GATA2 deficiency exacerbates chronic liver injury via disrupting hepatocyte death-regeneration balance: Clinical, histopathological, and molecular evidence

Abstract

BACKGROUND

GATA-binding protein 2 (GATA2) is a critical transcription factor that plays an essential role in maintaining the stemness of hematopoietic stem cells. Mutations in GATA2 are recognized as disease-causing mutations for aplastic anemia (AA) and myelodysplastic syndromes; however, the mechanisms linking these mutations to chronic liver injury (CLI) (e.g., cirrhosis) remain elusive.

AIM

To investigate the molecular mechanisms through which GATA2 deficiency promotes CLI.

METHODS

Whole genome sequencing was performed to identify loss-of-function mutations in GATA2 in a 55-year-old female patient who was suspected of and later diagnosed with AA and developed cirrhosis one year post-diagnosis. After establishing the genetic association between GATA2 mutations and cirrhosis through clinical case analysis, liver injury murine and hepatocyte models were developed to characterize GATA2 expression in response to liver injury. rAAV8-TBG-Gata2-shRNA-mediated liver-specific Gata2 knockdown was employed to elucidate the role of GATA2 in exacerbating liver injury.

RESULTS

The patient presented with a three-year medical history of dizziness, fatigue, and thrombocytopenia. Following conventional treatment, she developed severe fatigue, abdominal distension, and jaundice, indicative of liver disease. Multimodal evaluations, including bone marrow biopsy, flow cytometry, liver biopsy, and genetic testing, confirmed GATA2 mutation-associated AA complicated by cirrhosis. Experimental studies in cellular and animal models demonstrated compensatory upregulation of hepatocyte GATA2 in response to endoplasmic reticulum stress and oxidative stress. Mechanistically, liver-specific Gata2 knockdown exacerbated hepatocyte apoptosis, necroptosis, ferroptosis, and impaired regeneration. Furthermore, Gata2 knockdown suppressed the adaptive unfolded protein response, attenuated the antioxidant response, and inhibited fatty acid β-oxidation.

CONCLUSION

GATA2 deficiency drives the CLI by disrupting the hepatocyte death-regeneration balance.

Key Words: GATA2 deficiency; Aplastic anemia; Chronic liver injury; Hepatocyte death-regeneration balance; Unfolded protein response; Antioxidant response

Core Tip: GATA-binding protein 2 (GATA2) is known to play a crucial role in maintaining the stemness of hematopoietic stem cells, but its involvement in chronic liver injury (CLI), such as cirrhosis, remains unclear. Through an investigation into clinical cases and experimental models, this study demonstrates that GATA2 deficiency drives the progression of CLI by disrupting the hepatocyte death-regeneration balance. The findings suggest that GATA2 mutations contribute not only to aplastic anemia pathogenesis but also act as a causative factor in CLI, highlighting its potential as a therapeutic target for managing liver complications in aplastic anemia patients.

INTRODUCTION

Hematological diseases are closely linked to liver dysfunction[1,2]. Studies indicate that hematological diseases can lead to liver function abnormalities through mechanisms such as malignant cell infiltration, hemodynamic alterations, drug toxicity, and immune-mediated injury[3,4]. Aplastic anemia (AA) and myelodysplastic syndromes (MDS) are both associated with chronic liver injury (CLI), yet whether AA or MDS could share a common pathogenesis with chronic liver disease remains to be elucidated.

Clinicians often find it challenging to differentiate between AA and MDS. AA is induced by various factors that exhaust bone marrow hematopoietic function, leading to anemia, bleeding, and increased susceptibility to infection. Although the pathogenesis of AA is complex and remains incompletely understood, recent research indicates a genetic predisposition[5]. In patients with AA, peripheral blood and bone marrow exhibit an elevated proportion of lymphocytes and an imbalance in T cell subsets, such as the proportion of T helper cells type I and CD8+T suppressor cells. The increase in interleukin 2, interferon-gamma, tumor necrosis factor-alpha, and myeloid cell apoptosis caused the negative regulation of T cell secretion, for which immunosuppressive therapy is effective in most cases[6]. The most common cause of inhibited T helper lymphocyte activity is immune-mediated injury to hematopoietic stem cells. This injury suppresses bone marrow hematopoiesis, while immune stimulation boosts growth factor synthesis, thereby accelerating hematopoietic stem cell proliferation[7].

MDS, a group of acquired clonal diseases of hematopoietic stem cells characterized by severe hematopoietic dysfunction[8], shares similarities with AA, including bone marrow involvement, peripheral blood involvement, and clinical manifestations. The initial symptoms of MDS are nonspecific and primarily include anemia and one or two lines of hemocytopenia. Common incipient symptoms of MDS, such as anemia, insufficient bone marrow cells, dysplasia of erythrocytes and megakaryocyte lineage, and karyotype abnormalities, are observed in most patients[9,10]. Blood analysis complicates the distinction between AA and MDS. However, myelodysplasia is present in MDS but diminished in AA, and these conditions require distinct management strategies and have different prognoses[11,12].

Both AA and MDS patients frequently exhibit liver damage[13-15]. The hepatic injury in these diseases may involve multiple mechanisms: (1) Malignant cell infiltration: In MDS, blasts invade hepatic sinusoids, activating hepatic stellate cells and driving fibrosis; (2) Treatment-related hepatotoxicity: Immunosuppressants can induce sinusoidal obstruction, while iron overload exacerbates oxidative stress[16,17]; (3) Immune-inflammatory cascade: T helper cells type I cell polarization releases interferon-gamma/tumor necrosis factor-alpha, disrupting hepatocyte tight junctions and activating Kupffer cells, which amplifies systemic inflammation; and (4) Portal hemodynamic disturbances: Extramedullary hematopoiesis increases portal pressure, resulting in sinusoidal endothelial injury. Iron overload plays a pivotal role in liver injury[18,19], generating excessive reactive oxygen species (ROS) via the Fenton reaction and triggering oxidative stress. ROS induces oxidative stress, damaging cell membranes, proteins, and DNA, ultimately leading to dysfunction and apoptosis[20]. Additionally, ROS also activates hepatic immune cells (e.g., Kupffer cells), promoting pro-inflammatory cytokine release that exacerbates hepatic inflammation and fibrogenesis[21], potentially progressing to cirrhosis.

Genetic determinants play a crucial role in the pathogenesis of hematological disorders and CLI, with many conditions showing a strong hereditary predisposition. Notable examples include hereditary hemochromatosis and alpha-1 antitrypsin deficiency, which have well-established causal genes and pathogenic mechanisms. GATA-binding protein 2 (GATA2) gene mutations represent a rare genetic subtype associated with both AA and MDS. As a key regulator of hematopoietic stem/progenitor cells (HSPCs)[22,23], GATA2 controls the transcription of downstream target genes that govern HSPC proliferation and differentiation[24,25]. GATA2 mutations cause GATA2 deficiency syndrome, with its hematopoietic manifestations primarily including hematopoietic failure disorders such as AA and MDS and a subset of patients progressing to acute myeloid leukemia (AML)[23,26,27]. These mutations impair HSPC self-renewal capacity, induce differentiation blockade, and promote apoptosis, ultimately leading to bone marrow failure. In addition, these mutations may hinder liver regeneration[22,28]. GATA2 upregulates cell survival-related genes, such as those in the BCL-2 family, and its mutation can increase pro-apoptotic factors like BAX, exacerbating hepatocyte death[23,29]. Although GATA2 is expressed in fetal liver hematopoietic stem cells and progenitor cells[30], the direct role of GATA2 deficiency in the chronic progression of liver injury has not been investigated.

We conducted a multidimensional study on a clinical case of AA linked to a GATA2 missense mutation, which rapidly progressed to cirrhosis post-diagnosis. This unique clinical observation led us to hypothesize that GATA2 deficiency might contribute to CLI. The aim was to elucidate the molecular mechanisms underlying the genetic association between GATA2 mutations and cirrhosis development by constructing experimental liver injury models in mice and hepatocytes. The findings from this study may enhance the understanding of the pathogenic mechanisms of liver injury in AA patients with GATA2 deficiency and potentially offer therapeutic targets to improve patient outcomes.

MATERIALS AND METHODS

Patient

A 55-year-old woman was admitted to the Department of Hematology at the Affiliated Hospital of Zunyi Medical University (Zunyi, Guizhou Province, China) due to dizziness, fatigue, and yellow urine. She had experienced dizziness and fatigue for three years ago, accompanied by bilateral knee ecchymosis. However, she reported no cough, fever, or shivering. After necessary tests, “AA” was suspected. Symptomatic treatment was ineffective, and she later received traditional Chinese medicine (TCM) for complications. However, her symptoms worsened over the next two months, manifesting as chills, fever, chest tightness, shortness of breath, hematemesis, black feces, yellow urine, yellow eyes, nausea, vomiting, and intermittent gingival bleeding. TCM was discontinued after two months. Total bilirubin (TBil) peaked at 302.7 μmol/L, while direct bilirubin was at 156.1 μmol/L. The patient was then hospitalized for drug-induced liver injury and further evaluation. Her medical history indicated previous erythrocyte transfusions without adverse reactions and no other notable issues. She reported no smoking, drinking, or other harmful habits in her personal or family history.

Physical examination revealed stable vital signs within normal limits. Jaundice was observed in the skin and sclera, with no palmar erythema, spider nevus, petechiae, or ecchymosis. There was no visible enlargement of superficial systemic lymph nodes. The jugular vein was not distended, with no signs of hepatic jugular reflux. The cardiopulmonary examination revealed no apparent abnormalities. The abdomen was flat and non-tender, with no palpable hepatosplenomegaly, lethargy, or loss of mobility.

