Tumor Suppressor Role of INPP4B in Chemoresistant Retinoblastoma

The chemotherapy of retinoblastoma (RB), a malignant ocular childhood disease, is often limited by the development of resistance against commonly used drugs. We identified inositol polyphosphate 4-phosphatase type II (INPP4B) as a differentially regulated gene in etoposide-resistant RB cell lines, potentially involved in the development of RB resistances. INPP4B is controversially discussed as a tumor suppressor and an oncogenic driver in various cancers, but its role in retinoblastoma in general and chemoresistant RB in particular is yet unknown. In the study presented, we investigated the expression of INPP4B in RB cell lines and patients and analyzed the effect of INPP4B overexpression on etoposide resistant RB cell growth in vitro and in vivo. INPP4B mRNA levels were significantly downregulated in RB cells lines compared to the healthy human retina, with even lower expression levels in etoposide-resistant compared to the sensitive cell lines. Besides, a significant increase in INPP4B expression was observed in chemotherapy-treated RB tumor patient samples compared to untreated tumors. INPP4B overexpression in etoposide-resistant RB cells resulted in a significant reduction in cell viability with reduced growth, proliferation, anchorage-independent growth, and in ovo tumor formation. Caspase-3/7-mediated apoptosis was concomitantly increased, suggesting a tumor suppressive role of INPP4B in chemoresistant RB cells. No changes in AKT signaling were discernible, but p-SGK3 levels increased following INPP4B overexpression, indicating a potential regulation of SGK3 signaling in etoposide-resistant RB cells. RNAseq analysis of INPP4B overexpressing, etoposide-resistant RB cell lines revealed differentially regulated genes involved in cancer progression, mirroring observed in vitro and in vivo effects of INPP4B overexpression and strengthening INPP4B's importance for cell growth control and tumorigenicity.


Introduction
Retinoblastoma (RB) is one of the most common malignant ocular diseases in early childhood worldwide [1,2]. Various treatment options provide a patient survival rate of up to 95% [3,4]. Intra-arterial, intravitreal, or intracameral drug injections signifcantly increase eye preservation rates and reduce systemic chemotherapy [5,6]. Tumor treatment is, however, often limited by the massive side efects of the chemotherapeutics. Besides, the development of resistances against drugs of the commonly used VEC (vincristine, etoposide, and carboplatin) therapy ultimately lead to relapses or the emergence of secondary cancers [7]. Tus, new or adjacent RB therapy strategies are needed.
Our group identifed inositol polyphosphate 4phosphatase type II (INPP4B) as a downregulated gene in etoposide-resistant RB cell lines compared to their chemosensitive counterparts. Terefore, we assumed that INPP4B has a pivotal role in the development of RB resistance. A previous study by our group revealed that etoposide resistant RBs behave more aggressively than their chemosensitive counterparts. Tese cells display increased proliferation levels and, thus, higher growth kinetics in vitro and generate signifcantly more and larger tumors in vivo [8]. Other studies have already demonstrated that INPP4B is associated with chemoresistance in acute myeloid leukemia (AML; [9]) and induces chemosensitivity of human hepatocellular carcinoma cells lines [10].
Te structure of INPP4B contains an N-terminal C2 lipid binding domain, an internal nervy homology 2 (NHR2) domain, and a C-terminal catalytic phosphatase domain, and it is one of a plethora of enzymes maintaining homeostasis of phosphoinositides within the cell [11]. A phosphoinositide (PI) is a membrane-bound inositol lipid that acts as a docking site for signaling proteins involved in proliferation, survival, and apoptosis. INPP4B has been reported as a negative regulator of the phosphatidylinositol kinase 3 (PI3K)/AKT signaling pathway [12,13]. PI3K and phosphatidylinositol phosphate kinase 4 (PIP4K) phosphorylate phosphatidylinositol-3 (PI(3)P) and -4 phosphate (PI(4)P), which subsequently phosphorylate and thereby activate AKT, a potent driver of tumorigenic cell growth, which promotes cell proliferation, survival, and migration ( [11,[14][15][16] for review see: [17]). Phosphorylated AKT in turn activates the signal transduction of other downstream molecules in the PI3K/AKT signaling pathway [18][19][20]. Chen et al. demonstrated that INPP4B overexpression in cervical carcinoma cells inhibits the activation of the PI3K pathway by suppressing the phosphorylation of AKT as well as serum and glucocorticoid-regulated kinase-3 (SGK3) [21]. SGK3, another PI3K-dependent serine/threonine kinase, displays high structural and functional similarities with the AKT protein [22].
Te tumor suppressive function of INPP4B is most likely attributable to the negative regulation of PI3K/AKT signaling, whereas its oncogenic function is still unclear, and promotion of SGK3 signaling, inhibition of phosphatase and tensin homolog (PTEN)-dependent AKT activation, and enhancement of DNA repair mechanisms to confer chemoresistance have been proposed as potential mechanisms [17].
Tus, although its function in diferent human cancers remains controversial, INPP4B and the mediated PI3K/AKT downstream signaling pathway seem to play an important role in tumorigenesis and cancer progression. However, its biological role in retinoblastoma yet remain undiscovered.
Tus, in the study presented, we set out to unravel the role of INPP4B in RB chemoresistance by analyzing INPP4B expression in etoposide resistant RB cell lines and chemotherapy-treated RB patient tumors and investigating the efect of lentiviral INPP4B overexpression on etoposideresistant RB cell viability, proliferation, apoptosis, anchorage-independent growth, and tumor formation in vitro and in vivo. Besides, we studied the efect of INPP4B overexpression on the phosphorylation status of the known downstream signaling targets AKT and SGK3 and identifed novel downstream signaling targets in a RNA sequencing analysis of INPP4B overexpressing, etoposide-resistant RB cell lines.  [36], originally purchased from the Leibniz Institute DSMZ (German Collection of Microorganisms and Cell Cultures) as well as the RB cell line RB355 [37], formerly donated by K. Heise, and the corresponding etoposide-resistant RB cell lines Y79-Etop and RB355-Etop. Te cultivation of the above-mentioned cell lines as well as the human embryonic kidney cells (HEK293T) was described in detail previously [38]. No ethical approval was required for work with the human cell lines.
For stable transduction, 1 × 10 6 /0.8 × 10 6 RB cells (RB355/Y79) were seeded in DMEM medium as described previously [41]. Te medium was removed after 24 h, and cells were transfected with INPP4B virus or control virus particles, and 5 μL polybrene (H9268, Sigma-Aldrich, München, Germany) per ml lentivirus were administered. Twice the volume of the virus particles DMEM medium with supplements was added after 24 h. Forty-eight hours later we changed the medium completely and incubated the cells for another 72 h.

