Abstract

   Background/Objectives: Treatment with tyrosine kinase inhibitors (TKIs)
   in chronic myeloid leukemia (CML) has revolutionized disease management
   and has transformed CML from a life-threatening disease to a chronic
   condition for many patients. However, overcoming resistance,
   particularly related to leukemic stem cells (LSC) that can persist even
   when the bulk of the leukemic cells are eliminated, remains a
   significant challenge. Methods: K562 and KU812 cell lines were treated
   in vitro with the TKI Imatinib (IM). Gene expression, protein analysis,
   and metabolomic screening were conducted to investigate the ability of
   the drug to enhance stem cell (SC) features. Moreover, a gene ontology
   analysis was performed on different available datasets, to further
   consolidate our data. Results: 48 h of IM treatment can significantly
   increase the expression of genes related to SC self-renewal,
   particularly SOX2 and OCT 3/4. Interestingly, these modulations occur
   in cells that remain alive after drug treatment and that displayed
   features consistent with leukemia stem-like CML cells, suggesting that
   SC genes levels are crucial even in cell population survived upon TKI
   treatment. Moreover, after in silico analysis of available data, we
   observed an enrichment of SOX2/NANOG and OCT 3/4 signatures after TKI
   treatment, thus strengthening our results. Conclusions: Our results
   confirmed the relevance of LSC features after TKI treatment,
   highlighting the need for more effective and potentially curative
   strategies targeting LSCs to overcome resistance in CML.

   Keywords: SOX2, OCT 3/4, leukemic stem cells (LSC), TKI, resistance,
   ROS, glutathione, FAO

1. Introduction

   Chronic myeloid leukemia (CML) is a myeloproliferative disorder
   characterized by a significant expansion of primitive hematopoietic
   cells [[38]1]. CML is the first human pathology identified to involve a
   specific karyotype abnormality, the Philadelphia chromosome (Ph)
   [[39]2]. The detection of the Ph chromosome is currently used for
   diagnosing CML, molecular monitoring of residual disease, and follow
   up. The Ph chromosome results from a genetic translocation between
   chromosome 22 and chromosome 9, designated t (9;22), involving BCR and
   c-ABL genes, respectively [[40]3,[41]4]. The resulting BCR-ABL fusion
   protein exhibits oncogenic features, primarily due to its constitutive
   tyrosine kinase activity. These events lead to the transformation of
   hematopoietic stem cells and development of leukemic cells
   [[42]5,[43]6,[44]7].

   The functional significance of BCR-ABL, along with its oncogenic
   tyrosine kinase (PTK) activity, enabled the approval of the first
   generation of tyrosine kinase inhibitors (TKI), Imatinib (IM, Glivec,
   Novartis, Basel, Switzerland), in 2001. IM revolutionized the treatment
   of CML, given its ability to directly target the BCR-ABL enzymatic
   activity [[45]8].

   Second-generation TKIs such as Nilotinib (Tasigna; Novartis, Basel,
   Switzerland) and Dasatinib (Sprycel; Bristol–Myers Squibb, New York,
   NY, USA) were subsequently approved to treat patients resistant or
   intolerant to IM, as this class of inhibitors can bind and block 14 of
   15 different IM-resistant isoforms of BCR-ABL [[46]9]. Third generation
   TKIs like Ponatinib ((Iclusig), Takeda Pharmaceuticals U.S.A., Inc.,
   Cambridge, MA, USA) were developed to treat adult CML patients
   resistant to other TKIs, particularly those harboring T315I mutation
   [[47]10]. The use of TKIs as treatment led to the achievement of
   complete cytogenetic remission and major molecular response (CCgR and
   MMR) in the majority of CML cases. As a result, TKIs discontinuation
   has become a primary objective in the clinical management. However,
   resistance mechanisms, including BCR-ABL amplification, BCR-ABL
   mutations, and leukemic clonal evolution, remain a challenge
   [[48]11,[49]12]. Notably, the persistence of a primitive population of
   leukemia stem cells (LSCs) that can escape TKI treatment is a major
   cause of TKI resistance. LSCs possess self-regenerative capabilities,
   contributing to cancer expansion and driving disease relapse and
   progression [[50]13].