Chromatin immunoprecipitation sequencing and real-time quantitative polymerase chain reaction

Chromatin immunoprecipitation sequencing (ChIP-seq) was performed at Wuhan Seq Health Tech Co. (Wuhan, Hubei Province, China) using 1 × 10⁷ cells per sample. In brief, cells were washed with phosphate-buffered saline (PBS) and cross-linked with 1% formaldehyde for 10 minutes at room temperature, followed by quenching with 125 mmol/L glycine for 5 minutes. Cells were harvested and lysed on ice for 5 minutes using a cell lysis buffer containing 10 mmol/L HEPES (pH = 7.5), 0.1 mmol/L EDTA, 0.5% NP-40, and a protease inhibitor cocktail. Nuclei were collected by centrifugation at 2000 × g for 10 minutes at 4 °C, and chromatin was sheared to an average size of 100-500 base pairs (bp) using sonication. Of the sonicated chromatin, 10% was reserved and labeled as “input”, whereas 80% was used for immunoprecipitation with a specific anti-GATA2 antibody (Abcam, Cat# ab109241, Cambridge, MA, United States). An additional 10% was incubated with rabbit immunoglobulin G (IgG) (Cell Signaling Technology, Cat# 2729, Danvers, MA, United States) as a negative control and labeled as “IgG”. Washing involved low salt buffer (containing 20 mmol/L Tris-HCl pH 8.0, 2 mmol/L EDTA pH 8.0, 150 mmol/L NaCl, 0.1% SDS, 1% Triton X-100 in sterile water) and high salt buffer (with 20 mmol/L Tris-HCl pH 8.0, 2 mmol/L EDTA pH 8.0, 500 mmol/L NaCl, 0.1% SDS). DNA from both input and IP fractions was extracted via the phenol-chloroform method. The degree of chromosomal fragmentation was assessed by detecting input DNA via agarose gel electrophoresis. The success of the immunoprecipitation was further evaluated by western blot analysis. High-throughput DNA sequencing libraries were prepared using the VAHTS Universal DNA Library Prep Kit for MGI (Cat# NDM607, Vazyme, China). Library products corresponding to fragments ranging from 200 bp to 500 bp were enriched and quantified. Sequencing was conducted on a DNBSEQ-T7 (MGI; Vazyme, China) using the PE150 sequencing mode to generate paired-end reads.

To validate the GATA2 binding sites identified by ChIP-seq, ChIP-real-time quantitative polymerase chain reaction (qPCR) was performed using DNA from Input, IP, and IgG samples. Primers targeting the ChIP-seq peak regions were designed to amplify 100-150 bp amplicons using Primer-BLAST (Table 1). qPCR was carried out in triplicate using EnTurbo™ SYBR Green SuperMix (Cat# EQ001, Biotechnology, MA, United States) on a QuantStudio 6 Flex System, following these cycling conditions: 95 °C for 3 minutes, 40 cycles of 95 °C for 10 seconds, 58 °C for 30 seconds, and dissociation curve analysis. Enrichment was calculated using the double ΔCT method, normalizing IP signals against Input and IgG controls. Only results that met the quality criteria (amplification efficiency 90%-110%, R² > 0.99) were accepted.

Table 1 Primers used in this study.

Name	Sequences (5’-3’)		Tm	CG (%)	Product length (bp)
HSPD1	Sense	AGGGTTCAAATGATTCTCCTGC	60.0	45.5	143
	Antisense	CAAGGTCAGGAGTTCAAGACCAG	60.7	52.2
PPARA	Sense	TGAACTCCTGACCTCAAGTGATC	58.3	47.8	114
	Antisense	CCAGGTACACAGAAAGCATTAACT	57.8	41.7
GPX4	Sense	GAGCAGTACCAGCCACCGTT	59.9	60.0	176
	Antisense	AGACCTCTCAACCCAGTGTCTG	58.1	54.5

GPX4: Glutathione peroxidase 4; HSPD1: Heat shock protein family D member 1; PPARA: Peroxisome proliferator-activated receptor alpha.

Establishment of the CLI mouse model

All procedures involving experimental animals strictly adhered to laboratory animal welfare ethics guidelines and were approved by the Laboratory Animal Ethics Committee of Zunyi Medical University (Approval No. ZMU22-2303-029). Male BALB/c mice (5-6 weeks old, body weight 21.4 ± 1.7 g) were provided by the Experimental Animal Center of Zunyi Medical University (Guizhou Province, China) and were used to establish the CLI mouse model and subsequent in vivo studies. All mice were housed in a specific pathogen-free-level barrier facility at 23 ± 2 °C and 40%-60% humidity, with a 12-hour light/dark cycle, and animals had ad libitum access to autoclaved feed and acidified drinking water. After a 7-day acclimation, mice were randomly assigned to experimental groups using a random number table (12 mice per group).

CLI was induced in male BALB/c mice via intraperitoneal injection of carbon tetrachloride (CCl₄) prepared as a 20% stock solution in olive oil. Mice were randomly divided into two groups (n = 12/group): Solvent control group (olive oil, 5.0 mL/kg) and CLI model group (20% CCl₄-olive oil solution, 5.0 mL/kg), with treatments administered bi-weekly for eight consecutive weeks. CCl₄ was administered to the mice in a fume hood, ensuring that safety standards were upheld. Prior to each injection, mice were fasted for 6 hours for accurate dosing. At 24 hours post-final injection, blood samples were collected under anesthesia, followed by euthanasia and liver dissection. Liver tissues were fixed in a 4% paraformaldehyde solution for pathological examination, and snap-frozen at -80 °C for molecular analyses (e.g., western blot). Successful CLI modeling was confirmed through serum biochemistry and hepatic histopathology.

Hepatocyte-specific Gata2 knockdown in mice

Recombinant adeno-associated virus serotype 8 (rAAV8; Genechem, China) carrying a hepatocyte-specific thyroxine-binding globulin promoter was used to deliver Gata2-targeting short hairpin RNA (shRNA) or control shRNA (sequences listed in Table 2). BALB/c mice received tail vein injections of the viral constructions at a dose of 5 × 10¹⁰ to 1 × 10¹¹ vector genomes per mouse. For Gata2 knockdown validation, twenty-four mice were randomly divided into two groups (n = 12 each group): Control group (rAAV8-control shRNA with 4-week transfection) and Gata2 knockdown group (Gata2-KD) (rAAV8-Gata2 shRNA with 4-week transfection). Successful knockdown was confirmed by western blot analysis.

Table 2 Short hairpin RNA sequences used for in vivo knockdown of Gata2.

	Sequences (5’ to 3’)
Gata2 target	GAATCGGAAGATGTCCAGCAA
Gata2 shRNA	GAATCGGAAGATGTCCAGCAACGAATTGCTGGACATCTTCCGATTC
Control shRNA	AAACGTGACACGTTCGGAGAACGAATTCTCCGAACGTGTCACGTTT

shRNA: Short hairpin RNA.

To investigate the role of hepatic GATA2 in CLI, an additional 24 CLI mice were randomly assigned to two groups (n = 12 per group): CLI control group (CCl₄) (rAAV8-control shRNA with 4-week pre-transfection followed by 8-week CCl₄ administration) and Gata2 knockdown CLI group (Gata2-KD + CCl₄) (rAAV8-Gata2 shRNA with 4-week pre-transfection followed by 8-week CCl₄ administration). At the endpoint, blood samples were collected under anesthesia, followed by euthanasia and liver dissection. Liver tissues were processed: 0.5 cm³ fragments fixed in 4% paraformaldehyde solution for histopathology, 1 mm³ blocks fixed in 3% glutaraldehyde for transmission electron microscope (TEM), and remaining tissues snap-frozen at -80 °C for western blot analysis.

Determination of serum biochemical parameters

Peripheral blood samples from mice were subjected to coagulation-promoting treatment, followed by centrifugation at 4 °C and 3000 × g for 10 minutes to obtain serum. Serum alanine aminotransferase (ALT) and TBil levels were measured using a Beckman Coulter automatic biochemical analyzer (AU5800, United States), with duplicate wells for all tests. ALT activity was quantified via the rate method (enzyme-coupled reaction system, primary wavelength 340 nm, secondary wavelength 405 nm), whereas TBil concentration was determined by the diazonium salt colorimetric method (diazotized sulfanilic acid method, detection wavelength 546 nm). Experimental procedures were strictly adhered to the manufacturer’s protocols.

Histopathological analysis of liver tissue

Liver tissue specimens were fixed in 4% paraformaldehyde and subsequently dehydrated and embedded in paraffin. Serial sections, each 4 μm thick, underwent an xylene gradient dewaxing process, followed by gradient ethanol hydration. Routine hematoxylin and eosin staining was performed as described briefly below: Nuclei were stained with hematoxylin for 5 minutes, differentiated using 1% hydrochloric acid-ethanol, and cytoplasm was counterstained with eosin for 30 seconds. For Masson staining, sections were treated sequentially with Weigert’s iron hematoxylin for nuclear staining (5 minutes), acid fuchsin-ponceau mixture to stain muscle fibers, fibrin, and erythrocytes (10 minutes), followed by differentiation with 1% phosphomolybdic acid, and stained with 2% aniline blue to visualize collagen fibers (5 minutes), and final treatment with 1% glacial acetic acid for 1 minute. All stained sections were dehydrated through a gradient ethanol series, cleared in xylene, and mounted with neutral gum. Whole-slide digital scanning was conducted using a 3DHISTECH Pannoramic SCAN system (Budapest, Hungary). Images were analyzed using CaseViewer v2.4 software (3DHISTECH, Hungary) at magnifications ranging from × 1 to × 630, with target areas specifically observed at magnifications of × 100 to × 400.

Detection of hepatocyte apoptosis

Hepatocyte apoptosis was assessed in liver tissue sections using the Terminal deoxynucleotidyl transferase-mediated dUTP nick-end labeling (TUNEL) assay (Roche TUNEL Apoptosis Detection Kit, Cat#11684817910, Roche Diagnostics GmbH, Darmstadt, Germany) following the manufacturer’s protocol. Paraffin-embedded sections were dewaxed with xylene gradients and rehydrated through sequential ethanol gradients (100% to 70%). Tissue sections were treated with antigen retrieval (20 μg/mL proteinase K) at 37 °C for 15 minutes. For TUNEL labeling, tissue sections were incubated with the TdT enzyme-dUTP reaction mixture (prepared at a 2:29 v/v ratio) at 37 °C for 2 hours in a humidified chamber (protected from light). After PBS rinsing, nuclei were counterstained with 4’,6-diamidino-2-phenylindole (10 minutes, room temperature, dark) and then mounted with an anti-fade medium. Fluorescent images were acquired using a Nikon inverted fluorescence microscope (ECLIPSE TI-SR, Japan) equipped with the following filter sets: 4’,6-diamidino-2-phenylindole channel (excitation 330-380 nm/emission 420 nm, blue nuclear staining) and FITC channel (excitation 465-495 nm/emission 515-555 nm, green apoptotic signals). For each section, six non-overlapping fields were randomly captured at a 400 × magnification. The apoptotic index was calculated as: Apoptotic index (%) = (number of TUNEL-positive cells/total number of cells) × 100%.

