Abstract

Background & Aims

   Dysregulation of one-carbon metabolism is considered an early hallmark
   of mitochondrial dysfunction and cancer metabolism. ALDH1L2 belongs to
   the aldehyde dehydrogenase family and plays an important role in tumor
   progression. However, little is known about the precise role and
   underlying mechanisms of ALDH1L2 in hepatocellular carcinoma (HCC).

Methods

   Immunohistochemistry, western blotting, and immunofluorescence staining
   were used to evaluate ALDH1L2 expression in HCC samples (n = 90) and
   cell lines (n = 9). A series of in vitro and in vivo assays were
   performed to explore the role and molecular mechanism of ALDH1L2 in HCC
   progression.

Results

   ALDH1L2 upregulation is associated with poor prognosis in HCC (hazard
   ratio 1.923; 95% confidence interval 1.03–3.59; p = 0.04). ALDH1L2
   promotes tumor cell proliferation and metastasis by activating
   NRF2/IL-6/STAT3 signaling. ALDH1L2 promotes mitochondrial respiration,
   increases ATP production and protects HCC cells from reactive oxygen
   species-induced cellular damage via NRF2 stabilization. NRF2 also
   directly binds to the ALDH1L2 promoter and increases ALDH1L2
   transcription, thereby establishing a positive feedback loop to
   maintain the function of ALDH1L2. The interaction between
   tumor-associated macrophages and ALDH1L2-overexpressing HCC cells
   further promotes HCC progression. In addition, ALDH1L2 knockdown
   enhances the anti-HCC activity of the tyrosine kinase inhibitor
   sorafenib.

Conclusions

   These findings provide the first evidence indicating that ALDH1L2 is
   directly involved in tumor progression by interacting with
   tumor-associated macrophages through the Jak2/STAT3 signaling pathway
   and that ALDH1L2 may be a target molecule for HCC therapy.

Impact and implications:

   This research highlights that ALDH1L2 could serve as a predictive and
   prognostic marker in HCC. We found that a positive feedback loop
   between ALDH1L2 and NRF2 promotes HCC progression by activating the
   IL-6/Jak2/STAT3 signaling axis and tumor-associated macrophage
   polarization. In addition, we found that ALDH1L2 knockdown enhances the
   anti-HCC effect of sorafenib.

   Keywords: ALDH1L2, Feedback loop, TAM, Crosstalk, Progression,
   Hepatocellular carcinoma

Graphical abstract

   [47]Image 1
   [48]Open in a new tab

Highlights:

     * •
       ALDH1L2 promotes HCC cell proliferation, migration, invasion, and
       metastasis.
     * •
       Elevated ALDH1L2 expression is associated with poor prognosis in
       HCC.
     * •
       ALDH1L2 promotes HCC progression by activating the IL-6/Jak2/STAT3
       signaling axis and promoting TAM polarization.
     * •
       ALDH1L2 knockdown sensitizes HCC cells to sorafenib.

Introduction

   Hepatocellular carcinoma (HCC) is the most common type of liver cancer
   and the third leading cause of cancer-related death worldwide.[49]^1
   Although great progress has been made in research on HCC in the past
   few decades, HCC still has a poor prognosis, mainly because of
   metastasis, recurrence, and drug resistance [[50]49]. However, the
   exact mechanisms of tumor occurrence and metastasis remain largely
   unclear. Thus, the identification of potential regulatory genes
   involved in HCC may reveal therapeutic targets to improve the prognosis
   of HCC.

   One-carbon metabolism includes the folate and methionine cycles and
   supports cellular requirements for growth and proliferation.
   Dysregulation of one-carbon metabolism is an early hallmark of
   mitochondrial dysfunction and cancer metabolism.[51]^2^,[52]^3
   Accumulating evidence suggests that one-carbon metabolism is critical
   for cancer development. One-carbon metabolic enzymes are important
   targets for cancer treatment. Limitations in the supply of one-carbon
   units may have therapeutic benefits for cancer patients. Antifolates
   are among the oldest antimetabolite classes of anticancer agents and
   were one of the first modern anticancer drugs.[53]^4 Methotrexate is
   used as an essential component of current chemotherapeutic regimens
   that exhibit substantial efficacy in the treatment of numerous
   malignancies.[54]^5 However, these drugs have many deleterious side
   effects owing to the importance of tetrahydrofolate in healthy tissues.
   Therefore, the identification of new chemotherapeutic targets in folate
   metabolism is warranted. The desuccinylation of the mitochondrial
   serine hydroxymethyltransferase SHMT2 is a pivotal signal in cancer
   cells to reprogram serine metabolism to support rapid growth.[55]^6
   Phosphoglycerate dehydrogenase (PHGDH) is the major rate-limiting
   enzyme in the first step of the serine–glycine–one-carbon (SGOC)
   metabolic pathway.[56]^7 Dysregulation of PHGDH influences serine
   synthesis and downstream one-carbon metabolism. The presence of PHGDH
   heterogeneity in primary tumors could be considered a marker for tumor
   aggressiveness.[57]^8 PHDGH is a potential therapeutic target for the
   treatment of metastatic tumors[58]^9 and targeting PHGDH is an
   effective approach to overcome tyrosine kinase inhibitor (TKI) drug
   resistance in HCC.[59]^10 Methylenetetrahydrofolate dehydrogenase 1
   like plays an essential role in supporting HCC cell growth.[60]^11
   Methylenetetrahydrofolate dehydrogenase 2 contributes to the
   progression of renal cell carcinoma via a novel epitranscriptomic
   mechanism involving Hypoxia-inducible factor-2α.[61]^12 Understanding
   the importance of one-carbon metabolism in cancer cell growth is
   anticipated to lead to the development of more selectively inhibiting
   individual one-carbon pathway enzymes for cancer treatment and
   intervention.