   McCulloch and Till characterized stem cells (SC) by three specific
   properties: self-renewal, differentiation, and proliferative capacity
   [[51]14]. Key molecular regulators of SC-specific functions include
   POU5F1 (alias OCT 3/4) and SOX2, which cooperate with other
   transcription factors to support the expression of pluripotency genes
   essential for stemness. Dysregulation of these genes triggers cascades
   that contribute to cancer pathogenesis, including downregulation of
   tumor suppressors, induction of therapy resistance, and relapse
   [[52]15,[53]16,[54]17].

   Recent studies have highlighted the role of metabolic pathways
   alterations following TKI treatment in promoting LSC resistance and
   poor prognosis [[55]18,[56]19]. In particular, pathways related to
   glucose/glutamate metabolism, amino acid metabolism and lipid
   metabolism are significantly upregulated, suggesting a LSC adaptation
   in order to survive [[57]20,[58]21].

   In this context, there is an urgent need to better discern the
   mechanisms sustaining LSC maintenance and survival to minimize relapse
   and improve long-term outcomes. Therefore, the aim of our project was
   to investigate whether genes related to SC features, along with
   metabolic changes induced by drug treatment, contribute to TKI
   resistance. Using a CML cell line model, complemented by in silico
   analysis of available datasets, we described the unique association
   between specific SC-related genes and TKI treatment. These findings
   suggest potential novel approaches for molecular monitoring of CML
   progression.

2. Materials and Methods

2.1. Cells Line Treatment

   The K562 and KU812 cell lines were purchased from American Type Culture
   Collection (ATCC, Manassas, VA, USA), grown in RPMI 1640 medium
   supplemented with L Glutamine and sodium bicarbonate (Sigma-Aldrich,
   St. Louis, MO, USA), 10% inactivated fetal bovine serum (FBS,
   Sigma-Aldrich, St. Louis, MO, USA), and 0.1% penicillin/streptomycin
   (EuroClone, Milan, Italy), and maintained at 37 °C with 5% of CO[2]. To
   isolate living cells from dead ones, a Ficoll treatment using
   Lymphocyte Separation Media (Lymphosep, AUROGENE, Rome, Italy) was
   performed according to the manufacturer’s instructions. Live cells were
   collected separately and washed with PBS 1X through centrifugation at
   1300 rpm × 5 min.

2.2. Proliferation Assay

   Cells growth was evaluated by MTT assay (Cell Proliferation Kit I (MTT)
   Roche Diagnostics, Indianapolis, IN, USA) according to the
   manufacturer’s instructions. We set four different conditions of IM
   treatment (0, 0.5, 1, 2 µM). To begin, 30,000 cells/per well were
   seeded in triplicate for each condition in a 96-well plate. After 48 h
   of treatment, MTT reagent was added inside each well and, after
   appropriate incubation, the absorbance at 550–600 nm was measured.

2.3. Western Blot

   K562 and KU812 cells were treated with different concentrations of IM
   (0, 0.5, 1, 2 µM) for 48 h. Subsequently, cells were collected and
   lysed with RIPA buffer for 20 min. After a centrifugation at 4 °C, at
   12,000 rpm for 10 min, the supernatants were quantified using the
   Bradford method and 30 µg of proteins were loaded, and then resolved
   through SDS-PAGE 8% for 1.30 h at constant 130 V. Subsequently, gels
   were transferred on PVDF membrane (Amersham Hybond, Sigma-Aldrich, St.
   Louis, MO, USA) for 1 h at constant 400 mA. The membranes were
   saturated with skimmed milk 5% for 2 h at 37 °C and incubated overnight
   at 4 °C with primary antibodies of interest (p-Tyr: Sc- 7020, Actin:
   Sc-4778, BCR: sc-365728, Santa Cruz Biotechnology, Dallas, TX, USA).
   After appropriate secondary antibodies incubation (Anti-rabbit:
   MES176703, Anti-mouse: MBS176647, MyBioSource, San Diego, CA, USA),
   immuno-reactive bands were observed at Chemidoc by using
   chemiluminescent enhanced reagent (Clarity Western ECL Substrate
   #170-5061, Bio-Rad, Hercules, CA, USA) and analyzed using the Image Lab
   program 6.0.0 2017 (BioRad Laboratories, Hercules, CA, USA).