Analysis of liver ultrastructure

Liver ultrastructure was analyzed by TEM. In brief, fresh liver tissue samples (≤ 1 mm³) were immediately fixed in ice-cold 3% glutaraldehyde-0.1 M phosphate buffer (pH = 7.4), rinsed with PBS, and post-fixed with 1% osmium tetroxide (OsO₄) for 2 hours. Dehydration was conducted through a graded ethanol series (50%, 70%, 80%, 90%, 95%, 100%) and acetone, with subsequent infiltration and embedding using epoxy resin 812 (Epon 812, SPI Supplies, PA, United States; 37 °C/12 hours → 45 °C/12 hours → 60 °C/48 hours). Ultrathin sections (70 nm) were cut using a Leica UC7 ultramicrotome (Germany), double-stained with 2% uranyl acetate (15 minutes, dark) and Reynolds’ lead citrate (15 minutes CO₂-free), and imaged on a JEOL JEM-1400 TEM (80 kV accelerating voltage, 0.38 nm point resolution, JEOL Ltd., Tokyo) equipped with a Gatan Orius SC1000 CCD system (7.4 μm pixel size). Ultrastructure evaluation focused on nuclear membrane integrity, mitochondrial cristae morphology, rough endoplasmic reticulum (ER) dilation, and lipid droplet distribution. Four random, non-overlapping fields per sample were analyzed (scale bars: Nanometer-calibrated).

Quantification of hepatic malondialdehyde levels

Malondialdehyde (MDA) levels in liver tissues or hepatocytes were quantified using thiobarbituric acid colorimetry (MDA Assay Kit; Beijing BoxBio Technology Co., Ltd., China; Cat No. AKFA013C) according to the manufacturer’s instructions. Liver tissue sample (100 mg) was homogenized in 1 mL ice-cold MDA-specific extraction buffer (1:10 w/v) using a pre-chilled glass homogenizer, centrifuged at 8000 × g for 10 minutes at 4 °C, and the supernatant was collected for analysis. Then, the supernatant was mixed with the detection working solution (containing 0.67% thiobarbituric acid, 0.1 mmol/L EDTA, and an acetate buffer system with pH 3.5). After a 60-minute reaction at 95 °C, the resulting solution was rapidly cooled on ice, centrifuged at 10000 × g for 10 minutes at 4 °C, and the supernatant (0.2 mL per sample) was transferred to a 96-well plate. Absorbance values were measured using a BioTek microplate reader (Epoch, CA, United States) in a three-wavelength detection mode (primary wavelength: 532 nm; reference wavelengths: 450 nm and 600 nm). MDA levels were calculated using the formula: MDA (nmol/g) = 5 × [6.45 × (ΔA532 - ΔA600) - 1.29 × ΔA450]/W, where W represents the sample weight in grams (g).

Determination of hepatic glutathione levels

The levels of reduced glutathione (GSH) and oxidized GSH (GSSG) in mouse liver tissues or hepatocytes were determined using 5,5’-dithiobis-(2-nitrobenzoic acid) (DTNB) colorimetric method with the GSH assay kit (AKPR008M) and GSSG assay kit (AKPR009M) from Beijing Boxbio Science & Technology Co., Ltd. All procedures were performed under strict antioxidant conditions to prevent artificial oxidation. Briefly, 100 mg of liver tissue was homogenized in ice-cold antioxidant-preserved PBS (0.1 M, pH 7.4, containing 5 mmol/L EDTA and 0.1% N-ethylmaleimide) and then centrifuged at 12000 × g for 15 minutes at 4 °C. The supernatant was collected for analysis following the manufacturer’s protocols. GSH quantification relied on the DTNB reaction, which produces a yellow-colored product measured at 412 nm. For GSSG detection, GSH interference was eliminated by pretreatment with 2-vinylpyridine prior to the DTNB chromogenic reaction. Absorbance values were recorded using a BioTek Epoch microplate reader, and GSH and GSSG concentrations were calculated based on standard calibration curves.

Examination of the enzyme activity of mitochondrial electron transfer chain complex I

The activities of electron transfer chain complex I (ETC-CI) (NADH-ubiquinone oxidoreductase) in liver tissue or hepatocytes mitochondria were determined using detection kits (ADS-W-FM006-48, Adisson Biotechnology Co., Ltd., China). Fresh liver tissue (100 mg) was homogenized on ice in 1 mL of pre-cooled mitochondrial isolation buffer (225 mmol/L mannitol, 75 mmol/L sucrose, 1 mmol/L EDTA, 10 mmol/L HEPES, pH 7.4). The homogenate was then subjected to stepwise centrifugation at 4 °C (700 × g for 10 minutes to remove nuclei and cell debris, subsequently 12000 × g for 10 minutes to pellet mitochondria). Complex I activity was measured in a 200 μL reaction mixture containing 50 μg mitochondrial protein, 200 μmol/L NADH, and 100 μmol/L ubiquinone analog (decylubiquinone). Absorbance measurements were conducted at 340 nm using a BioTek Epoch microplate reader (PA, United States). All procedures, including sample preparation, detection, absorbance reading, and enzyme activity calculations, were performed according to the respective kit manuals.

Western blot analysis

Western blot analysis was performed to examine protein expression levels. Briefly, liver tissues or hepatocytes were homogenized in pre-cooled RIPA lysis buffer (containing 50 mmol/L Tris-HCl pH 7.4, 150 mmol/L NaCl, 1% NP-40, 0.5% sodium deoxycholate, 0.1% SDS, 1 mmol/L EDTA, 1 mmol/L PMSF, and 1 × phosphatase inhibitor) (R0010; Beijing Solarbio Science & Technology Co., Ltd.), followed by ice-bath ultrasonication. The lysates were centrifuged at 12000 × g for 15 minutes at 4 °C, and supernatants were collected. Protein concentration was quantified using the BCA assay, with 40 μg of total protein mixed with 4 × loading buffer (containing 2% SDS, 10% glycerol, and 5% β-mercaptoethanol) at a 3:1 volume ratio and denatured at 95 °C for 5 minutes. Proteins were separated on 12% SDS-PAGE gels (70 V for 30 minutes in stacking gel, 100-120 V for 90 minutes in separating gel) and transferred to PVDF membranes (0.45 μm for > 30 kDa; 0.22 μm for < 30 kDa; Millipore IPFL00010) using a wet transfer method (Bio-Rad system). After blocking with 5% skim milk in TBST for 1 hour, membranes were incubated overnight at 4 °C with primary antibodies (Table 3) in TBST containing 5% BSA, followed by 1.5 hours at 4 °C temperature incubation with HRP-conjugated secondary antibodies (anti-mouse IgG, sc-516102, Santa Cruz, TX, United States, 1:10000; anti-rabbit IgG, sc-2357, Santa Cruz, TX, United States, 1:10000). Signals were visualized using ECL chemiluminescence reagent (MA0186, Meilunbio, Dalian, China) and captured by a multifunctional gel imaging system (Shanghai Qinxiang Scientific Instrument Co., Ltd.). Gray value analysis was conducted with Image J software, normalizing target protein expression levels to GAPDH and presenting as relative values. All experiments were repeated three times, with quantitative data expressed as mean ± SD.

Table 3 Primary antibodies used in this study.

Antibody	Source	Lot number	Manufacturer	Molecular weight (kDa)
Cleaved caspase-3	Rabbit mAb	9664	Cell Signaling Technology, Danvers, MA, United States	17
GAPDH	Mouse mAb	MAB45855	Bioswamp, Wuhan, China	36
GATA2	Mouse mAb	sc-267	Santa Cruz, TX, United States	50
GRP78	Rabbit mAb	ab108615	Abcam, Cambridge, MA, United States	78
GPX4	Mouse mAb	sc-166437	Santa Cruz, TX, United States	21
UCP2	Mouse mAb	sc-390189	Santa Cruz, TX, United States	70
HSP60	Mouse mAb	sc-376240	Santa Cruz, TX, United States	60
PCNA	Mouse mAb	sc-25280	Santa Cruz, TX, United States	36
MCAD	Mouse mAb	sc-365109	Santa Cruz, TX, United States	45
MLKL	Rabbit pAb	PA5-34733	Thermo Fisher Scientific, Waltham, MA, United States	54
PPARα	Mouse mAb	MA5-37652	Thermo Fisher Scientific, Waltham, MA, United States	55
p-MLKL	Rabbit mAb	37333S	Cell Signaling Technology, Danvers, MA, United States	54

GATA2: GATA binding protein 2; GRP78: Glucose-regulated protein 78; GPX4: Glutathione peroxidase 4; UCP2: Uncoupling protein 2; HSP60: Heat shock protein 60; PCNA: Proliferating cell nuclear antigen; MCAD: Medium chain acyl-CoA dehydrogenase; MLKL: Mixed lineage kinase domain-like protein; PPARα: Peroxisome proliferator-activated receptor alpha; p-MLKL: Phosphorylated mixed lineage kinase domain-like protein; mAb: Monoclonal antibody; pAb: Polyclonal antibody.