RNA Extraction and Quantitative Real-Time PCR.
Te NucleoSpin RNA II Kit (Macherey & Nagel, Düren, Germany) and the miRNeasy Kit (Qiagen, Hilden, Germany) were used for RNA isolations of RB cells. Complementary DNA for quantitative real-time PCR analyses was synthesized with the QuantiTect Reverse Transcription Kit (Qiagen, Hilden, Germany) following the manufacturer´s protocol. A SYBR ™ green PCR assay (Applied Biosystem, Darmstadt, Germany) and the following specifc primers were used to analyze INPP4B expression (see Table 1 Journal of Oncology 3
2.9. Cell Viability Assays. We determined cell viability by seeding 4 × 10 4 cells in 100 μL medium in a 96-well plate in two triplicates. Ten microliters of a water-soluble tetrazolium (WST-1) salt solution (Sigma-Aldrich, München, Germany) were added to each well after 72 h of incubation, and cells were incubated at 37°C for diferent time points. Quantifcation was performed in a microplate reader by measuring the absorbance at 450 nm.

Caspase-Glo 3/7 Assay.
A caspase-Glo 3/7 assay (Promega, Madison, WI, USA) was used to analyze the caspase 3 and 7 cleavage activity after INPP4B overexpression, following the manufacturer´s instructions. Terefore, we seeded 9 × 10 4 INPP4B overexpressing and control cells in 300 μL growth medium supplemented with 2% FBS in a 24-well plate format and incubated them overnight. After 24 h, cell suspensions were mixed 1 : 1 with the caspase-3/7 reagent and seeded in a white 96-well plate for 2 h at room temperature protected from light. Luminescence was measured with an Orion II microplate luminometer (Berthold Detection Systems, Pforzheim, Germany). Measurements were performed three times in fve replicates.

