechanism underlying the metabolic dynamics between HCC cells and macrophages within the TME. This interaction not only facilitates immune evasion but also significantly impacts the efficacy of ICI therapy. Ultimately, understanding these metabolic pathways could enhance therapeutic strategies and improve patient outcomes in the treatment of HCC.