Determination of hepatic triglyceride levels

Liver triglyceride (TG) content was measured using a TG assay kit (AKFA003M, Boxbio, China) based on the glycerol phosphate oxidase-peroxidase-aminoantipyrine phenol method, following the manufacturer’s instructions. Briefly, 100 mg of liver tissue was homogenized in 1 mL of pre-chilled isopropanol-n-heptane TG extraction solution (1:10 w/v). The homogenate was immediately boiled for 5 minutes to inactivate enzymes, then centrifuged at 8000 × g for 10 minutes at 4 °C, and the supernatant was collected. Free glycerol was removed from the supernatant using magnesium silicate adsorbent. The treated supernatant was mixed with the assay working solution and incubated at 65 °C for 15 minutes, then cooled to room temperature. Aliquots of 0.2 mL were transferred into a 96-well plate, and optical density (OD) values were measured at 500 nm using a BioTek Epoch microplate reader (Winooski, VT, United States). Liver TG content was calculated from a standard curve and expressed as mg/g protein.

Cell cultures and experiments

Human hepatocellular carcinoma HepG2 cells and mouse hepatocyte AML-12 cells were purchased from the American Type Culture Collection and authenticated via short tandem repeat profiling. HepG2 cells were cultured in Dulbecco’s Modified Eagle Medium supplemented with 10% fetal bovine serum and 1% penicillin-streptomycin. The experiments involving human liver cells were approved by the Medical Ethics Committee of Zunyi Medical University, protocol No. ZMLS[2023] 1-172. AML-12 cells were maintained in Dulbecco’s Modified Eagle Medium/F12 medium containing 10% fetal bovine serum, 0.5% insulin-transferrin-selenium (PB180429, Proncell, Wuhan, Hubei, China), 40 ng/mL dexamethasone, and 1% penicillin/streptomycin.

Thapsigargin (THA; Sigma, St. Louis, MO, United States) was used to induce ER stress. Briefly, AML-12 cells were treated with THA (1.0 μmol/L) for 48 hours in the THA group or with dimethyl sulfoxide in the vehicle control group. To induce oxidative stress, both AML-12 and HepG2 cells were exposed to hydrogen peroxide (H₂O₂; 150.0 μmol/L) for 48 hours in the H₂O₂ group or to an equivalent volume of PBS in the solvent control group.

A HepG2 cell model with GATA2 overexpression was established using a plasmid-based system. The open reading frame of human GATA2 (NM_001145661) was cloned into the pCMV-MCS-EF1a-ZsGreen-SV40-Neomycin vector (Lifecore Biotechnology, Suzhou, China). An empty control vector was used as a negative control. HepG2 cells were transfected with the GATA2 construct. Following transfection, cells were subjected to selection with neomycin to generate a stable polyclonal population. Successful overexpression of GATA2 was validated via western blot analysis. Both parental HepG2 cells and the resulting GATA2-overexpressing cells were treated with 1.0 μmol/L THA for 48 hours.

Cell viability assay

Hepatocyte viability was examined using the Cell Counting Kit-8 (A311, Vazyme, Nanjing Province, China) according to the manufacturer’s instructions. Briefly, approximately 5000 hepatocytes (AML-12 or HepG2 cells) per well were seeded in a 96-well plate based on experimental groups, with five replicates per sample. In the blank control group, an equal volume of complete culture medium was added to each well to maintain consistent experimental conditions. 10 μL of Cell Counting Kit-8 solution was added to each well. The plate was then incubated at 37 °C with 5% CO₂ in a humidified atmosphere for 1 hour in the dark. After incubation, OD values were measured at 450 nm using a microplate reader (BioTek Epoch, PA, United States), with 630 nm serving as the reference wavelength for background subtraction. Mean values from replicates were calculated for data analysis.

Determination of ferrous iron levels

The impact of GATA2 deficiency on hepatocyte ferrous iron (Fe²⁺) concentrations was quantitatively assessed using the Ferrous Ion Content Assay Kit (AKIC004M, Boxbio, China). Liver tissues were lysed with a concentrated sulfuric acid-chloroform system to ensure complete disruption of cellular architecture for maximal iron ion release. Under acidic conditions, Fe²⁺ selectively formed a stable blue complex with tripyridyltriazine, characterized by a distinct absorbance maximum at 593 nm. Spectrophotometric measurement at this wavelength enabled precise and reproducible quantification of free Fe²⁺. A standard curve was constructed using FeSO₄ solutions of known concentrations (0.005-0.16 μmol/mL) for calibration. To account for inter-sample variability, all data were normalized to wet liver tissue mass (μmol/g) or total cell count (μmol/10⁶ cells).

Statistical analysis

Statistical analysis was performed using IBM SPSS Statistics 29.0 (CA, United States). The normality of continuous variables was assessed using the Shapiro-Wilk test. Normally distributed data were presented as mean ± SD, with intergroup comparisons conducted using either the independent samples t-test or Welch’s corrected t-test, depending on Levene’s test for homogeneity of variance. For non-normally distributed data, values were reported as median (interquartile range) and analyzed using the Mann-Whitney U test. All statistical comparisons in this study were based on two-group analyses, with a significance threshold set at α = 0.05.

RESULTS

Case presentation

A 55-year-old female patient presented with dizziness, fatigue, and thrombocytopenia that had persisted for three years. Following ineffective supportive treatment, she developed significant fatigue, abdominal distension, and jaundice, indicating liver injury. During the diagnosis of AA disease, the patient had taken TCM for three months and received two blood transfusions. Examination revealed abnormal liver function, characterized by increased TBil and direct bilirubin, inverted albumin-to-globulin ratio (Table 4), and decreased red blood cells, white blood cells, and platelets as indicated by routine blood tests (Table 5). Further analysis showed reduced active bone marrow erythroid hyperplasia, predominantly in middle and late juvenile red blood cells, a scarcity of megakaryocytes, and an absence of platelets (Table 6).

Table 4 Dynamic changes in the patient’s liver biochemical indexes.

Date	ALT (U/L)	AST (U/L)	TBil (μmol/L)	DBil (μmol/L)	ALP (U/L)	GGT (U/L)	ALB (g/L)	GLB (g/L)
	Reference: 7-40	Reference: 13-35	Reference: 5-21	Reference: 0-3.4	Reference: 50-135	Reference: 7-45	Reference: 40-55	Reference: 20-40
December 3, 2021	73↑	53↑	115.1↑	56.4↑	166↑	153↑	28↓	26
December 5, 2021	51↑	43↑	106.4↑	49↑	158↑	129↑	27.6↓	27
December 9, 2021	26	34	83.2↑	38.8↑	149↑	118↑	27.7↓	26
December 12, 2021	21	32	76.9↑	35.1↑	131	105↑	30.9↓	24
December 14, 2021	15	27	84.4↑	39.7↑	142↑	97↑	28.7↓	27
December 17, 2021	13	26	73.3↑	31↑	150↑	85↑	33.8↓	25
August 4, 2022	21	26	15.4	4.5↑	176↑	62↑	37.4↓	25
August 6, 2022	44↑	15	20.9↑	5.5↑	157↑	48↑	32.5↓	19↓
August 8, 2022	76↑	26	20.3	6.2↑	153↑	45	32.7↓	19↓

“↑” indicates values above the normal range, and “↓” denotes values below the normal range. ALT: Alanine transaminase; AST: Aspartate aminotransferase; TBil: Total bilirubin; DBil: Direct bilirubin; ALP: Alkaline phosphatase; GGT: Gamma-glutamyl transferase; ALB: Albumin; GLB: Globulin.

Table 5 Dynamic changes in the patient's blood routine indexes.

Date	WBC (× 109/L)	RBC (× 1012/L)	HB (g/L)	PLT (× 109/L)	NEUT (%)	Lym, %
	Reference: 3.5-9.5	Reference: 3.8-5.1	Reference: 115-150	Reference: 125-350	Reference: 0.40-0.75	Reference: 0.20-0.50
December 1, 2021	2.0↓	1.77↓	70↓	9↓	0.66	0.27
December 3, 2021	1.92↓	1.94↓	79↓	41↓	0.49	0.40
December 5, 2021	1.46↓	1.99↓	81↓	21↓	0.40	0.49
December 9, 2021	1.51↓	1.87↓	74↓	12↓	0.35↓	0.54↑
December 12, 2021	1.62↓	1.71↓	70↓	30↓	0.41	0.48
December 14, 2021	4.26	1.71↓	71↓	21↓	0.72	0.22
December 15, 2021	2.15↓	1.65↓	70↓	46↓	0.64	0.30

“↑” indicates values above the normal range, and “↓” denotes values below the normal range. WBC: White blood cell; RBC: Red blood cell; HB: Hemoglobin; PLT: Platelets; NEUT: Neutrophils; Lym: Percentage of lymphocytes.

Table 6 Dynamic changes in bone marrow examinations.

Date	G/E	NRBC	MKC (a)	PLT (a)	LYM%	Bone marrow granules	Blood
June 19, 2020	1.39:1↓	Hyperplasia of active	11	Rare and scattered	Increased	Relatively full	White blood cells reduce
August 31, 2020	2.86:1	Hyperplasia of active	Not	Not	Increased	Not seen	White blood cells reduce
June 15, 2021	5.53:1↑	Hyperplasia of active	Not	Rare	Increased	Not seen	White blood cells reduce
January 18, 2022	1.72:1↓	Hyperplasia of active	Not	Not	Increased	Not seen	White blood cells reduce

G/E: Grain:red; NRBC: Nucleated red blood cell; MKC: Megakaryocyte; PLT: Platelets; LYM: Lymphatic ratio.

Clotting was normal after hospitalization, but C-reactive protein was elevated at 36.6 mg/L (normal range: 0.068-8.2 mg/L). Systematic screening for common hepatotoxic pathogens, including hepatitis A virus, hepatitis C virus, hepatitis E virus, Epstein-Barr virus, and cytomegalovirus, revealed no current infections. Thyroid function tests showed an increase in T3 levels, while other markers were within normal ranges. Abdominal computed tomography (CT) indicated no abnormal obstruction of hepatic venous return, ruling out cirrhosis with hepatic jugular obstruction. Analysis of iron metabolism revealed elevated serum iron and transferrin levels, decreased total iron-binding capacity and unsaturated iron-binding capacity, and increased transferrin saturation, indicating an increased iron load in the patient (Table 7). Antinuclear antibodies were positive (1:100 positive; nuclear plus cytoplasmic particle type), while all other autoimmune hepatitis antibodies were negative. Repeat liver function tests revealed no progressive elevation in bilirubin, gamma-glutamyl transferase, or alkaline phosphatase levels, indicating no signs of primary biliary cholangitis.