   Aldehyde dehydrogenase 1 family member L1 (ALDH1L1) and aldehyde
   dehydrogenase 1 family member L2 (ALDH1L2) are two key
   folate-metabolizing enzymes. ALDH1L1 is a cytosolic
   10-formyltetrahydrofolate dehydrogenase, whereas ALDH1L2 is the
   mitochondrial homolog of 10-formyltetrahydrofolate
   dehydrogenase.[62]^13 The biological roles of ALDH1L1 and ALDH1L2 are
   quite different although the two proteins share approximately 72% amino
   acid sequence identity and are close structural and enzymatic
   homologs.[63]^13^,[64]^14 ALDH1L1 inhibits tumor cell motility via
   dephosphorylation of cofilin by PP1 and PP2A in A549 cells.[65]^15
   Downregulation of ALDH1L1 expression is associated with poor prognosis
   in HCC.[66]^16 ALDH1L1 knockout promotes liver tumor growth.[67]^17
   These results suggest that ALDH1L1 can function as a tumor suppressor.
   Previous studies have shown that high expression of ALDH1L2 is
   associated with poor prognosis in pancreatic ductal
   adenocarcinoma.[68]^18 ALDH1L2 is upregulated in human colorectal tumor
   tissues compared with normal tissues.[69]^19 Recent reports have shown
   that colorectal cancer patients with low ALDH1L2 expression exhibit
   radioresistance.[70]^20 Previous reports have shown that oxidative
   stress inhibits the distant metastasis of human melanoma cells. These
   reports also showed that ALDH1L2 was much more highly expressed in
   liver metastatic nodules than in subcutaneous tumors in both donor and
   recipient mice.[71]^21 However, new research indicates that the loss of
   ALDH1L2 can support metastatic progression by promoting formate and
   formyl-methionine (fMet) production in breast cancer cells.[72]^22
   These results indicate that ALDH1L2 plays a cellular context-dependent
   role in different tumor types. However, evidence on the role and
   clinical significance of ALDH1L2 expression in HCC is limited. Thus,
   the function and molecular mechanism underlying the role of ALDH1L2 in
   HCC and the relationships between ALDH1L2 expression and
   clinicopathological parameters remain unclear. Here, we examined the
   potential role of ALDH1L2 in HCC progression and explored the mechanism
   by which ALDH1L2 affects HCC cell proliferation and metastasis.

Materials and methods

   Further information can be found in the [73]Supplementary material,
   Methods section.

   The sequences of shRNA targets are listed in [74]Table S1.

   The primer sequences are listed in [75]Table S2.

   Information on the antibodies is provided in [76]Table S3.

Mitochondrial oxygen consumption assay

   The oxygen consumption rate (OCR) was evaluated with a Mito Stress Test
   Kit (Agilent, Cat. No. 103015–100, California, CA, USA) as described
   previously.[77]^23 Mitochondrial respiration inhibitors (oligomycin,
   FCCP, and antimycin A + rotenone) were used to treat the cells in the
   assay system.

Orthotopic and subcutaneous tumor model models and treatment

   For the in vivo xenograft experiment, a microsyringe was used to
   orthotopically inoculate nude mice in the left hepatic lobe through an
   8-mm transverse incision made in the upper abdomen under
   anesthesia[[78]48]. A total of 2 × 10^6 HCC cells suspended in 40 μl of
   a mixture of serum-free DMEM/Matrigel (1:1 volume) (BD Biosciences, MA,
   USA) were inoculated into each male nude mouse.

   For the in vivo isograft experiment, male C57BL/6J mice were
   subcutaneously injected with 5 × 10^6 Hepa1-6 derived cells. After 24
   days, the mice were sacrificed.

   For in vivo drug studies, 6-week-old male BALB/c nude mice were
   injected subcutaneously with 2 × 10^6 cells. When the tumor volume was
   approximately 100 mm^3, the mice were randomized into four groups. The
   mice were treated with sorafenib (10 mg/kg, once every other day) via
   intraperitoneal injection. The tumor dimensions were measured with a
   Vernier calipers every 3 days, and the tumor volume was calculated as
   follows: tumor volume = (length × width^2)/2.

   Four or six weeks later, the mice were sacrificed, and the tumors along
   with the liver and lung tissues from individual mice were excised and
   fixed with 4% phosphate-buffered neutral formalin for at least 72 h.
   Metastatic tissues were analyzed via H&E staining. All animal
   experiments were approved by the Laboratory Animal Ethics Committee of
   Renji Hospital, Shanghai Jiao Tong University School of Medicine
   (approval number: RT2022-122u).

Statistical analysis

   All data are shown as the mean ± standard deviation (SD) values and
   were analyzed with IBM SPSS Statistics 21 (IBM, Chicago, IL, USA).
   Graphs were generated with GraphPad Prism 7.0 (GraphPad Software, San
   Diego, CA, USA). Statistical comparisons of the data were performed via
   two-tailed Student t tests or via one-way ANOVA for multiple
   comparisons among more than two groups. Overall survival curves were
   generated via the Kaplan–Meier method and compared via the log-rank
   test. Univariate and multivariate analyses were performed with a
   stepwise Cox proportional hazard regression model. A value of p <0.05
   was considered to indicate statistical significance (∗p <0.05, ∗∗p
   <0.01).

Results

ALDH1L2 is a marker of poor prognosis in HCC patients

   ALDH1L2 expression was upregulated in HCC tissues compared with
   noncancerous tissues according to analysis of The Cancer Genome Atlas
   (TCGA) and Gene Expression Omnibus (GEO) datasets and the ICGC-LIRI-JP
   cohort ([79]Fig. 1A–E). These results were also confirmed by western
   blotting (WB) and immunohistochemistry (IHC) staining ([80]Fig. 1F and
   G). Furthermore, in the TCGA dataset, ALDH1L2 expression was
   upregulated in patients with nodal metastasis compared with patients
   without nodal metastasis ([81]Fig. 1H). In addition, ALDH1L2
   upregulation was confirmed to be related to tumor metastasis in
   datasets for other solid cancers from the TCGA ([82]Fig. S1A) and GEO
   datasets ([83]Fig. S2).

Fig. 1.

   [84]Fig. 1
   [85]Open in a new tab

   ALDH1L2 upregulation is associated with poor prognosis in HCC.

   (A) Heatmap showing ALDH1L2 expression in the TCGA dataset, GEO dataset
   and ICGC-LIRI-JP cohorts. (B–E) ALDH1L2 expression in HCC tissues was
   compared with that in the corresponding noncancerous liver tissues in
   the TCGA dataset (n = 50) (B), [86]GSE22058 dataset (C), [87]GSE67680
   dataset (D), and ICGC-LIRI-JP cohort (E). (F) ALDH1L2 expression in HCC
   tissues was compared with that in the corresponding noncancerous liver
   tissues via Western blotting. (G) Immunohistochemical analysis of
   ALDH1L2 expression in HCC samples and corresponding noncancerous liver
   tissues. Representative images are shown. (H) ALDH1L2 expression in
   noncancerous liver tissues, HCC tissues and metastatic cancer tissues
   was analyzed via TCGA data. (I) Representative images of cells with
   high and low ALDH1L2 expression. (J,K) OS and DFS of patients with HCC
   stratified by the ALDH1L2 expression level, Log-rank test. (L,M)
   Univariate and multivariate Cox proportional hazards regression
   analyses were conducted to evaluate the effect of ALDH1L2 on overall
   survival in patients with HCC. Data are represented as mean ± SD. ∗p
   <0.05; ∗∗p <0.01. Unpaired t-test. DFS, disease-free survival; GEO,
   Gene Expression Omnibus; HCC, hepatocellular carcinoma; OS, overall
   survival; TCGA, The Cancer Genome Atlas; ICGC-LIRI-JP, International
   Cancer Genome Consortium liver cancer-RIKEN Japan.