2.4. RNA Extraction and qRT-PCR Analysis

   The RNA was extracted using TRI Reagent (Sigma, St. Louis, MO, USA)
   according to the manufacturer’s instructions. All samples were
   quantified using Nanodrop (ThermoFisher, Waltham, MA, USA) and 500 ng
   of total RNA was used as template for the reverse transcription
   reaction with High-Capacity cDNA Reverse Transcription Kit (Applied
   Biosystem, Foster City, CA, USA) following the specified protocol.
   Expression level of Sox2 (Hs01053049_s1) and OCT3/4 (Hs0999634_gH) were
   investigated with TaqMan technology using a SensiFAST Probe No-ROX kit
   (Meridian Bioscience, Milan, Italy), through the C1000 Thermal Cycler
   CFX96 Real-Time System (Bio-Rad, Hercules, CA, USA). qRT-PCR data were
   analyzed by Bio-Rad CFX Manager 3.1 software (Bio-Rad, Hercules, CA,
   USA). The analysis was performed in triplicate and gene expression was
   expressed after normalization with GusB (Hs00939627_m1) housekeeping
   gene.

2.5. Immunofluorescence

   Immunofluorescence was performed after executing a cytospin of 50,000
   cells [[59]22]. Cells were fixed with 4% paraformaldehyde (PFA)
   (Sigma-Aldrich, St. Louis, MO, USA), quenched with NH4Cl for 20 min,
   then subjected to cell permeabilization with 0.1% of Triton, and then
   saturated with blocking solution (PBS, 1% BSA, 0.3 M glycine) for 40
   min. Subsequently, slides were incubated for 2 h with primary
   antibodies of interest (Santa Cruz Biotechnology, Dallas, TX, USA).
   Alexa Fluor 488 secondary antibodies (Invitrogen, Waltham, MA, USA)
   were added for 40 min for protein detection. Slices were mounted with
   Vectashield Mounting Medium with DAPI (Vector Laboratories, Newark, CA,
   USA) and visualized with a confocal microscope (LSM 800; Carl Zeiss
   MicroImaging Inc., Oberkochen, Germany) using a 63X objective. Figures
   were analyzed and quantified using Fiji 2 software
   ([60]https://imagej.net/software/fiji/downloads, accessed on 28
   December 2024).

2.6. Apoptosis Assay

   Apoptosis assay was performed using cells after Ficoll treatment and
   evaluated by flow cytometry (FACS) [[61]23]. Briefly, cells were washed
   with PBS 1X and incubated in Annexin buffer 1X with FITC-conjugated
   annexin V and propidium iodide (PI) (Biolegend, San Diego, CA, USA). BD
   FACSDiva v9.0 Software (BD Biosciences, Franklin Lakes, NJ, USA) was
   used for data analysis of Annexin V-positive cells.

2.7. ROS Measurement

   The ROS level was measured with the ROS-sensitive fluorescent probe
   5-(and-6)-chloromethyl-2′,7′-dichlorodihydro-fluorescein
   diacetate-acetoxymethyl ester (DCFDA-AM) [[62]22]. The results are
   expressed as nmol/mg.

2.8. Samples Preparation for HPLC-HRMS Analysis

   Pellets derived from five million cells under different experimental
   conditions were washed twice with PBS buffer. The dry pellet was
   resuspended in 250 µL of MilliQ water and lysed by alternating cold
   (liquid nitrogen) and hot (45 °C) baths. Subsequently, 250 µL of
   acetonitrile (ACN) with 1% formic acid was added, followed by vortexing
   for 2 min. The mixture was then stored at −20 °C for 30 min to
   facilitate protein precipitation. After centrifugation at 13,000 rpm
   for 10 min, 200 µL of the supernatant was collected and transferred to
   a vial for HPLC-MS analysis.