Table 7 Dynamic changes in serum-related iron indexes.

Date	Zn, μmol/L	Cu, μmol/L	Fe, μmol/L	UIBC, μmol/L	TIBC, μmol/L
	Reference: 10.7-17.7	Reference: 12.56-23.55	Reference: 9-27	Reference: 22.4-57.8	Reference: 50-77
November 25, 2021	4.5↓	17.12	32.3↑	1.5↓	34↓
May 31, 2022	-	-	24.6	18.1↓	42.7↓

“↑” indicates values above the normal range, and “↓” denotes values below the normal range. UIBC: Unsaturated iron binding capacity; TIBC: Total iron binding capacity.

The patient’s condition improved with liver protection and symptomatic treatment. Given the severity of the liver injury and the possibility of a significant underlying etiology, a liver biopsy was performed following a platelet transfusion that raised the count to 33 × 10⁹/L. The histopathological analysis revealed chronic active hepatitis, moderate inflammation, early signs of cirrhosis, and grade 2 liver iron deposition (Figure 1).

Figure 1 Pathological findings of liver biopsy in the patient. A: Moderate interface inflammation with infiltration of lymphocytes and neutrophils; B: Hepatocytes containing yellow-brown granules and focal confluent necrosis; C: Masson trichrome stain: Portal-based fibrous expansion and septum (blue collagen); D: Prussian blue stain: Iron granules (blue-black deposits); E: Cytokeratin 7 immunohistochemistry: Biliary differentiation; F: Sirius red stain: Pericellular/periportal fibrosis (red/orange birefringence).

Abdominal color Doppler ultrasound demonstrated a normal-sized and shaped liver with a full capsule, though there was localized thickening. The splenic intercostal width measured approximately 45 mm (normal: ≤ 40 mm), and the subcostal length was approximately 114 mm (normal: < 110 mm). The spleen surface was smooth, and the interior echoes were uniform. These findings indicated thickening of the splenomegaly. Endoscopic evaluation showed the diagnoses of chronic non-atrophic gastritis and duodenitis.

Identification of GATA2 mutation and diagnosis of AA

A whole genomic test revealed a heterozygous missense mutation, NM_032638.5:C.1286G>C (p.Ser429Thr), in the coding region of the GATA2 gene (Figure 2A). Chromosomal analysis using G-banding was performed on the peripheral karyotype, with 20 divisions counted across 5 karyotypes, and no abnormalities in chromosome number or structure were found (Figure 2B). Given that mutations in the GATA2 gene are closely linked to the pathogenesis of both AA and MDS and that distinguishing AA from MDS represents a clinical challenge, flow cytometry and bone marrow biopsy were used for disease screening. The results showed no increase in naive cells and only rare megakaryocytes (Table 8, Figure 2C and D), confirming the diagnosis of AA. During a 5-year follow-up, the patient’s hematological symptoms remained stable under supportive treatment, with no evidence of progressive disease deterioration, further supporting the diagnosis of AA.

Figure 2 Identification of GATA2 mutation and diagnosis of aplastic anemia in the patient. A: Sanger sequencing of the gene; B: Karyotype analysis; C: Pathological findings of bone marrow biopsy. The proliferation of bone marrow nucleated cells was low; the erythroid lineage was mainly composed of intermediate and late-stage cells; megakaryocytes were rare; lymphocytes were scattered in small numbers; CD34 small vessels (+); few CD61 megakaryocytes (+); D: Flow cytometric analysis. The proportion of CD34+ cells in nuclear cells was about 0.49%, with no obvious abnormality in the immunophenotype. The relative proportion of granulocytes was normal, and the immunophenotypes CD13, CD16, CD15, and CD11b were disordered. Approximately 2.60% of CD19+CD10+ immature B lymphocytes, consistent with normal B progenitor cell proliferation.

Table 8 Results of flow cytometric immunofluorescence analysis.

Cell population	Percentage of total (%)	Cell series/phenotype analysis
Lymphoid	6.53	The relative ratio decreased
CD45 weakly expressed cells	5.75	Some of them were normally hyperplastic B progenitor cells
CD45 negative expression cells	24.57	They were mainly nucleated red cells and cell debris

Development of decompensated cirrhosis one year post-discharge

The patient was discharged after improvement with symptomatic supportive therapy, including reduced GSH, monoammonium glycyrrhizinate, and ursodeoxycholic acid (Yucivir) for hepatoprotection, without the need for immunosuppressive therapy. Following discharge, the patient was scheduled for regular follow-up visits. Over a 5-year follow-up period, abnormal transaminase levels were detected during the first-year assessment, while bilirubin levels were normal. Additionally, ferritin and transferrin levels increased, and total iron-binding capacity decreased (Table 7). Follow-up imaging with upper abdominal CT (Figure 3A) and magnetic resonance imaging (Figure 3B) revealed cirrhosis, splenomegaly, and ascites. During the subsequent 4-year follow-up, the patient’s cirrhosis remained relatively stable without significant disease progression.

Figure 3 Abdominal magnetic resonance imaging and computed tomography findings of cirrhosis during follow-up. A: Upper abdominal computed tomography (plain and enhanced) scans revealed irregular margins with slightly widened fissures, and strip low-density shadows around the portal vein in the liver; B: Upper abdominal magnetic resonance imaging revealed heterogeneous signal intensity with an irregular contour, slightly widened fissures, heterogeneous enhancement, and a long T2 signal around the portal vein in the liver. The spleen was enlarged. The upper abdominal magnetic resonance imaging findings indicated cirrhosis with splenomegaly and ascites. No abnormal enhancement was observed on the enhanced scan. The spleen was enlarged.

Identification of GATA2 binding sites in gene promoters through ChIP-seq and prediction of its regulatory functions

Intrigued by the rapid progression to cirrhosis in our patient with AA attributed to the heterozygous missense mutation in the GATA2 gene (NM_032638.5:C.1286G>C, p.Ser429Thr), we performed further in vitro studies to investigate the potential role of GATA2 in the development of liver injury. ChIP-seq was conducted in HepG2 cells, indicating a strong correlation between biological replicates (Spearman = 0.8; Figure 4A). A total of 42127 GATA2 binding peaks were identified, with an average length of 380 ± 297 bp (Figure 4B). The distribution of these peaks across chromosomes is illustrated in Figure 4C. The coverage depth of reads surrounding the transcription start sites of peak-associated genes is depicted in Figure 4D. To investigate the potential regulatory functions of GATA2 binding, genes with peaks located in the promoter-transcription start sites region and ranked among the top 2000 most significantly enriched peaks (criteria: Input reads ≥ 2 and P-value < 0.05) were subjected to functional enrichment analysis.

Figure 4 Identification of GATA2 binding sites in gene promoters through chromatin immunoprecipitation sequencing and prediction of its regulatory functions. A: Scatter plot showing the correlation between biological replicates; B: Histogram depicting the distribution of peak lengths; C: Distribution of peaks across chromosomes (X-axis: Chromosome length; left Y-axis: Peak intensity; right Y-axis: Chromosome numbers); D: Read distribution surrounding transcription start sites (TSS) of peak-associated genes (top: Line chart of average sequencing depth around TSS; bottom: Heatmap of sequencing depth around TSS for peak-associated genes, ranked by average depth); E: Gene Ontology functional enrichment analysis of peak-associated genes; F: Kyoto Encyclopedia of Genes and Genomes pathway enrichment analysis of peak-associated genes; G: Fold enrichment of selected genes associated with partial regeneration, molecular chaperones, antioxidant response, and fatty acid β-oxidation as determined by chromatin immunoprecipitation sequencing; H: Validation of GATA2 binding to the promoters of heat shock protein family D member 1, glutathione peroxidase 4, and peroxisome proliferator-activated receptor alpha by chromatin immunoprecipitation real-time quantitative polymerase chain reaction. Data are presented as fold enrichment relative to immunoglobulin G; I: Western blot analysis of GATA2, heat shock protein 60, glutathione peroxidase 4, and peroxisome proliferator-activated receptor alpha protein expression following transfection with the Gata2 construct. ^bP < 0.01 vs immunoglobulin G or control group. ChIP-seq: Chromatin immunoprecipitation sequencing; IgG: Immunoglobulin G; PCNA: Proliferating cell nuclear antigen; HSPD1: Heat shock protein family D member 1; GPX4: Glutathione peroxidase 4; UCP2: Uncoupling protein 2; PPARA: Peroxisome proliferator-activated receptor alpha.

Gene Ontology enrichment analysis revealed that these genes were enriched in functions related to RNA binding, nucleoplasm, cytosol, viral processes, mRNA binding, nucleus, and cytoplasmic stress granules (Figure 4E). Furthermore, Kyoto Encyclopedia of Genes and Genomes pathway enrichment analysis highlighted their involvement in pathways such as RNA transport, ubiquitin-mediated proteolysis, protein processing in the ER, hepatocellular carcinoma, chronic myeloid leukemia, and the cell cycle (Figure 4F).

ChIP-seq analysis revealed enrichment of genes associated with partial regeneration [proliferating cell nuclear antigen (PCNA)], molecular chaperones [Heat shock protein family A member 5 (HSPA5) and HSPD1], antioxidant response [GSH peroxidase 4 (GPX4) and uncoupling protein 2], and fatty acid β-oxidation [peroxisome proliferator-activated receptor alpha (PPARA) and acyl-CoA dehydrogenase] (Figure 4G). This finding was validated by ChIP-real-time quantitative polymerase chain reaction, which confirmed GATA2 binding to the promoters of HSPD1, GPX4, and PPARA, with data presented as fold enrichment over the IgG control (Figure 4H; all P < 0.01).

Furthermore, western blot analysis demonstrated that transfection with the GATA2 construct significantly increased the protein expression levels of GATA2, HSP60 (a mitochondrial oxidative stress marker, encoded by HSPD1), GPX4, and PPARα (encoded by PPARA) compared to the empty vector control (Figure 4I; all P < 0.01).