   The clinicopathological features of HCC patients (n = 90) are shown in
   [88]Table S4. On the basis of the IHC results, the patients were
   divided into two groups on the basis of ALDH1L2 expression in the tumor
   tissue ([89]Fig. 1I). ALDH1L2 expression was positively associated with
   TNM stage and capsule invasion status. However, there were no
   correlations between ALDH1L2 expression and other clinicopathological
   factors, such as sex, age, tumor grade, cirrhosis status,
   serum alpha-fetoprotein level, portal vein tumor thrombus, vessel
   carcinoma embolus, tumor size, and HBV positivity status
   ([90]Table S5). We next evaluated the prognostic value of ALDH1L2 in
   HCC. Survival analysis of the HCC patients based on the IHC staining
   results was performed, and the results indicated that patients with
   higher ALDH1L2 expression had markedly worse overall survival (OS; p =
   0.035) and disease-free survival (DFS; p = 0.044) than patients with
   lower ALDH1L2 expression ([91]Fig. 1J and K). Furthermore, univariate
   and multivariate Cox proportional hazard analyses suggested that higher
   expression of ALDH1L2 was more strongly associated with worse survival
   in HCC patients than low ALDH1L2 expression in HCC patients (p <0.05;
   [92]Fig. 1L and M). In addition, high expression of ALDH1L2 was
   associated with shorter OS times in patients with various types of
   tumors according to analysis of TCGA data ([93]Fig. 1B). Thus, these
   findings suggested that ALDH1L2 could be a valuable prognostic factor
   for survival in patients with HCC.

ALDH1L2 promotes HCC cell growth in vitro and in vivo

   On the basis of the endogenous expression of ALDH1L2 in HCC cell lines,
   the Huh7, MHCC-97H, Hep3B, SNU387, and Li7 cell lines were selected for
   gain- and loss-of-function studies ([94]Fig. S3). Overexpression of
   ALDH1L2 increased the proliferation of HCC cells, whereas knockdown of
   ALDH1L2 decreased the proliferation of HCC cells ([95]Fig. 2A–H,
   [96]Fig. S4). Overexpression of ALDH1L2 promoted the G1/S transition in
   Huh7 cells ([97]Fig. 2I, [98]Fig. S5).

Fig. 2.

   [99]Fig. 2
   [100]Open in a new tab

   ALDH1L2 increases HCC cell proliferation, invasion and metastasis.

   The expression level of ALDH1L2 was measured by Western blotting in
   ALDH1L2-overexpressing (A) and ALDH1L2-knockdown (B) HCC cells. Cell
   proliferation was assessed by a CCK8 assay (C,D), a colony formation
   assay (E,F) and an EdU incorporation assay (G,H). (I) The cell cycle
   distribution of ALDH1L2-overexpressing Huh7 cells was analyzed by flow
   cytometry. (J) Liver tissues from animals bearing xenografts derived
   from MHCC-97H cells with stable ALDH1L2 overexpression or from Li7
   cells with stable ALDH1L2 knockdown. The dot plots show the results of
   the quantitative analysis of liver weight. (K) Ki67 expression in
   xenografts derived from MHCC-97H ALDH1L2-overexpressing or
   Li7-shALDH1L2 cells and the corresponding control vector-transduced
   cells was evaluated via IHC staining. (L,M) The effects of ALDH1L2
   overexpression (L) and knockdown (M) on HCC cell migration and invasion
   were assessed via Transwell assays. (N) Representative images of
   intrahepatic nodules and lung nodules formed by ALDH1L2-overexpressing
   MHCC-97H cells, ALDH1L2-knockdown Li7 cells and the corresponding
   control cells are shown. The numbers of intrahepatic metastatic nodules
   and lung metastatic nodules are shown in the right panel. (O) IHC
   analysis of metastasis with human-specific anti-mitochondria
   antibodies. Bar = 50 μM. Data are represented as mean ± SD. ∗p <0.05;
   ∗∗p <0.01. Unpaired t-test or one-way ANOVA. CCK8, Cell Counting Kit-8;
   HCC, hepatocellular carcinoma; IHC, immunohistochemistry.

   In the in vivo mouse xenograft model, overexpression of ALDH1L2
   increased HCC tumor growth, whereas knockdown of ALDH1L2 decreased HCC
   tumor growth ([101]Fig. 2J). Furthermore, Ki67 expression was higher in
   ALDH1L2-overexpressing tumor tissues and lower in ALDH1L2-knockdown
   tumor tissues ([102]Fig. 2K). Therefore, these results show that
   ALDH1L2 promotes HCC cell proliferation.

ALDH1L2 promotes HCC cell invasion and metastasis

   Overexpression of ALDH1L2 dramatically increased cell migration and
   invasion, whereas knockdown of ALDH1L2 had the opposite effect
   ([103]Fig. 2L and M, [104]Fig. S6). Matrix metalloproteinases (MMPs)
   are proteolytic enzymes that play major roles in tumor invasion. We
   found that the expression of MMP9 and MMP2 was upregulated in
   ALDH1L2-overexpressing cells and downregulated in ALDH1L2-knockdown
   cells ([105]Fig. S7).

   We next investigated the effect of ALDH1L2 on HCC metastasis in vivo.
   Histological examination of lung and liver tissues revealed that more
   numbers of intrahepatic metastatic nodules and lung metastases in mice
   bearing tumors derived from ALDH1L2-overexpressing cells than in mice
   bearing tumors derived from control cells, with the opposite pattern
   found for ALDH1L2 knockdown ([106]Fig. 2N). To further confirm
   metastasis, we used a human-specific antimitochondrial antibody to
   confirm the human origin of the cells and found more positively stained
   regions in the tumors derived from ALDH1L2-overexpressing cells. In
   contrast, compared with tumors derived from shNC control cells, tumors
   derived from ALDH1L2-knockdown cells exhibited a regular boundary and
   fewer positively stained regions ([107]Fig. 2O). Taken together, these
   findings suggest that ALDH1L2 promotes HCC metastasis.

ALDH1L2 is critical for maintaining mitochondrial-associated metabolism in
HCC cells

   Cellular NADPH is essential for maintaining redox homeostasis and
   maintaining reductive biosynthesis, which are indispensable for
   tumorigenesis and cancer progression.[108]^24 Furthermore, the results
   of the TCGA analysis revealed that the expression of ALDH1L2 was
   associated with the activity of the reactive oxygen species (ROS)
   pathway in multiple tumor types including HCC ([109]Fig. S8). The
   colocalization of ALDH1L2 with a mitochondrial indicator (pDsRed1-Mito)
   was also detected in HCC cells ([110]Fig. S9). Therefore, we examined
   the effects of ALDH1L2 on the NADPH/NADP^+ ratio and intracellular ROS
   levels. Overexpression of ALDH1L2 decreased intracellular ROS levels
   and increased the intracellular NADPH/NADP^+ ratio in HCC cells,
   whereas ALDH1L2 knockdown had the opposite effects ([111]Fig. 3A–D,
   [112]Fig. S10). Furthermore, we found that the overexpression of
   ALDH1L2 suppressed H[2]O[2]-induced ROS generation in HCC cells,
   whereas ALDH1L2 knockdown had the opposite effect ([113]Fig. 3E and F).
   Collectively, these data indicate that ALDH1L2 is critical for
   maintaining the NADPH levels for redox homeostasis and thereby supports
   HCC cell proliferation.