2.9. HPLC-HRMS Metabolomic Analysis

   To investigate the molecular composition of these samples, an HPLC-HRMS
   analysis, through HPLC-timsTOF Pro 2 instrument (Bruker Daltonics GmBH
   & Co., Bremen, Germany), was conducted. The HPLC was equipped with a
   Luna Polar C18 column (2.1 × 150 mm, 3 μm) (Phenomenex, Bologna, Italy)
   and the system was operated in binary gradient mode. Water at 0.1% of
   formic acid was used as solvent A and ACN at 0.1% of formic acid as
   solvent B. The chromatographic gradient was set as follows: flow, 0.3
   mL/min; t: 0 min B: 5%; t: 15 min B: 100%; t: 16 min B: 100%; column
   reconditioning was conducted for 4 min at B: 5%. HRMS instrument was
   coupled to HPLC with a VIP-HESI source with the following parameters
   both in positive and negative ion modes: End Plate Offset of 500 V,
   Capillary Voltage of 4.5 kV; Nebulizer and Dry gas flow rate were,
   respectively, 2 bar and 8 L/min; Dry and Sheath gas temperature were,
   respectively, 230 °C and 400 °C. The mass spectrometer was operated in
   full-scan mode in the range 20–1300 m/z, with a resolution of 30,000 in
   FTMS. DDA tandem mass experiments were performed in the range 20–1300
   m/z for both polarities. Auto MS/MS was used to collect the MS/MS
   spectra with a spectra dynamic control of spectra rate from 16 to 20
   Hz. Collision energy was set at 20 eV for the range of 50–1000 m/z and
   at 30 eV for 1300 m/z. All spectra were acquired in profile mode.
   MetaboScape 2023b software (Bruker Daltonics GmBH & Co., Bremen,
   Germany) was used for data elaboration and calculation.

2.10. Data Pre-Processing and Analysis

   A MetaboScape principal component analysis (PCA) was performed to
   assess the statistical stability and separation of the data groups;
   then the features table was filtered to include only the annotated
   ones. MetaboAnalyst 6.0
   ([63]https://www.metaboanalyst.ca/MetaboAnalyst/ModuleView.xhtml,
   accessed on 28 December 2024) was subsequently used to identify
   significant factors and to highlight the metabolites with differential
   expressions. From these analyses, a heatmap was generated to visualize
   the expression patterns of the metabolites, and pathway enrichment
   analysis was conducted to identify key biological pathways associated
   with the observed metabolic changes.

2.11. Gene Expression Analysis of CML Models

   Gene expression profiling of Imatinib-treated K562 cells or CD34+
   derived from CML cells were retrieved from GEO ([64]GSE120932 and
   [65]GSE198576) [[66]24,[67]25]. In the first dataset, K562-IR
   (K562-derived drug-resistant sub-clone) samples ([68]GSM3421755,
   [69]GSM3421756, [70]GSM3421757) treated with 10µM imatinib and
   untreated standard K562 cell line samples ([71]GSM3421746,
   [72]GSM3421747, [73]GSM3421748) were analyzed by the GEO2R tool. In the
   second dataset, CD34+ derived from CML patients at diagnosis
   ([74]GSM5952552, [75]GSM5952547, [76]GSM5952542) or treated in vitro
   with IM for 72 h ([77]GSM5952545, [78]GSM5952550, [79]GSM5952540) were
   analyzed by the GEO2R tool. Identified significant modulated genes
   obtained by GEO2R (Log2FC > 2 or <−2 and FDR adjusted p-value < 0.001
   for [80]GSE120932, Log2FC > 1 or <−1 and FDR adjusted p-value < 0.05)
   were submitted to gene ontology analysis using the tool EnrichR
   ([81]https://maayanlab.cloud/Enrichr/, accessed on 28 December 2024)
   [[82]26]. The first 10 enriched terms from the ChEA Transcription
   Factor Targets Dataset (ChEA) 2022 were ordered by p-value and reported
   using the Appyter web-tool integrated in EnrichR.