Elevated hepatic GATA2 protein levels in CLI mice

We established a mouse model of liver injury and compared GATA2 protein levels between the vehicle control group and the CCl₄-induced liver injury group. CCl₄-treated mice exhibited significantly elevated serum ALT and TBil levels compared to the vehicle control group (Figure 5A and B; all P < 0.01). Histological analysis confirmed the successful establishment of the CLI model, evidenced by disrupted hepatic lobular architecture, pseudolobule formation, extensive necrosis, portal collagen deposition, and inflammatory infiltration in CCl₄-treated livers (Figure 5C; P < 0.01). Western blot analysis demonstrated a significant upregulation of GATA2 protein expression in CLI livers compared to controls (Figure 5D; P < 0.01).

Figure 5 Liver injury assessment and hepatic GATA2 protein expression in chronic liver injury mice. A: Serum alanine aminotransferase levels; B: Serum total bilirubin levels; C: Representative hematoxylin and eosin and Masson’s trichrome staining images and quantitative analysis of histopathological necrotic areas; D: Western blotting and quantitative analysis of GATA2 protein expression. ^bP < 0.01 vs control group. ALT: Alanine aminotransferase; TBil: Total bilirubin; H&E: Hematoxylin and eosin.

Upregulation of GATA2 protein in hepatocytes under ER stress and oxidative stress

Following the findings of elevated GATA2 protein levels associated with liver injury in mice, we evaluated its expression in mouse and human hepatocyte cell lines, AML-12 and HepG2, under ER and oxidative stress, respectively. Treatment of AML-12 mouse hepatocytes with THA for 48 hours significantly reduced cell viability compared to solvent controls (Figure 6A; P < 0.01). The cell culture model of ER stress was established, evidenced by a marked increase in glucose-regulated protein 78 (GRP78, an ER stress marker, encoded by HSPA5) in THA-induced cells (Figure 6B; P < 0.01). Notably, GATA2 protein expression was significantly upregulated in AML-12 cells under ER stress compared to control cells (Figure 6B; P < 0.01). Similarly, treatment of AML-12 and HepG2 cells with H₂O₂ for 48 hours significantly reduced cell viability compared to controls (Figure 6C and E; all P < 0.01). Western blot analysis demonstrated marked increases in HSP60 in H₂O₂-treated cells (Figure 6D and F; P < 0.01), confirming the successful establishment of cell culture models for oxidative stress. GATA2 protein levels were significantly elevated in oxidative stress-induced hepatocytes compared to controls (Figure 6D and F; all P < 0.01).

Figure 6 Elevated GATA2 protein expression in hepatocytes under endoplasmic reticulum stress and oxidative stress. A: Cell viability assessed by Cell Counting Kit-8 assay in thapsigargin-treated AML-12 cells; B: Semi-quantitative analysis of thapsigargin-induced GATA2 protein expression by western blotting (representative blots shown); C and E: Cell viability measured by Cell Counting Kit-8 assay in AML-12 (C) and HepG2 cells (E) treated with H₂O₂; D and F: Semi-quantitative analysis of H₂O₂-induced GATA2 protein expression by western blotting in AML-12 (D) and HepG2 cells (F) (representative blots shown). ^bP < 0.01 vs control group. HSP60: Heat shock protein 60.

GATA2 overexpression alleviated hepatocyte death and promoted regeneration under ER stress

Fluorescence microscopy demonstrated that both the GATA2 construct and empty vector plasmids achieved efficient transfection in HepG2 cells (Figure 7A). Transfection GATA2 construct significantly elevated GATA2 protein levels in THA-induced HepG2 cells (Figure 7B; P < 0.01). Compared to the THA control group, GATA2 overexpression markedly enhanced cell viability, GSH levels, and ETC-CI activities, while reducing MDA, GSSG, and TG levels in THA-induced HepG2 cells (Figure 7C-H; all P < 0.05). Mechanistically, GATA2 overexpression upregulated the proliferation marker PCNA and downregulated ferroptosis-associated Fe²⁺ levels, the apoptosis marker cleaved caspase-3, and the necroptosis marker phosphorylated mixed lineage kinase domain-like protein in THA-induced HepG2 cells (Figure 7I and J; P < 0.01).

Figure 7 GATA2 overexpression protected HepG2 cells from endoplasmic reticulum stress-induced injury. A: Fluorescence microscopy images confirming transfection efficiency; B: Western blot analysis of GATA2 protein expression, with representative blots displayed alongside quantitative results; C: Assessment of cell viability using the Cell Counting Kit-8 assay; D: Intracellular malondialdehyde content; E: Glutathione levels; F: Oxidized glutathione levels; G: Electron transfer chain complex I activity; H: Triglyceride levels; I: Ferrous ion levels; J: Western blot analysis of proliferating cell nuclear antigen, cleaved caspase-3, and phosphorylated mixed lineage kinase domain-like protein expression (representative blots shown) with quantification. ^cP < 0.05 vs the thapsigargin group; ^dP < 0.01 vs the thapsigargin group. THA: Thapsigargin; MDA: Malondialdehyde; GSH: Glutathione; GSSG: Oxidized glutathione; ETC-CI: Electron transfer chain complex I; TG: Triglyceride; p-MLKL: Phosphorylated mixed lineage kinase domain-like protein; MLKL: Mixed lineage kinase domain-like protein.

Hepatocyte-specific knockdown of Gata2 exacerbated hepatocyte death-regeneration imbalance in CLI mice

Next, we employed hepatocyte-specific Gata2 knockdown mice to investigate a potentially direct role of GATA2 in liver injury. Compared with the control group, Gata2 shRNA transfection did not significantly alter serum ALT levels in mice (Figure 8A; P > 0.05). Histopathological analysis of liver tissues from Gata2 shRNA-transfected mice demonstrated intact hepatic architecture with normal lobular morphology, with no signs of hepatocyte necrosis, portal collagen deposition, or inflammatory infiltration (Figure 8B). Western blot analysis revealed significantly reduced hepatic GATA2 protein expression in Gata2 shRNA-transfected mice (Figure 8C; P < 0.01), indicating GATA2 knockdown.

Figure 8 Hepatocyte-specific Gata2 knockdown exacerbates hepatocyte death-regeneration imbalance in chronic liver injury mice. A-C: Gata2 shRNA-transfected healthy mice: Serum alanine aminotransferase levels (A); hepatic histopathology by hematoxylin and eosin staining (B); western blot analysis of GATA2 protein expression (representative blots shown) with quantification (C); D-I: Gata2 knockdown intervention in chronic liver injury mice: Effect of Gata2 knockdown on serum alanine aminotransferase levels in chronic liver injury mice (D); serum total bilirubin levels (E); quantitative analysis of necrotic areas with representative hematoxylin and eosin and Masson’s trichrome staining images (F); fluorescent Terminal deoxynucleotidyl transferase-mediated dUTP nick-end labeling staining analysis of hepatocyte apoptosis rate (representative images shown) (G); ferrous ion content (H); western blot analysis of GATA2 protein expression (blots displayed) with quantification (I); J: Western blot analysis of proliferating cell nuclear antigen, cleaved caspase-3, and phosphorylated mixed lineage kinase domain-like protein expression (representative blots shown) with quantification. ns: P > 0.05, ^bP < 0.01 vs control group; ^cP < 0.05 vs chronic liver injury group (CCl₄), ^dP < 0.01 vs chronic liver injury group (CCl₄). ALT: Alanine aminotransferase; H&E: Hematoxylin and eosin; TBil: Total bilirubin; PCNA: Proliferating cell nuclear antigen; p-MLKL: Phosphorylated mixed lineage kinase domain-like protein; MLKL: Mixed lineage kinase domain-like protein.

In the CLI mouse model, the Gata2 knockdown group exhibited two mortality events, while the control group had none. Hepatocyte-specific Gata2 knockdown in CLI mice showed significantly higher serum ALT and TBil levels compared to the control CLI group (Figure 8D and E; P < 0.01). Histopathological analysis revealed aggravated hepatic structural disruption in Gata2 knockdown mice, marked by increased pseudolobule formation, extensive hepatocyte necrosis, enhanced portal collagen deposition, and intensified inflammatory infiltration (Figure 8F; P < 0.05). The TUNEL assay indicated a higher number of apoptotic hepatocytes in the Gata2 knockdown group vs CLI controls (Figure 8G; P < 0.05). Mechanistically, Gata2 deficiency upregulated Fe²⁺ levels, cleaved caspase-3, and phosphorylated mixed lineage kinase domain-like protein levels, while downregulating the protein levels of GATA2 and the proliferation marker PCNA in CLI livers (Figure 8H-J; P < 0.01), suggesting a disruption in the balance between hepatocyte death and regeneration.

Hepatocyte-specific knockdown of Gata2 impaired hepatic stress adaptation and inhibited fatty acid β-oxidation in CLI mice

Compared to the CLI group, hepatocyte-specific Gata2 knockdown significantly increased MDA, GSSG, and TG levels, alongside decreased hepatic GSH levels and reduced activities of ETC-CI (Figure 9A-E; all P < 0.05). TEM revealed mitochondrial matrix condensation, disruption of cristae, vesiculated dilation of the rough ER, and increased deposition of cytoplasmic lipid droplets in hepatocytes of Gata2-deficient mice (Figure 9F). Western blot analysis demonstrated that Gata2 knockdown significantly reduced the protein expression of GRP78, HSP60, GPX4 (antioxidant), uncoupling protein 2, PPARα, and medium chain acyl-CoA dehydrogenase (MCAD, encoded by acyl-CoA dehydrogenase while increasing the protein expression levels of the ER stress pro-apoptotic factor CHOP (Figure 9G-I; P < 0.01).