Fig. 3.

   [114]Fig. 3
   [115]Open in a new tab

   ALDH1L2 induces ROS accumulation via the regulation of NRF2 stability.

   (A) The level of ROS in ALDH1L2-overexpressing and control cells was
   assessed by flow cytometry. (B) The intracellular NADPH/NADP^+ ratio
   was determined in ALDH1L2-overexpressing and control cells. (C) The
   level of ROS in ALDH1L2-knockdown cells was assessed via flow
   cytometry. (D) The intracellular NADPH/NADP^+ ratio was determined in
   ALDH1L2 knockdown cells. (E,F) The level of ROS was assessed by flow
   cytometry in ALDH1L2-overexpressing (E) and ALDH1L2-knockdown (F) cells
   treated with H[2]O[2]. (G) Mitochondrial ROS levels in cells stained
   with MitoSOX Red were evaluated by flow cytometry. (H) The mtMP was
   assessed with the fluorescent mitochondrial probe JC-1. (I–M) The OCR
   was measured with a Seahorse XF24 Analyzer as detailed in the Materials
   and methods section (I). The basal mitochondrial respiration rate (J),
   maximal respiratory capacity (K), ATP production (L), and spare
   respiratory capacity (M) were calculated and statistically analyzed.
   (N,O) NRF2 expression in ALDH1L2-overexpressing (N) and
   ALDH1L2-knockdown (O) cells was measured by Western blotting. (P) NRF2
   expression was evaluated by immunofluorescence staining. (Q) Stability
   of NRF2 in ALDH1L2-knockdown HCC cells treated with cycloheximide (CHX)
   at the indicated time points. (R) NRF2 expression in ALDH1L2-knockdown
   HCC cells treated with MG132 for 6 h. (S) Lysates of
   ALDH1L2-overexpressing and control HCC cells were subjected to
   immunoprecipitation with an anti-NRF2 antibody, and the immunocomplexes
   were analyzed by immunoblotting with an antibody specific for
   ubiquitin. Data are represented as mean ± SD.∗p <0.05; ∗∗p <0.01.
   Unpaired t-test or one-way ANOVA. NRF2, nuclear factor E2-related
   factor 2; ROS, reactive oxygen species; OCR, oxygen consumption rate.

   As mitochondria are a major source of ROS, we next examined ROS levels
   in mitochondria via MitoSOX Red staining. The mitochondrial superoxide
   anion level was significantly higher in ALDH1L2-knockdown cells than in
   the corresponding control cells ([116]Fig. 3G). As the mitochondrial
   membrane potential (mtMP) is a key indicator of mitochondrial activity,
   we examined the mtMP in ALDH1L2-knockdown HCC cells via a JC-1 assay.
   Our results showed that the mtMP in ALDH1L2-knockdown cells was lower
   than that in the corresponding control cells ([117]Fig. 3H).

   To determine whether the knockdown of ALDH1L2 affects mitochondrial
   respiration, we next evaluated the OCR, ATP production, and maximal
   respiratory capacity. The results revealed that the knockdown of
   ALDH1L2 reduced basal cellular respiration, decreased the intracellular
   ATP levels, and reduced the spare respiratory capacity and the
   mitochondrial OCR ([118]Fig. 3I–M). Therefore, these results indicate
   that ALDH1L2 promotes cellular respiration and the generation of ATP
   and oxidative metabolites in tumor cells.

ALDH1L2 increases the expression of nuclear factor E2-related factor 2 and
prevents its degradation in HCC cells

   As a leucine zipper transcription factor, nuclear factor E2-related
   factor 2 (NRF2) plays an important role in the maintenance of redox
   homeostasis.[119]^25 We next explored the effect of ALDH1L2 on NRF2
   expression in HCC cells. Overexpression of ALDH1L2 promoted NRF2
   protein expression in HCC cells, whereas ALDH1L2 knockdown had the
   opposite effect ([120]Fig. 3N and O). However, ALDH1L2 overexpression
   and knockdown did not affect NRF2 mRNA expression in HCC cells
   ([121]Fig. S11). The results of the immunofluorescence assay revealed
   that the expression and nuclear translocation of NRF2 were increased in
   ALDH1L2-overexpressing HCC cells compared with vector-transduced cells
   ([122]Fig. 3P, [123]Fig. S12).

   To confirm the influence of ALDH1L2 on NRF2 protein stability, we
   incubated ALDH1L2-knockdown and control HCC cells with cycloheximide
   (CHX) (50 μg/ml), which blocks de novo protein synthesis. In the
   presence of cycloheximide, knockdown of ALDH1L2 resulted in a higher
   rate of NRF2 degradation than detected in control cells, suggesting
   that the expression of the ALDH1L2 protein increases the stability of
   the NRF2 protein ([124]Fig. 3Q). The half-life of the NRF2 protein in
   the cells decreased from 1 h to 15 min as a consequence of ALDH1L2
   knockdown ([125]Fig. 3Q). To investigate the involvement of the
   ubiquitin‒proteasome pathway in the proteolytic degradation of NRF2, we
   treated ALDH1L2-knockdown HCC cells with MG132, a reversible proteasome
   inhibitor. Our results revealed that NRF2 degradation induced by
   ALDH1L2 knockdown was inhibited in the presence of MG132
   ([126]Fig. 3R). In addition, overexpression of ALDH1L2 resulted in
   decreased NRF2 ubiquitination ([127]Fig. 3S). NRF2 expression is
   regulated primarily by Kelch-like ECH-associated protein 1 (KEAP1), a
   component of the Cullin 3 (CUL3)-based E3 ubiquitin ligase
   complex.[128]^26 We found that the overexpression of ALDH1L2 suppressed
   KEAP1 expression, whereas the knockdown of ALDH1L2 had the opposite
   effect ([129]Fig. S13A). To confirm the role of KEAP1 in the promoting
   effect of ALDH1L2 on NRF2 stability, we evaluated the expression of
   NRF2 in ALDH1L2-knockdown and ALDH1L2-overexpressing cells following
   the introduction of a KEAP1 overexpression plasmid or a KEAP1 shRNA.
   The ALDH1L2 overexpression-induced increase in NRF2 protein expression
   was reversed by the overexpression of KEAP1. Conversely, ALDH1L2
   knockdown-induced decrease in NRF2 protein expression was reversed by
   the knockdown of KEAP1. Furthermore, ALDH1L2 attenuated NRF2
   ubiquitination caused by overexpression of KEAP1 in HCC cells
   ([130]Fig. S13B and C). Collectively, these findings suggest that
   ALDH1L2 can stabilize NRF2 by directly preventing its
   ubiquitin–proteasome-mediated degradation.