2.12. Statistical Analysis

   Statistical analysis was performed using the paired t-test. All the
   experiments were performed in triplicate and significance was marked as
   follows: * p ≤ 0.05, ** p ≤ 0.01, and *** p ≤ 0.001.

3. Results

3.1. Imatinib Treatment Induces a Significant Increase of Genes Related to
Stemness

   Imatinib treatment completely revolutionized the prognosis of CML even
   if it does not eradicate LSC. We treated K562 and KU812 with different
   concentrations of drug (0.5, 1, and 2 µM) for 48 h to evaluate the
   modulation of stem cell (SC)-related genes that may be responsible for
   IM resistance. IM reduces the proliferation of K562 and KU812 in a dose
   dependent manner and causes a strong decrease in BCR/ABL
   phosphorylation level, as demonstrated by MTT assay and Western blot,
   respectively ([83]Figure 1A,B). Subsequently, we decided to investigate
   the expression of genes related to SC self-renewal, thus SOX2 and OCT
   3/4 on cells extracted after Ficoll density gradient centrifugation. By
   using this method, we can isolate the fraction of cells that survived
   after IM treatment, eliminating dead cells that might interfere with
   gene expression quantification.

Figure 1.

   [84]Figure 1
   [85]Open in a new tab

   Imatinib treatment induces a significant increase of genes related to
   stemness. (A) K562 and KU182 cell lines were treated with different
   concentrations of IM (0, 0.5, 1, 2 μM) to evaluate cell proliferation
   rate. Not treated cells are considered 100% vitality. (B) Western blot
   analysis of BCR/ABL phosphorylation level demonstrates its decrease in
   a dose dependent manner, confirming the activity of TKI. (C,D)
   FITC-Annexin V and PI staining was performed to confirm by FACS
   analysis the vitality of cells after Ficoll stratification. (E,F)
   OCT3/4 and SOX2 gene expression were evaluated through RT-qPCR, after
   48 h of IM treatment cells extracted with Ficoll. Abbreviations: NT:
   not treated; IM: Imatinib; Ann V, Annexin V. * p ≤ 0.05, ** p ≤ 0.01,
   *** p ≤ 0.001.

   After confirming that cells collected after Ficoll stratification are
   not under significant apoptosis, as shown in [86]Figure 1C,D, we
   proceeded with RNA extraction and gene expression analysis. [87]Figure
   1E,F confirmed that in the fraction of alive K562 and KU812, SOX2 and
   OCT 3/4 levels increased significantly after treatment (IM alive),
   suggesting that SC gene levels are crucial even in cells unaffected by
   TKIs.

3.2. Imatinib Treatment Induces Nuclear Localization of Both Sox2 and OCT 3/4
and Is Associated with an Increase in the LSC Phenotype

   Since the mRNAs of the investigated genes related to stem cells
   increased after IM treatment in the fraction of viable cells, we
   decided to assess the levels and localization of the corresponding
   proteins through immunofluorescence analysis. Sox2 exhibited
   predominant nuclear localization, with its levels significantly
   increasing after 48 h of 1 µM IM treatment, as indicated by a sharp
   green nuclear accumulation ([88]Figure 2A) in both the cell lines
   analyzed. A similar pattern was observed for OCT 3/4 ([89]Figure 2B),
   which displayed a significant increase in nuclear localization with a
   spotted distribution, in accordance with its ability to form a complex
   with Sox2 and co-bind to specific DNA sequences.

Figure 2.