Figure 9 Hepatocyte-specific Gata2 knockdown impairs hepatic stress adaptation and exacerbates triglyceride metabolic dysregulation in chronic liver injury mice. A: Hepatic malondialdehyde content; B: Glutathione levels; C: Oxidized glutathione levels; D: Electron transfer chain complex I activity; E: Triglyceride levels; F: Representative transmission electron microscope images showing ultrastructural morphology: Nucleus (N), mitochondria (M) with cristae disruption, dilated endoplasmic reticulum (ER), and lipid droplets (L); G: Analysis of endoplasmic reticulum stress-related protein expression; H: Western blot analysis of oxidative stress-related protein expression levels; I: Western blot analysis of triglyceride metabolism-related protein expression levels. ^cP < 0.05, ^dP < 0.01 vs chronic liver injury group (CCl₄). MDA: Malondialdehyde; GSH: Glutathione; GSSG: Oxidized glutathione; ETC-CI: Electron transfer chain complex I; TG: Triglyceride; p-MLKL: Phosphorylated mixed lineage kinase domain-like protein; MLKL: Mixed lineage kinase domain-like protein; HSP60: Heat shock protein 60; GRP78: Glucose-regulated protein 78; GPX4: Glutathione peroxidase 4; UCP2: Uncoupling protein 2; PPARα: Peroxisome proliferator-activated receptor alpha; MCAD: Medium chain acyl-CoA dehydrogenase.

DISCUSSION

The relationship between AA and CLI, including cirrhosis, remains poorly understood. This study explored the potential genetic link between GATA2 mutations - recognized as disease-causing mutations for AA - and CLI, demonstrating that these mutations exacerbated the imbalance between hepatocyte death and regeneration, thereby driving CLI progression. The key novel findings are summarized as follows: (1) A clinical case study in a 55-year-old woman with AA and rapid progression to cirrhosis identified a heterozygous missense mutation, NM_032638.5:C.1286G>C (p.Ser429Thr), in the coding region of the GATA2 gene, confirming its role as a disease-causing mutation for AA. This prompted our further in-depth investigation into the potential direct role of GATA in liver injury using experimental models in mice and hepatocytes; (2) In these models, we demonstrated that hepatic GATA2 expression was compensatorily upregulated during ER stress and oxidative stress associated with liver injury; (3) Hepatocyte-specific Gata2 knockdown exacerbated hepatocyte apoptosis, necroptosis and ferroptosis while suppressed hepatocyte regeneration; and (4) Further analysis revealed that Gata2 depletion impaired the adaptive capacity of the unfolded protein response (UPR), reduced antioxidant defense, and exacerbated fatty acid β-oxidation disorders. Collectively, these findings confirmed that GATA2 deficiency disrupted the balance between hepatocyte death and regeneration, providing initial scientific evidence for a genetic link of AA-associated GATA2 mutations with CLI.

AA complicated by idiopathic cirrhosis

The patient developed liver damage following a brief course of TCM treatment prior to the diagnosis of AA. She had received TCM for three months and two blood transfusions. The examination found concomitant chronic liver damage. Other potential causes of liver damage, such as viruses like the hepatitis virus, Epstein-Barr virus, and cytomegalovirus, were ruled out through negative findings. Similarly, thyroid function tests showed elevated T3, with all other markers in the normal ranges. Abdominal CT indicated no apparent abnormal obstruction of hepatic venous return, excluding cirrhosis secondary to obstruction of hepatic jugular return. Iron metabolism-related indices in our patient revealed elevated serum iron and transferrin levels, indicating an iron load. Antibodies to autoimmune hepatitis were positive for antinuclear antibodies, while all other antibodies were negative. Liver function tests revealed normal bilirubin levels, ruling out primary biliary cholangitis. Liver biopsy revealed chronic active hepatitis with moderate inflammation, a local tendency toward cirrhosis, and grade 2 hepatic iron deposition[31,32]. Although the patient had a three-month TCM treatment, the progression of cirrhosis cannot be attributed solely to TCM exposure. The rapid clinical deterioration from the pre-cirrhotic stage to decompensated cirrhosis within one year suggests the involvement of multiple synergistic pathogenic factors. Notably, iron overload may serve as a critical synergistic mechanism accelerating the progression from CLI to end-stage liver disease.

Iron overload alone cannot fully explain the etiology of cirrhosis in this patient

While iron is an essential trace element, excessive amounts are toxic[33]. Dietary iron is absorbed via divalent metal transporter 1 in intestinal cells and transported to portal circulation via ferroportin, where it binds to transferrin and is taken up by hepatocytes, macrophages, and bone marrow cells via transferrin receptor 1[34,35]. Iron homeostasis relies on the regulation of transferrin receptor 1, ferroportin, and ferritin expression[36]. Systemic iron overload leads to tissue deposition and injury[37]. In this patient, iron overload was thoroughly considered. Iron excess may also indicate hereditary hemochromatosis, an acquired disorder caused by excessive iron intake, absorption, or repeated transfusions[38,39]. The liver is primarily responsible for iron metabolism and plays a crucial role in iron homeostasis. Excessive hepatic iron deposition can cause CLI[40,41], liver fibrosis[42], cirrhosis[43], and even cause liver failure or hepatocellular carcinoma[44]. Iron overload may arise from congenital disabilities in iron metabolism, chronic anemia, hemolysis[45], or transfusion diseases[46], resulting in excessive ferritin accumulation in hepatocytes. When cellular iron in hepatocytes surpasses a certain threshold, high ROS levels overwhelm antioxidant defenses, resulting in oxidative damage to hepatocytes[47]. Thus, we hypothesized that patients with systemic and liver diseases are influenced by iron metabolism, which, in turn, affects the hepatic iron load, thereby aggravating liver damage. However, in this patient, the cause of the abnormal iron metabolism could not be explained simply by laboratory iron metabolism indexes and associated medical history, despite two erythrocyte transfusions and short-term TCM use. The patient’s whole-genome sequencing was performed to assess potential genetic involvements in liver injury and clarify the crosstalk between iron metabolism in the liver and blood. Our findings revealed a heterozygous dominant mutation in GATA2 gene. Given its association with MDS, bone marrow aspiration, biopsy, and flow cytometry were conducted, consequently ruling out MDS.

The heterozygous missense variant c.1286G>C (p.Ser429Thr) in GATA2 identified in this case co-segregates with the disease phenotype; however, its pathogenic mechanism requires cautious interpretation. This c.1286G>C (p.Ser429Thr) alteration constitutes a missense mutation within the coding region of GATA2 (National Center for Biotechnology Information ClinVar; VCV000404078.24, https://www.ncbi.nlm.nih.gov/clinvar/variation/VCV000404078.24, accessed October 14, 2025). According to the large-scale population frequency database gnomAD, this variant has been observed in 22 heterozygous individuals, with no reported homozygotes, resulting in an overall minor allele frequency of 0.00007778 and an East Asian population frequency of 0.001052 (National Center for Biotechnology Information ClinVar; VCV000404078.24). Literature reports its detection in a child with MDS[48]. Based on current evidence, this variant is classified as a Variant of Uncertain Significance. To avoid overinterpretation of causal relationships based solely on individual loci, further investigation into the potential functional impact of this mutation is warranted.

Despite GATA2 mutations being a shared pathogenic factor for both AA and MDS, their treatment responses and prognoses differ significantly

The GATA2 gene encodes a critical transcription factor that regulates the proliferation and maintenance of HSPCs[49,50]. GATA2 expression was highly expressed in HSPCs, but decreased after expression in erythroid progenitor cells[51]. GATA2 plays a key role in controlling lineage-limiting genes during erythroid differentiation[52]. Notably, the regulation of GATA expression in erythroid cells differs from that of mast cells[53]. During myeloid differentiation, GATA2 is typically downregulated, a process inhibited by forced GATA2 overexpression[54]. Furthermore, GATA2 is essential for dendritic cell (DC) differentiation, maturation, and monocyte differentiation into inflammatory DC[55]. In bone metabolism, GATA2 inhibits osteogenesis while promoting osteoclast differentiation, potentially contributing to bone metabolic disorders[56,57]. Clinically, GATA2 mutations are found in up to 10% of patients with congenital neutropenia or AA cases[58] and represent a novel susceptibility factor for familial MDS and AML[59]. Pathogenic GATA2 variants can lead to MDS[60,61], Emberger syndrome[62], or AML[63,64]. Given the overlapping hematologic features of AA and MDS - yet their markedly different prognoses - accurate differentiation is crucial[65,66]. Our patient underwent serial blood tests, indicating that the three systems were reduced. Multiple bone marrow examinations revealed that the red blood system primarily comprised middle- and late-stage erythroblasts, with only rare megakaryocytes and platelets. Additionally, there was no increase in naive cells or megakaryocytes, and neither acute leukemia nor immunophenotypic abnormalities associated with high-risk MDS were observed. These findings, along with the GATA mutation, confirmed the diagnosis of GATA2 mutation-associated AA complicated by cirrhosis, while excluding MDS.

Mutations in the GATA2 gene may be a critical factor in accelerating liver injury progression

GATA2 is a transcriptional regulator involved in hematopoietic and non-hematopoietic embryonic stem cell differentiation. In vitro, mesenchymal stem cells (MSCs) can be induced to differentiate into hepatocyte-like cells, acquiring a partial liver cell phenotype and function. These differentiated MSCs can settle and survive in the splenic parenchyma, with some migrating to the host liver[67]. Recent studies have shown that GATA2 deletion not only impairs bone marrow hematopoiesis but also reduces myeloid cell populations, including monocytes, B lymphocytes, natural killer, and DCs. The GATA2 gene plays a crucial role in hematopoietic stem cell differentiation, as well as the development and activation of myeloid cells[68]. In patients with AA, splenic abnormalities, dysregulated iron metabolism, and increased myelodysplasia contribute to diminishing the compensatory capacity of the bone marrow stem cells. The role of GATA2 in liver injury is not fully understood. Our study showed that hepatocyte-specific knockdown of Gata2 exacerbates CLI in mice.