ALDH1L2 regulates HCC cell functions via the Jak2/STAT3 pathway

   To explore the mechanism of ALDH1L2 in HCC, RNA-seq was performed in
   ALDH1L2-overexpressing cells. Kyoto Encyclopedia of Genes and Genome
   (KEGG) pathway enrichment analysis revealed that ALDH1L2 can regulate
   the Jak/signal transducer and activator of transcription 3 (STAT3)
   pathway in HCC cells ([131]Fig. 4A). The IL-6/Jak/STAT3 signaling
   pathway was also enriched according to analysis of data obtained from
   the TCGA ([132]Fig. S14). In addition, NRF2 has been reported to
   directly induce IL-6 transcription in hepatocytes.[133]^27 Furthermore,
   ROS modulate various cellular signaling pathways, including the STAT3
   pathway.[134]^28 Therefore, we investigated the effects of ALDH1L2 on
   the IL-6/Jak/STAT3 pathway. Our results revealed that Jak2 but not Jak1
   was activated in ALDH1L2-overexpressing HCC cells ([135]Fig. 4B,
   [136]Fig. S15A). The overexpression of ALDH1L2 increased the
   phosphorylation of STAT3, whereas ALDH1L2 knockdown had the opposite
   effect ([137]Fig. 4B and C). Moreover, ALDH1L2 overexpression
   facilitated STAT3 nuclear entry in HCC cells ([138]Fig. 4D,
   [139]Fig. S15B). Furthermore, knockdown of ALDH1L2 inhibited the
   transcription of STAT3 target genes ([140]Fig. S16).

Fig. 4.

   [141]Fig. 4
   [142]Open in a new tab

   ALDH1L2 regulates HCC cell functions via the Jak2/STAT3 pathway.

   (A) KEGG pathway analysis based on ALDH1L2 in HCC. (B,C) The protein
   levels of Jak2, p-Jak2, STAT3, and p-STAT3 in ALDH1L2-overexpressing
   (B) and ALDH1L2-knockdown (C) HCC cells were measured by Western
   blotting. (D) Effects of ALDH1L2 overexpression on the nuclear
   translocation of STAT3 as assessed by an immunofluorescence assay.
   (E,F) ALDH1L2-overexpressing HCC cells were treated with BP-1-102 or
   DMSO as indicated, and the level of STAT3 was measured by Western
   blotting (E) and evaluated by immunofluorescence staining (F). (G–J)
   ALDH1L2-overexpressing HCC cells were treated with BP-1-102 as
   indicated, and cell proliferation, migration and invasion were
   evaluated via a CCK8 assay (G), a colony formation assays (H), an EdU
   incorporation assay (I) and Transwell assays (J). (K)
   ALDH1L2-overexpressing HCC cells were transduced with STAT3 shRNA as
   indicated, and the level of STAT3 was measured by Western blotting.
   Cell proliferation, migration and invasion were evaluated via a CCK8
   assay (L), an EdU incorporation assay (M,N), a colony formation assay
   (O), and Transwell assays (P). Data are represented as mean ± SD. ∗p
   <0.05; ∗∗p <0.01. Unpaired t-test or one-way ANOVA. CCK8, Cell Counting
   Kit-8; HCC, hepatocellular carcinoma; KEGG, Kyoto Encyclopedia of Genes
   and Genome; STAT3, signal transducer and activator of transcription 3.

   Next, HCC cells were treated with a STAT3 inhibitor (BP-1-102) or
   transduced with STAT3 shRNA. The ALDH1L2 overexpression-induced
   increases in cell proliferation, migration, and invasion were reversed
   by treatment with BP-1-102 or STAT3 shRNAs ([143]Fig. 4E–P,
   [144]Fig. S15D–K). Furthermore, BP-1-102 suppressed STAT3 nuclear
   translocation in ALDH1L2-overexpressing HCC cells ([145]Fig. 4F).

   In addition, we found that the overexpression of ALDH1L2 upregulated
   IL-6 expression, whereas the knockdown of ALDH1L2 downregulated IL-6
   expression in HCC cells ([146]Fig. 5A). To confirm whether IL-6R
   participates in ALDH1L2-mediated activation of the Jak2/STAT3 pathway,
   ALDH1L2-overexpressing HCC cells were transduced with IL-6R shRNA. The
   increase in Jak2/STAT3 activity induced by ALDH1L2 overexpression was
   reversed by IL-6R shRNA transduction ([147]Fig. S17). Therefore, these
   results suggest that ALDH1L2 regulates HCC cell functions by activating
   the IL-6/Jak2/STAT3 signaling pathway.

Fig. 5.

   [148]Fig. 5
   [149]Open in a new tab

   ALDH1L2 promotes HCC progression by affecting TAM polarization.

   (A) The expression of IL-6 and IL-10 was analyzed via qRT-PCR in
   ALDH1L2-overexpressing and ALDH1L2-knockdown HCC cells. (B,C) After
   coculturing HCC cells (ALDH1L2-overexpressing or control cells) with
   THP-1-derived macrophages for 48 h, the relative mRNA levels of CD206
   and CD163 in the macrophages were measured via qRT-PCR. Migration and
   invasion were assessed via Transwell assays. (D) CD206 expression was
   examined by flow cytometry. (E,F) THP-1-derived macrophages cells were
   pretreated with IL-6 for 24 h and were then cocultured with HCC cells
   in a Transwell apparatus for 48 h. Cell migration and invasion were
   examined via a Transwell assay. (G-K) Hepa1-6 cells stably transduced
   with ALDH1L2 or the corresponding control construct were subcutaneously
   injected into C57BL/6J mice. The dot plots show the results of the
   quantitative analysis of tumor weight (H). The concentrations of IL-6
   and IL-10 in the serum of C57BL/6J mice were measured via ELISA (I).
   Flow cytometric analysis of the CD45^+CD11b^+ Ly6G^-F4/80^+ CD206^+-M2
   TAM population in tumor tissue (J). The expression of F4/80 and CD206
   in xenograft and isograft mouse tumor tissues was evaluated via IHC
   staining (K). (L) The expression of CD206 and ALDH1L2 in HCC tissue was
   evaluated via IHC staining. The correlation between ALDH1L2 and CD206
   expression in HCC tissues was analyzed. (M,N) After coculturing HCC
   cells (ALDH1L2-overexpressing or control cells) with THP-1-derived
   macrophages with or without IL-6R knockdown for 48 h, NF-κB protein
   expression was assessed via WB (M), and CD206 expression was examined
   via flow cytometry (N). Data are represented as mean ± SD. ∗p <0.05;
   ∗∗p <0.01; ns, non-significant. Unpaired t-test or one-way ANOVA. HCC,
   hepatocellular carcinoma; IHC, immunohistochemistry; qRT-PCR,
   quantitative real-time-PCR; TAM, tumor-associated macrophage.