   [90]Figure 2
   [91]Open in a new tab

   Imatinib treatment increases protein levels of SOX2 in K562 and KU812.
   (A,B) The green signal corresponding to Sox2 or OCT 3/4 showed a strong
   nuclear increase after IM treatment in alive cells extracted with
   Ficoll gradient. Red propidium (PI) is used to detect nuclei (63X
   magnification). (C) ROS levels in cytosolic extract of cells treated
   with IM, ROS is expressed as nmol/mg of proteins. (D) Heatmap of the
   most significant expressed metabolites in the groups of alive and total
   cells after Imatinib treatment. Significantly reduced activities are
   represented in blue while those with a significantly increased activity
   are in red. K562 IM alive R1, 2, and 3 corresponded to viable cells
   condition triplicates while K562 IM tot R1, 2, and 3 represented the
   total cells after treatment condition triplicates. (E) Pathway
   enrichment analysis obtained with MetaboAnalyst and KEGG for pathway
   identification. The significance of a particular pathway was expressed
   as enrichment ratio. The p-value is represented with a red scale. * p ≤
   0.05, ** p ≤ 0.01.

   Overall, these results confirm the activation of transcription factors
   in the viable pool of cells, a process that could be essential for
   inducing the expression of crucial genes in the stemness compartment.
   Subsequently, since ROS levels are crucial to define an increase of
   stem characteristic, we measured it in cells treated with IM for 48 h
   and in alive cells collected after treatment. ROS levels were
   significantly reduced in cytosolic extracts of viable cells compared to
   the total amount of treated cells ([92]Figure 1E). This result is
   important since low ROS levels are currently used as a parameter to
   isolate LSC, thus supporting the hypothesis that our population of
   cells surviving IM treatment might acquire specific features resembling
   those of resistant stem cells [[93]27].

   Subsequently, to investigate whether the pool of viable cells exhibited
   additional signatures resembling that of stem cells, we performed an
   untargeted metabolomic analysis using the HPLC-timsTOF Pro 2 instrument
   (Bruker Daltonics GmBH & Co., Bremen, Germany). The resulting
   metabolites were analyzed using MetaboAnalyst to identify those
   significantly expressed.

   After conducting an ANOVA test between the two groups, the resulting
   heatmap revealed sharp clustering, suggesting distinct metabolic
   features between the two groups. Interestingly, glutathione emerged as
   one of the most important metabolites in the pool of viable cells,
   alongside L-Palmitoylcarnitine, 3-Hydroxybutyrylcarnitine, and
   adenosine monophosphate (AMP). Their roles in protecting cells from
   oxidative stress, maintaining low ROS levels, fueling energy demands,
   and preserving the cellular metabolic and energetic homeostasis are
   well-known in LSCs.

   Finally, pathway enrichment analysis was conducted with MetaboAnalyst
   and KEGG to identify key biological pathways associated with the
   observed metabolic changes. Interestingly, the identification of
   pathways related to redox homeostasis, energy production through fatty
   acid oxidation, amino acid metabolism, and oxidative metabolism in the
   viable cells strengthened our findings.

3.3. Imatinib Treatment Activate a Few SC Transcription Factors in Cell Lines
and in CD34+ CML Cells After In Vitro Treatment

   To further strengthen our findings, we decided to evaluate a potential
   involvement of SC genes in cells resistant to TKI treatment by
   analyzing a public dataset comparing two groups: K562 cells sensitive
   and resistant to IM [[94]28]. We investigated the pathway enrichment by
   EnrichR and ChEA 2022 as source by using all the 313 genes that emerged
   significantly deregulated. Notably, Sox2 was identified as the top
   transcription factor activated in the resistant cell group ([95]Figure
   3A), as it regulates 82 out of the 313 detected genes ([96]Figure 3B).
   The EnrichR analysis of these differentially expressed genes
   highlighted oxidative stress and redox pathway as among the most
   affected, alongside focal adhesion and TGF-beta signaling.
   Interestingly, CD44 and Cav-1, emerging among the most significantly
   highly expressed, are known to promote quiescence of LSCs and can
   support cell cycle progression through G2/M phase when deleted, thus
   enforcing our analysis [[97]29,[98]30].

Figure 3.

   [99]Figure 3
   [100]Open in a new tab

   Imatinib treatment activate few SC transcription factors in cell lines
   and in CD34+ CML cells after in vitro treatment. (A) Differentially
   expressed genes among K562 sensitive and resistant to IM clusters
   enriched according to Gene Ontology terms and ordered by adjusted
   p-value. (B) Heatmap of all genes modulated by Sox2 emerged after
   comparison between control and Imatinib treatment K562 cells. (C)
   Differentially expressed genes after CD34+ CML cells treatment in vitro
   with IM clusters enriched according to Gene Ontology terms and ordered
   by adjusted p-value. * adjusted p-value.