Colony Formation Soft Agarose Assay.
Colony formation capacity was determined in soft agarose assays by suspending 5 × 10 3 INPP4B overexpressing and control RB cells in 2 ml DMEM/F12 medium (Sigma-Aldrich, München, Germany) supplemented with 10% fetal bovine serum, 100 μ penicillin/ml and 100 μg streptomycin/ml, 4 mM L-glutamine, 50 μM β-mercaptoethanol, 10 μg insulin/ml, and 0.7% agarose (Roth, Karlsruhe, Germany). Cell suspensions were layered on 2 ml 1% agarose as described previously [42] and maintained over a period of 3 weeks. Colony formation was quantifed after 3 weeks of incubation, and assays were repeated three times. We determined the number of colonies for each cell line by counting colonies in eight visual felds at a 10x magnifcation in triplicates. A Nikon Eclipse TS2 microscope with a digital camera and IC measure 1.0 software (Nikon, Düsseldorf, Germany) was used to determine colony size and eight colonies per well were surveyed.
2.14. CAM Assays. Te efects of INPP4B overexpression on tumor formation and migration capacity in vivo were studied in the chicken chorioallantoic membrane (CAM) assay. INPP4B overexpressing RB cells and control cells were inoculated on the CAM on embryonic development day (EDD) 10 mainly following the protocols published by [43,44]. Ten eggs were inoculated with 1 × 10 6 cells suspended in 50 μL PBS in at least three independent experiments. At EDD17, grown tumors were excised, measured, weighted, and photographed as described previously [38]. Besides, GFP-labelled etoposide resistant Y79 and RB355 control and INPP4B overexpressing cells were injected into a CAM vene at EDD12 as described previously by our group [38]. Five days after injection, we sacrifced the chicken embryos and collected punches of the ventral CAM, opposing the injection site. Te migrated, GFP-labelled cells were identifed via fuorescence microscopy of the CAM punches.
Whole amount immunofuorescent staining of CAM vessels was described previously by our group [45]. A Nikon ECLIPSE E600 microscope and NIS Elements Imaging 5.20.02 software (Nikon, Düsseldorf, Germany) were used for imaging.

Statistical Analysis.
We performed all assays at least in triplicates and used GraphPad Prism 4 for statistical analyses. Te data represent means ± SEM of three independent experiments from independent RB cell cultures. Te results were analyzed by a Student's t-test and considered signifcantly diferent if p value <0.05 ( * ), p value <0.01 ( * * ), or p value <0.001 ( * * * ). Te growth curve statistics were performed using a free web interface https://bioinf.wehi.edu.au/ software/compareCurves/, which uses the "compare growth curves" function from a statistical modeling package called statmod, available from the "R Project for Statistical Computing:" https://www.r-project.org, previously described elsewhere [46]. All experiments were performed at least in triplicate, and results were depicted with a standard error of the mean (SEM). A student's t-test was used to calculate statistical diferences (ns p > 0.05; * p < 0.05; * * p < 0.01; * * * p < 0.001) between control and experimental groups.

INPP4B Overexpression Induces Caspase Dependent Apoptosis in Etoposide Resistant Y79 and RB355 Cell Lines.
As revealed by DAPI cell counts, INPP4B overexpression resulted in a signifcant increase in apoptosis levels in etoposide resistant Y79 as well RB355 cell lines (Figure 4(a)). Caspase assays revealed that INPP4B likewise signifcantly increased caspase-3/7 activity in both RB cell lines investigated (Figure 4(b)). Additional immunocytochemical stains with an antibody against active, cleaved caspase-3 confrmed an increase in caspase-3 activity, indicated by a higher number of caspase-3 positive cells in INPP4B overexpressing cells. Together, these data indicate that INPP4B overexpression activates caspase-3/7 dependent apoptosis signaling.

INPP4B Overexpression Diminishes Anchorage-Independent Growth of Etoposide-Resistant Y79 and RB355
Cell Lines. We next tested INPP4B overexpressing RB cells for alterations in anchorage-independent growth, known as the capacity of transformed, carcinogenic cells to grow without adherence to a solid surface [47]. Soft agarose assays revealed that both INPP4B overexpressing, etoposide resistant RB cell lines investigated form signifcantly fewer colonies than their parental, chemosensitive counterparts ( Figure 5 . ß-actin served as a loading control and micro manager 1.4 software was used to calculate the relative intensity ratios. All experiments were performed at least in triplicates and results depicted with a standard error of the mean (SEM). A student's t-test was used to calculate statistical diferences ( * p < 0.05; * * * p < 0.001) between control and experimental groups.  reveal that INPP4B overexpression results in a diminished tumorigenic and migratory potential in vivo.