ALDH1L2 promotes HCC progression by affecting tumor-associated macrophage
polarization

   The results of the gene ontology functional enrichment analysis
   revealed that ALDH1L2 can regulate the cytokine‒cytokine receptor
   interaction pathway in HCC cells ([150]Fig. 4A). Moreover, we found
   that overexpression of ALDH1L2 increased the expression of IL-6 and
   IL-10, whereas ALDH1L2 knockdown had the opposite effect
   ([151]Fig. 5A). Dysregulation of the inflammatory microenvironment can
   influence the crosstalk between cancer cells and components of the
   tumor microenvironment. Tumor-associated macrophages (TAMs) are
   important negative regulators of host innate immunity in the tumor
   microenvironment. IL-6 is recognized to be involved in alternative
   macrophage polarization.[152]^29^,[153]^30 To examine whether cancer
   cell invasion is affected by macrophages, coculture experiments with
   THP-1-derived macrophages were performed. The coculture experiments
   revealed upregulated expression of M2 polarization-related genes
   (CD163, CD206, and IL-10) in ALDH1L2-overexpressing cells
   ([154]Fig. 5B–D). Furthermore, HCC cell migration and invasion were
   significantly increased in the presence of THP-1-derived macrophages
   pretreated with IL-6 ([155]Fig. 5E–F).

   To confirm the role of TAMs in ALDH1L2-related HCC progression,
   ALDH1L2-overexpressing Hepa1-6 mouse HCC cells (Hepa1-6/ALDH1L2 cells)
   were transplanted into immunocompetent C57BL/6J mice. Overexpression of
   ALDH1L2 promoted tumor growth ([156]Fig. 5G–H-). Furthermore,
   overexpression of ALDH1L2 increased IL-6 and IL-10 secretion in
   C57BL/6J mice ([157]Fig. 5I-). Next, flow cytometry was used to
   evaluate TAM, myeloid-derived suppressor cell (MDSC), and regulatory T
   cell (Treg cell) infiltration in the harvested tumors. Implantation of
   Hepa1-6/ALDH1L2 cells significantly increased M2 TAM (marked by the
   CD45^+/CD11b^+/F4/80^+/CD206^+/Ly6G^- signature) infiltration
   ([158]Fig. 5J). Infiltration of MDSCs (marked by the
   CD45^+/CD11b^+/Ly6G^+ signature) and Treg cells (CD45^+/CD4^+/FOXP3^+)
   was not observed ([159]Fig. S18). Furthermore, IHC revealed that
   CD206-positive M2 macrophages were more abundant in tumors derived from
   ALDH1L2-overexpressing MHCC-97H cells but less abundant in tumors
   derived from ALDH1L2-knockdown Li7 cells than in tumors derived from
   the corresponding control cells ([160]Fig. 5K). Moreover, ALDH1L2
   expression is associated with the expression of M2 polarization-related
   genes (CD163, CD206, and IL-10) and M2 macrophage infiltration in many
   cancer types, including HCC, according to analysis of TCGA data
   ([161]Figs S19–S21). Furthermore, IHC staining revealed a positive
   correlation between ALDH1L2 and CD206 expression in HCC ([162]Fig. 5L).
   Therefore, these results demonstrate that ALDH1L2 impacts the
   recruitment and polarization of TAMs.

   We next examined NF-κB expression in THP-1-derived macrophages. The
   results revealed that the NF-κB pathway was activated in THP-1-derived
   macrophages treated with conditioned medium derived from
   ALDH1L2-overexpressing cells ([163]Fig. 5M). IL-6R silencing was
   insufficient to promote NF-κB activation in THP-1-derived macrophages
   treated with conditioned medium derived from ALDH1L2-overexpressing
   cells ([164]Fig. 5M). Furthermore, IL-6R silencing inhibited CD206
   expression in THP-1-derived macrophages treated with conditioned medium
   from ALDH1L2-overexpressing cells ([165]Fig. 5N).

   M2-polarized TAMs secreted an increased amount of IL-10. The NF-κB p65
   subunit can bind upstream of the IL10 gene and regulate the production
   of the IL-10 protein.[166]^31 Furthermore, we found that IL-10
   significantly promoted the migration and invasion of HCC cells as well
   as STAT3 activation in these cells. IL-10RA silencing reversed the
   effects of IL-10 on the activation of STAT3 in HCC cells
   ([167]Fig. S22). Therefore, ALDH1L2 impacts the polarization of TAMs,
   which secrete IL-10 to mediate the activity of a positive feedback loop
   in HCC cells.

NRF2 promotes the transcription of ALDH1L2 via a positive feedback loop

   We found that the expression of ALDH1L2 was upregulated during
   H[2]O[2]-induced ROS generation in HCC cells ([168]Fig. 6A,
   [169]Fig. S23A). Sulforaphane (SFN) is an activator of NRF2. We found
   that SFN promoted ALDH1L2 expression in a dose- and time-dependent
   manner in HCC cells ([170]Fig. 6B and C, [171]Fig. S23B). Furthermore,
   knockdown of NRF2 suppressed ALDH1L2 expression in HCC cells
   ([172]Fig. 6D and E). There are two putative antioxidant response
   elements (AREs) in the promoter of ALDH1L2 (-1,200 bp/+1 bp,
   [173]Fig. 6F and G). Serial truncations of the ALDH1L2 promoter were
   generated on the basis of the location of the NRF2-binding sites
   ([174]Fig. 6H). SFN was found to increase ALDH1L2 promoter activity
   ([175]Fig. 6I). Knockdown of NRF2 inhibited ALDH1L2 promoter activity
   in HCC cells ([176]Fig. 6J). The chromatin immunoprecipitation (ChIP)
   assay revealed that AREs (−280 to −270, −703 to −693) were
   immunoprecipitated with the anti-NRF2 antibody. However, the site with
   the strongest binding was site 1, which was closest to the
   transcription start site ([177]Fig. 6K, [178]Fig. S24). Furthermore,
   neither NRF2 knockdown nor SFN treatment regulated the activity of the
   ALDH1L2 promoter containing a putative NRF2-mutated binding site 1
   ([179]Fig. 6L, [180]Fig. S25). Therefore, binding site 1 in the ALDH1L2
   promoter is critical for NRF2 binding. Furthermore, we detected a
   significant positive correlation between the expression of ALDH1L2 and
   that of NRF2 in HCC tissues ([181]Fig. 6M and N). Similar results were
   obtained in many cancer types, including HCC via analysis of TCGA data
   ([182]Fig. S26). Therefore, these results suggest that ALDH1L2/NRF2 can
   form a positive regulatory loop and promote HCC progression.

Fig. 6.

   [183]Fig. 6
   [184]Open in a new tab

   NRF2 binds to the ALDH1L2 promoter and increases ALDH1L2 expression.