   In addition, several crucial stem cell factors, including OCT 3/4 and
   Nanog, emerged in another re-analysis of public dataset of CD34+
   derived from CML at diagnosis and treated in vitro with IM for 72 h.
   Even if the number of upregulated genes is reduced compared to cell
   lines specimens, these findings reinforce the link between these
   pathways and TKI treatment in a clinical context, thereby corroborating
   our results ([101]Figure 3C) [[102]25].

4. Discussion

   LSCs are a subpopulation of cells with distinct features, including
   unique cell surfaces antigens, molecular pathways, and metabolism that
   support their expansions [[103]13,[104]31]. Moreover, these cells have
   a remarkable ability to resist elimination by standard TKI treatments
   and may serve as potential source for disease relapse [[105]32]. This
   represents the most significant challenge in current leukemia
   treatment. In this context, data from multiple clinical trials suggest
   that tyrosine kinase inhibitor therapy can be discontinued in
   approximately half of CML patients who achieve a deep molecular
   response, following a program known as therapy-free remission (TFR)
   [[106]33,[107]34]. During this time, the probability of LSC emergence
   and an increase in the bulk of leukemia cells is high. Several studies
   have confirmed the persistence of LCS using highly sensitive technique
   like digital polymerase chain reaction (ddPCR), suggesting that even if
   patients remain in TFR or have activated a program of TKI
   discontinuation, a quote of resilient LSCs can persist for a long time
   [[108]34,[109]35]. These studies highlight the complexity of LSCs and
   their resistance to TKI. Understanding the molecular processes involved
   in LSCs and in their resistance to Imatinib has been the focus of
   several studies in the last decades, and it will be crucial for
   developing more effective targeted therapies and improving outcomes for
   leukemia patients in the near future.

   In our study, we observed that after in vitro treatment of K562 and
   KU812, commonly used CML cell lines, the expression of certain key
   genes, specificallySox2 and OCT 3/4, was significantly increased.
   Interestingly, this increase was particularly noticeable in the
   fractions of samples that remain alive after treatment, suggesting that
   TKIs may activate a program aimed at maintaining the stem cells
   characteristics. Moreover, the sharp nuclear localization of Sox2 and
   OCT 3/4 confirmed the activation of these transcription factors,
   indicating that they could act as a source of disease relapse.

   These genes play a significant role in SC maintenance and self-renewal,
   and their involvement in resistance mechanisms has been explored in
   several studies [[110]36,[111]37]. Indeed, both OCT 3/4 and Sox2 are
   directly involved in Wnt and Notch signaling pathway, contributing to
   self-renewal and resistance of LSCs [[112]36,[113]38].

   The role of SOX2 and OCT 3/4 as prognostic markers in the diagnosis and
   relapse of ALL has been highlighted in a recent studies, in agreement
   with previous research focused on AML specimens [[114]39,[115]40]. High
   levels of these markers are predictors of shorter overall survival (OS)
   and DFS in both ALL and AML patients and are associated with abnormal
   karyotypes, contributing to therapy resistance and higher risk
   stratification [[116]40]. In CML, their reduction can decrease the LSCs
   population and sensitize these cells to targeted therapies
   [[117]41,[118]42]. Although the biological understanding of these
   processes is still limited in the field of leukaemia, they deserve
   further investigations. Targeting these pathways in combination with
   conventional therapies could improve treatment outcomes.

   Furthermore, the identification of low levels of ROS in the pool of
   alive cells after treatment represents an important relevance in
   association with SC genes increase.