GATA2 deficiency may disrupt the regulatory network of hepatocyte homeostasis

Hepatocyte homeostasis is essential for maintaining structural integrity, functional compensatory capacity, and metabolic dynamic equilibrium of the liver. Hepatocytes, making up approximately 80% of the liver’s volume, are crucial for the dynamic balance between hepatocyte death and regeneration, vital for preserving hepatic lobule architecture. Excessive hepatocyte loss without adequate regeneration activates hepatic stellate cells and drives fibrogenesis. Apoptosis and necroptosis represent the two primary pathways of programmed hepatocyte death. Apoptosis is a caspase cascade-driven, non-inflammatory process characterized by chromatin condensation and cell shrinkage, whereas necroptosis, triggered by inflammatory signals, involves mixed lineage kinase domain-like protein phosphorylation[69,70], leading to pore formation in the plasma membrane and the release of cellular contents that activate Kupffer cells and amplify inflammation[71,72]. Notably, ferroptosis is a manifestation of dysregulated iron metabolism, representing an iron-dependent form of non-apoptotic cell death[73]. Its defining characteristics include excessive accumulation of lipid ROS and functional inactivation of GPX4[74-76]. In this study, GATA2 depletion induced iron overload in hepatocytes, which may subsequently lead to the substantial production of lipid-free radicals and directly activate ferroptosis signaling cascades. Hepatocyte regeneration is the primary repair mechanism of the liver in response to injury, with its effectiveness, together with other compensatory pathways (such as stem cell activation), determining the outcome of liver injury[77]. In acute liver injury, mature hepatocytes proliferate to replace damaged cells[78]. However, chronic injury diminishes regenerative capacity, leading to repair dysfunction and promoting fibrosis and cirrhosis progression through mechanisms such as inhibition of MSC function[79]. Pluripotent stem cells demonstrate significant therapeutic efficacy in treating liver failure, holding substantial potential for clinical translation within the realm of liver regenerative medicine[80]. PCNA, a key biomarker for hepatocyte regeneration, correlates positively with regenerative efficiency in early-stage chronic liver disease. PCNA expression intensity positively correlates with the regenerative efficiency of residual hepatocytes but declines as the disease progresses, indicating exhausted regenerative capacity[81,82]. Inhibition of PCNA expression at this stage leads to regenerative failure and accelerates hepatic failure. In mice, global Gata2 knockout increases apoptosis of hematopoietic stem cells and impairs regenerative capacity[83,84]. In AML cell lines, shRNA-mediated suppression of GATA2 attenuates cell proliferation and induces apoptosis[85]. Our study showed that GATA2 deficiency enhances hepatocyte apoptosis, necroptosis, and ferroptosis and reduces cellular regeneration. This ferroptotic phenotype exhibited a strong correlation with clinical observations of grade 2 hepatic iron deposition and rapid fibrosis progression in the patient’s liver tissue. These findings suggest that GATA2 deficiency may accelerate the chronic progression of liver injury by disrupting the balance between hepatocyte death and regeneration. However, the mechanism by which Gata2 deficiency affects the balance between hepatocyte death and regeneration remains unexplored.

Lipid metabolic homeostasis plays a crucial role in maintaining hepatocellular stability, balancing fatty acid oxidation, and TG metabolism to ensure functional integrity of hepatocytes[86]. Hepatocytes regulate fatty acid β-oxidation through the PPARα pathway[87], modulate lipid synthesis via the sterol regulatory element-binding protein-1c pathway, and achieve metabolic adaptive regulation through autophagic lipolysis (lipophagy) of lipid droplets[88,89]. Dysregulation can lead to lipotoxicity and hinder hepatocyte regeneration by downregulating PCNA expression and blocking G1/S phase transition via saturated fatty acids, worsening chronic liver disease. GATA2 overexpression in adipose tissue suppresses lipogenic gene expression[90], suggesting that GATA2 deficiency may promote hepatic steatosis. MCAD is crucial for medium-chain fatty acid (C6-C12) β-oxidation; its deficiency leads to impaired metabolism of medium-chain fatty acids, causing hypoglycemia, energy crisis, and lipid accumulation, which can be life-threatening[91]. In our study, hepatocyte-specific knockdown of GATA2 results in elevated intrahepatic TG levels, downregulation of PPARα and MCAD protein expression, suggesting that GATA2 deficiency may disrupt TG metabolism, at least in part, by reducing fatty acid β-oxidation.

The liver, as a metabolically active organ, maintains oxidative-reductive homeostasis through its antioxidant defense system, which eliminates ROS via enzymatic components (superoxide dismutase, GPx, catalase) and non-enzymatic antioxidants (GSH, vitamins C and E). This system balances ROS generated from metabolic processes such as cytochrome P450-mediated detoxification and fatty acid β-oxidation[92,93]. However, abnormal TG accumulation induces lipotoxicity, leading to the buildup of free fatty acids (FFAs) and ceramide (Cer), disrupting homeostasis[94,95]. FFAs activate de novo sphingolipid synthesis via serine palmitoyltransferase, promoting toxic Cer production, and inhibit sarcoplasmic/ER Ca²⁺-ATPase (SERCA) function, thereby disturbing ER calcium homeostasis[96]. When FFAs enter mitochondria, increased electron leakage from the ETC elevates ROS production, directly acting on lipids to induce peroxidation (e.g., MDA formation), which compromises membrane protein cross-linking and permeability. Concurrently, cardiolipin and polyunsaturated fatty acids in mitochondrial and plasma membranes undergo peroxidation, generating lipid peroxidation products that inhibit ETC-CI function and ATP synthesis, ultimately resulting in energy metabolism failure. ROS further disrupts the ER GSH/GSSG redox system, impairing protein folding and activating pro-inflammatory and pro-fibrotic pathways, which drive hepatocellular progression from steatosis to hepatitis, fibrosis, cirrhosis, and even carcinogenesis. Additionally, FFAs and Cer exacerbate oxidative stress by suppressing antioxidant detoxification systems (e.g., cytochrome P450, GSH, GPX4). A reduction in GPX4 expression may impair the cellular clearance of lipid peroxides and lead to ferroptosis in hepatocytes. Our results show that hepatocyte-specific Gata2 knockdown increases MDA levels, reduces ETC-CI activity, and downregulates HSP60 and GPX4 expression, indicating that GATA2 deficiency severely impairs mitochondrial function, diminishes antioxidant capacity, and amplifies oxidative stress. This collapse of redox homeostasis and mitochondrial damage enhances lipotoxicity, thereby accelerating CLI progression.

Hepatocyte homeostasis imbalance can trigger ER stress, leading to the activation of UPR[97], a conserved adaptive mechanism that restores ER homeostasis by inhibiting protein synthesis and enhancing protein processing capacity. If ER homeostasis is not quickly reestablished, persistent UPR may worsen TG metabolic disorders, through multiple signaling pathways, including inositol-requiring enzyme 1, activating transcription factor 6 (ATF6), and protein kinase R-like ER kinase/eukaryotic translation initiation factor 2 alpha/ATF4 pathways[98,99]. Under normal conditions, GRP78, a member of the HSP70 family, binds to inositol-requiring enzyme 1, ATF6, and protein kinase R-like ER kinase to inhibit their activation[100]; however, during ER stress, it binds to unfolded/misfolded proteins, activating UPR[101]. This process modulates protein synthesis inhibition, folding capacity enhancement, or apoptosis initiation, influencing cell fate[102]. Additionally, GRP78 translocates to the plasma membrane, where it regulates cell proliferation and apoptosis, and acts as a potential receptor during pathogen infections[103]. CHOP and caspase-12 specifically mediate ER stress-mediated apoptotic pathways[104,105]. It is activated by the accumulation of unfolded proteins or calcium homeostasis disruption, cleaving downstream effector procaspase-3 to trigger apoptosis. Our results indicate that liver-specific knockdown of GATA2 reduces hepatic GRP78 levels and increases CHOP protein expression, suggesting that GATA2 deficiency impairs UPR’s compensatory function, subsequently worsening ER stress.

In addition, the patient in our study did not progress to cirrhosis until age 56, and her cirrhotic condition remained stable with symptomatic supportive treatment. In animal studies with a Gata2 knockdown mouse model, despite no significant pathological alterations in hepatic tissue, CLI progression was markedly exacerbated, suggesting that GATA2 deficiency itself may not directly induce liver injury but increases susceptibility to external insults by impairing the liver’s ability to maintain homeostasis. As a critical transcription factor for maintaining the stemness of hematopoietic stem cells, GATA2 deficiency can impair cell function[106,107]. Given the pivotal role of hematopoietic stem cells in liver regeneration and repair[108], GATA2 deficiency may compromise the liver’s regenerative capacity, thereby acting as a potential mechanism that exacerbates the progression of CLI.

Limitations and future directions

Despite the strengths of this study demonstrating how GATA2 deficiency in AA drives the progression of CLI by disrupting hepatic homeostasis, providing a complete “clinical case-animal model-cellular mechanism” evidence chain, and advancing the molecular understanding of genetic metabolic liver diseases, several limitations remain. First, the conclusions rely on a single case study, making them hypothesis-generating rather than definitive. Validation in larger patient cohorts is required to confirm the association between GATA2 mutations and cirrhosis. Second, the specific regulatory targets of GATA2 and TG metabolic pathways, as well as the interaction networks with stress pathways such as UPR and antioxidant systems, have not been clarified. There is a lack of interventional studies based on mechanisms to verify potential therapeutic strategies. In future studies, it is necessary to screen the GATA2 mutation spectrum in patients with AA/MDS combined with liver injury, establish a genotype-phenotype database, and map liver-specific transcriptional networks to identify direct GATA2 targets. Additionally, based on the above mechanisms, future studies may develop small-molecule drugs targeting UPR or antioxidant pathways and evaluate their therapeutic potential. This will advance the “clinical phenotype → multi-organ mechanism” paradigm for hereditary cross-organ injuries.

CONCLUSION

GATA2 deficiency drives CLI progression by disrupting hepatocyte homeostasis, particularly through an imbalance between hepatocyte death and regeneration, impaired adaptive UPR, compromised antioxidant defenses, and dysregulated TG metabolism. These findings suggest a shared molecular mechanism linking the hereditary hematologic disorder AA to liver damage. As such, this study enhances our understanding of the pathogenic mechanisms underlying liver injury in AA patients with GATA2 deficiency and may offer potential therapeutic targets to improve patient outcomes.