   (A) ALDH1L2 expression in HCC cells treated with H[2]O[2] was assessed
   via WB. (B–E) ALDH1L2 expression in HCC cells treated with SFN (B,C)
   and in NRF2-knockdown HCC cells (D-E) was measured via qRT-PCR and WB.
   (F,G) STAT3-binding motif. (H–J) Huh7 and HEK293T cells were
   transfected with different luciferase reporter vectors. The
   corresponding luciferase activities were measured via reporter gene
   assays (H). The cells were treated with SFN (I) or transfected with
   NRF2 shRNA or the control (J). The corresponding luciferase activities
   were measured via reporter gene assays. (K) ChIP followed by qRT-PCR
   analysis was performed to evaluate NRF2 binding to the ALDH1L2
   promoter. (L) Huh7 and HEK293T cells were transfected with ALDH1L2
   luciferase reporter vectors (containing wild-type or mutant
   NRF2-binding sites, -585/-173). The cells were transduced with an NRF2
   shRNA vector or the corresponding control vector. The corresponding
   relative luciferase activities were determined via reporter gene
   assays. (M) ALDH1L2 and NRF2 expression in HCC tissues was measured via
   WB. (N) The correlation between ALDH1L2 and NRF2 expression in HCC
   tissues was analyzed (Spearman’s correlation). Data are represented as
   mean ± SD. ∗p <0.05; ∗∗p <0.01. Unpaired t-test or one-way ANOVA. ChIP,
   chromatin immunoprecipitation; HCC, hepatocellular carcinoma; NRF2,
   nuclear factor E2-related factor 2; qRT-PCR, quantitative
   real-time-PCR; SFN, sulforaphane; STAT3, signal transducer and
   activator of transcription 3; WB, western blotting.

Knockdown of ALDH1L2 sensitizes HCC cells to sorafenib treatment

   Sorafenib is a first-line standard therapeutic agent for the treatment
   of advanced HCC. Sorafenib is a multikinase inhibitor that targets
   Raf-1 and B-Raf and exhibits kinase activity toward proteins in the
   Ras/Raf/MEK/ERK signaling pathway, platelet-derived growth factor
   receptor, vascular endothelial growth factor receptor 2, hepatocyte
   factor receptor (c-KIT), and other proteins to inhibit tumor
   angiogenesis.[185]^32 However, acquired or intrinsic resistance of
   cancer cells to sorafenib remains a major obstacle to improving the
   prognosis of patients with HCC.[186]^32 Epigenetic factors, drug
   metabolism, dysregulation of apoptosis and characteristics of the tumor
   microenvironment are the main mechanisms involved in the initiation and
   development of sorafenib resistance in HCC.[187][33], [188][47]
   Therefore, we next investigated whether the knockdown of ALDH1L2
   sensitizes HCC cells to sorafenib. When ALDH1L2 expression was silenced
   by shRNA transduction, Li7 and SNU387 cells exhibited increased
   sorafenib sensitivity. The IC50 values of sorafenib in Li7/shALDH1L2-1
   cells (7.05 μM) and Li7/shALDH1L2-2 cells (7.41 μM) were lower than
   those in Li7/shNC (14.23 μM) cells. In addition, the IC50 values of
   sorafenib in SNU387/shALDH1L2-1 cells (8.21 μM) and Li7/shALDH1L2-2
   cells (8.58 μM) were lower than those in SNU387/shNC (13.65 μM) cells
   ([189]Fig. 7A). ALDH1L2 knockdown synergized with sorafenib to suppress
   the growth of HCC cells and promote ROS release and apoptosis
   ([190]Fig. 7B–F, [191]Fig. S27A–D). Conversely, ALDH1L2 overexpression
   significantly decreased the sorafenib sensitivity of Huh7 cells (from
   10.31 to 17.59 μM) and MHCC-97H cells (from 10.59 to 16.57 μM)
   ([192]Fig. 7G). ALDH1L2 overexpression effectively reversed the
   promoting effects of sorafenib on cell proliferation inhibition and
   apoptosis ([193]Fig. 7H and I).

Fig. 7.

   [194]Fig. 7
   [195]Open in a new tab

   Knockdown of ALDH1L2 sensitizes HCC cells to sorafenib.

   (A) The effect of ALDH1L2 knockdown on the IC50 of sorafenib in Li7 and
   SNU387 cells was evaluated. (B–F) ALDH1L2 knockdown and control cells
   were incubated with sorafenib (5 μM). Cell proliferation was evaluated
   using a CCK8 assay (B), a colony formation assay (C), and an EdU
   incorporation assay (D). ROS levels (E) and apoptosis (F) were
   evaluated by flow cytometry. (G) The effect of ALDH1L2 overexpression
   on the IC50 of sorafenib in Huh7 cells was evaluated. (H,I)
   ALDH1L2-overexpressing and control cells were incubated with sorafenib
   (5 μM). Cell proliferation was evaluated by an EdU incorporation assay
   (H). Apoptosis was evaluated by flow cytometry (I). (J–M) Li7 cells
   stably expressing shNC or shALDH1L2 were injected subcutaneously into
   the flanks of nude mice, and the mice were treated with or without
   sorafenib (10 mg/kg). Tumor growth was measured every 3 days (J).
   Representative images of excised tumors are shown (K). The tumors were
   weighed, and the weights were plotted (L). The expression of Ki67 in
   xenograft tissues was evaluated by IHC staining (M). (N) Model of the
   mechanisms of action of ALDH1L2 in HCC (figure created with
   [196]Biorender.com). Data are represented as mean ± SD. ∗p <0.05; ∗∗p
   <0.01. One-way ANOVA. CCK8, Cell Counting Kit-8; HCC, hepatocellular
   carcinoma; ROS, reactive oxygen species.

   Furthermore, ALDH1L2 knockdown increased the sensitivity of HCC cells
   to sorafenib in vivo in xenograft models. ALDH1L2 knockdown combined
   with sorafenib treatment reduced the overall tumor volume and mass
   ([197]Fig. 7J–L). In addition, Ki67 expression was more significantly
   decreased in the combination group than in the ALDH1L2-knockdown group
   or the group treated with sorafenib alone ([198]Fig. 7M). These results
   indicate that the knockdown of ALDH1L2 sensitizes HCC cells to the TKI
   sorafenib.

Discussion

   Modulation of oxidative stress is currently recognized as an effective
   approach for anticancer therapy. NRF2 is a basic leucine zipper
   transcription factor that plays a pivotal role in maintaining redox
   homeostasis.[199]^34 ROS production in response to oxidative stress is
   regulated by NRF2.[200]^35 A large body of evidence shows that NRF2
   activation in cancer cells promotes cancer progression and
   metastasis.[201][36], [202][37], [203][38] In this study, we found that
   ALDH1L2 suppresses ROS generation via NRF2 activity. Knockdown of
   ALDH1L2 inhibits NRF2 expression via the ubiquitin–proteasome pathway
   in HCC cells. In addition, we found that NRF2 directly binds to the
   ALDH1L2 promoter and induces ALDH1L2 expression in HCC cells. Taken
   together, our results reveal that ALDH1L2 and NRF2 are involved in a
   positive feedback loop that enables them to maintain cellular redox
   homeostasis and facilitate HCC cell proliferation and metastasis.