   Indeed, SOX2 and OCT 3/4 play a crucial role in maintaining low ROS
   levels by regulating antioxidant defences, reprogramming metabolism,
   and suppressing pro-oxidative pathways [[119]27,[120]43]. This
   capability supports the survival and persistence of stem-like cells,
   protecting them against chemotherapy-induced oxidative damage, thereby
   contributing to drug resistance and relapse. Disrupting the ability of
   SOX2 and OCT 3/4 to regulate ROS levels could sensitize LSCs to
   oxidative stress and therapy, offering a potential treatment strategy.
   Through an untargeted analysis of the metabolites characterizing the
   surviving and the corresponding total cell population after drug
   treatment, we were able to identify several metabolites and pathways
   traceable to LSCs. Among these, glutathione and ubiquinone biosynthesis
   emerged as particularly important [[121]44,[122]45]. Their contribution
   to protecting cells from oxidative stress is critical for LSCs, as they
   help to maintain low ROS levels, supporting the quiescent state of
   LSCs, and preventing DNA damage. L-Palmitoylcarnitine and
   3-Hydroxybutyrylcarnitine were identified as key metabolites involved
   in beta-oxidation and fatty acid catabolism (FAO), essential processes
   that sustain the energy needs of LSCs [[123]46,[124]47].

   Additionally, AMP, an indicator of low energy status, is well known in
   the context of LSCs for its ability to activate AMPK (AMP activated
   protein kinase), which is crucial for maintaining the metabolic and
   energetic homeostasis of LSCs [[125]48,[126]49]. TCA cycle is often
   active in LSCs that rely on oxidative phosphorylation for energy
   production and focusing on this dependency could represent a promising
   approach to eliminate OXPHOS-dependent LSCs [[127]50,[128]51]. Finally,
   the significant number of pathways related to amino acid metabolism
   confirmed their importance in maintaining the stem-like characteristics
   of LSCs by supporting energy production, redox balance, self-renewal,
   and resistance to therapy [[129]52,[130]53].

   Exploring the regulation of these metabolic markers could provide
   valuable targets for the development of clinical therapies aimed at
   contributing to the eradication of LSCs. Subsequently, thanks to an
   in-silico analysis of available datasets on two specific group of
   cells, K562 sensitive and resistant to Imatinib, we observed that among
   the most activated transcription factors in the resistant group there
   is Sox2, even if the different phenotypes that they assume do not match
   perfectly with LSCs characteristics, suggesting that cells can probably
   activate a program with some intermediate traits. Finally, the ChEA
   performed on CD34+ cells derived from CML patients after in vitro
   Imatinib treatment suggest important findings: only a few lists of
   deregulated genes emerge between the two specimens (not treated and
   Imatinib treatment), and the presence of Sox2, Nanog, and OCT 3/4 make
   our assumptions even more meaningful since they are validated on a tool
   of human samples [[131]25,[132]28]. Interestingly, the present ChEA
   analysis asserts that CML CD34+ cells after treatment display a
   potentially selective activation of TP53. The activation of p53
   reflects the ability of cells to induce downstream TP53 effectors like
   BAX, NOXA, and BID and to disrupt anti-apoptotic proteins such as Mcl-1
   and Bcl-2, favoring the induction of apoptosis after TKI treatment
   [[133]54,[134]55]. On the other hand, TP53 seems to be highly expressed
   in LSC, and, for this reason, developing strategies aimed at regulating
   orthogonal pathways, such as those associated with p53, could enhance
   the activity of TKI against CML stem cells [[135]56,[136]57].

5. Conclusions

   All these evidences, including our results, suggest that several
   mechanisms significantly contribute to resistance of LSCs
   [[137]58,[138]59]. In our case, targeting Sox2, OCT 3/4 and their
   downstream pathways may offer potential therapeutic strategies to
   overcome resistance and improve treatment of CML patients. Moreover,
   further research is needed to fully understand the complex regulatory
   mechanisms involving SC features in LSC and their implications in
   treatment, prognosis, and in guiding different patients’ behavior. This
   should include shedding light on individuals that discontinue TKI
   therapy and experience CML recurrence and others that do not could
   contribute to optimizing the best therapeutic strategies.

   Finally, both SOX2 and OCT3/4 could serve as potential markers useful
   for disease monitoring, relapse identification, and stratification of
   patients.

Acknowledgments