   Previous studies have shown that NRF2 binds to the IL-6 gene and
   activates its expression in hepatic cells.[204]^27 Elevated levels of
   IL-6 are observed in many solid tumors, including HCC.[205]^39 IL-6 is
   produced by multiple cell types within the tumor microenvironment,
   including tumor cells, tumor-infiltrating immune cells and stromal
   cells.[206]^40 IL-6 acts directly on tumor cells to induce STAT3
   expression, which promotes angiogenesis, metastasis, immunosuppression,
   and accelerates tumor progression.[207]^40 In this study, the plasma
   IL-6 level was found to be increased in ALDH1L2-overexpressing
   syngenetic Hepa1-6 model mice. Furthermore, we found that ALDH1L2
   activated the Jak2/STAT3 pathway in HCC cells. Pharmacological
   inhibition and knockdown of STAT3 attenuate ALDH1L2-mediated cell
   proliferation and invasion in HCC cells. In addition, previous studies
   have shown that the ability of STAT3 to promote IL-6 gene expression
   results in a feedforward autocrine feedback loop.[208]^41 Therefore,
   targeting the ALDH1L2/NRF2 axis is a promising therapeutic approach for
   treating HCC.

   Cancer cell-derived cytokines, such as IL-6, IL-1β, and IL-10,
   frequently direct differentiating tumor-infiltrating immune cells
   toward a tumor-promoting phenotype.[209]^42 Previous studies have shown
   that IL-1β promotes TAM and MDSC infiltration to promote HCC
   metastasis.[210]^43 Increased infiltration of TAMs has been associated
   with poor prognosis in HCC.[211]^43^,[212]^44 TAMs expressing IL-10
   have been found to facilitate immune evasion.[213]^45 IL-10 is involved
   in STAT3 activation, which can contribute to tumor formation. In this
   study, we found that ALDH1L2 stimulates IL-6 and IL-10 secretion from
   HCC cells. NF-κB is considered a central mediator of immune and
   inflammatory responses.[214]^46 Our results show that an increase in
   the tumor expression level of ALDH1L2, which activates NF-κB signaling
   in TAMs to generate an inflammatory microenvironment, promotes
   malignant progression. Furthermore, increased CD206 and CD163
   expression was observed in macrophages cocultured with
   ALDH1L2-overexpressing HCC cells. In addition, CD206-positive M2
   macrophages were more abundant in tumors derived from
   ALDH1L2-overexpressing cells but less abundant in tumors derived from
   shALDH1L2-transduced cells than in tumors derived from the
   corresponding control cells. These results indicate that ALDH1L2
   promotes the M2 polarization of TAMs, which accelerates HCC
   progression.

   In conclusion, we demonstrated that NRF2-induced ALDH1L2 overexpression
   promotes HCC progression through an IL-6/Jak/STAT3 positive feedback
   loop, which results in persistent crosstalk between HCC cells and TAMs
   and accelerates HCC progression ([215]Fig. 7N). Interference with this
   oncogenic loop may have great potential therapeutic application in HCC.

Abbreviations

   AREs, antioxidant response elements; CCK8, Cell Counting Kit-8; ChIP,
   chromatin immunoprecipitation; DFS, disease-free survival; GEO, Gene
   Expression Omnibus; HCC, hepatocellular carcinoma; IHC,
   immunohistochemistry; KEAP1, Kelch-like ECH-associated protein 1; KEGG,
   Kyoto Encyclopedia of Genes and Genome; MDSC, myeloid-derived
   suppressor cells; MMPs, matrix metalloproteinases; NF-κB, nuclear
   factor-kappa B; NRF2, nuclear factor E2-related factor 2; OCR, oxygen
   consumption rate; OS, overall survival; PHGDH, phosphoglycerate
   dehydrogenase; qRT-PCR, quantitative real-time-PCR; ROS, reactive
   oxygen species; SFN, sulforaphane; STAT3, signal transducer and
   activator of transcription 3; TAMs, tumor-associated macrophages; TCGA,
   The Cancer Genome Atlas; TKI, tyrosine kinase inhibitor; Treg cell,
   regulatory T cell; WB, Western blotting; ICGC-LIRI-JP, International
   Cancer Genome Consortium liver cancer-RIKEN Japan.

Financial support

   This work was supported by grants from the National Natural Science
   Foundation of China (82273102, 81972581, 82103201, 82300165, 82072892),
   Shanghai Sailing Program (21YF1445400), State Key Laboratory of Systems
   Medicine for Cancer Research Foundation (ZZ-94-2312), Guangxi Natural
   Science Foundation (2024GXNSFAA010061, 2023GXNSFAA026276), the Key
   Discipline Construction Project of Jiading District Health System (No.
   XK202405), Guangxi Training Program for Medical High-level Academic
   Leaders (No. Guiweikejiaofa [2020]-15), Bose Talent Highland (No.
   2020-3-2), Building Projects of Guangxi Bagui Scholars (No. 2024),
   Building Projects from the Key Laboratory of Molecular Pathology (for
   Hepatobiliary Diseases) of Guangxi (No. Guiweikejiaofa [2020]-17).

Authors’ contributions

   Designed and supervised research, designed and performed experiments,
   analyzed data, and wrote the manuscript: HT. Performed experiments:
   JJL, CZ, QQZ, JYC, QQL, LLM, WT, YY, CG, YTS. Provided technical
   assistance: JW, XDL.

Data availability statement

   All data are available from the corresponding authors upon reasonable
   request. RNA-seq data of tumor and non-tumor tissues were obtained from
   The Cancer Genome Atlas (TCGA; [216]http://cancergenome.nih.gov), GEO
   ([217]GSE22058, [218]GSE67680), and ICGC-LIRI-JP
   ([219]https://dcc.icgc.org/projects/LIRI-JP).

Conflicts of interest

   The authors declare no conflict of interest.

   Please refer to the accompanying ICMJE disclosure forms for further
   details.

Footnotes

   Author names in bold designate shared co-first authorship

   Supplementary data to this article can be found online at
   [220]https://doi.org/10.1016/j.jhepr.2024.101217.

Contributor Information

   Xi-Dai Long, Email: sjtulongxd@263.net.

   Jun Wu, Email: jun.wu@shsmu.edu.cn.

   Hua Tian, Email: htian@shsci.org.

Supplementary data

   The following are the supplementary data to this article:
   Multimedia component 1
   [221]mmc1.pdf^ (3.7MB, pdf)
   Multimedia component 2
   [222]mmc2.docx^ (57.5KB, docx)
   Multimedia component 3
   [223]mmc3.pdf^ (618.1KB, pdf)
   Multimedia component 4
   [224]mmc4.pdf^ (7.2MB, pdf)

References