INPP4B Overexpression Induces Phosphorylation of SGK3
but Did Not Afect AKT Signaling. AKT, a serine/threonine kinase, plays an important role in the PI3K signaling pathway and is known to be negatively regulated by INPP4B [13]. It has been shown that in cervical carcinoma cells INPP4B overexpression reduces the phosphorylation of AKT and SGK3, sharing structural and functional similarities [21]. Terefore, we investigated AKT, p-AKT, SGK3, and p-SGK3 expression after INPP4B overexpression in order to reveal the AKT and/or SGK3 pathways as potential INPP4B signaling mechanisms in etoposide-resistant RB cells. In both etoposide-resistant RB cell lines investigated, no signifcant changes in AKT levels were discernible following INPP4B overexpression and p-AKT expression was not consistently altered, being upregulated as well as downregulated in the respective cells (Figure 8(a) and Supplementary Figures 1(A) and 1(B). Total SGK3 levels were decreased, while p-SGK3 levels increased upon INPP4B overexpression (Figure 8(b)). Changes did not reach signifcance ( Supplementary  Figures 1(C) and 1(D), however, indicate that INPP4B potentially triggers the SGK3 signaling pathway in etoposide-resistant RB cells.   (Figure 9(a)), and RB355-Etop cells (Figure 9(b)), with INPP4B being identifed as the most upregulated gene in both cell lines. UMAP analysis after INPP4B overexpression revealed a clear separation of an INPP4B overexpressing group compared to a control group in both etoposide-resistant RB cell lines investigated (Figure 10(a)). It became, however, evident, that the three biological replicates for each cell line themselves were diferent, most likely due to inner experimental variabilities (Figure 10(a)). Tus, subsequently, only genes with the same expression changes in all three biological replicates were included in further downstream analyses in order to exclude nonspecifc gene expression changes.
To unravel INPP4B responsive genes, the mean of the three biological replicates was determined for each cell line and corresponding genes were selected (p < 0.05). Additionally, a minimum fold change (FC) of 1.5 relative to the controls were used as the criterion for an INPP4B responsive gene selection (RGS) cut of. A heatmap of the RGS for both cell lines analyzed is shown in Figures 10(b) and 10(c).
Te RGS identifed were used to perform pathway enrichment analysis in order to identify gene-related functions. DAVID analyses of diferentially expressed genes fltered as RGS revealed four GO-terms with p < 0.05. Signifcantly related GO-terms were "positive regulation of gene expression" and "negative regulation of apoptotic signaling pathway" as well as "cytokine activity" for INPP4B over- pathway connection could be identifed for RGS of both RB cell lines investigated. In order to narrow the list of DEGs, we subsequently reanalyzed the RGS and fltered only genes with a known gene function, excluding pseudogenes and noncoding RNAs. Tereupon, the number of genes signifcantly regulated after INPP4B overexpression was reduced to 16 upregulated and 8 downregulated genes for Y79-Etop cells ( Table 2) and 25 upregulated and 6 downregulated genes for RB355-Etop cells (Table 3).
RNAseq data were validated for selected genes by quantitative real-time PCR, confrming that INPP4B overexpression causes a signifcant increase for all genes initially identifed as upregulated (Figures 11(a) and 11(b), red bars).   Besides, a slight yet not signifcant decrease in expression of HYAL3 and EMP1 in Y79-Etop (Figure 11(a)) and PLK5 in RB355-Etop cells (Figure 11(b)) could be monitored following INPP4B overexpression. Downregulation of GALP (Figure 11(a)) and PTAFR (Figure 11(b)) could not be verifed by Real-time PCR. Our RNAseq data indicated that INPP4B overexpression in etoposide resistant RB cells led to several gene expression changes, potentially related to tumor progression. Especially the connected GO-Term "negative regulation of apoptotic signaling pathway" correlates well with the efects seen after INPP4B overexpression. Additionally, a subset of the highly signifcant regulated genes may be functionally involved in INPP4B-mediated efects on apoptosis and cell growth in vitro and therefore involved in etoposide-resistant RB tumor progression in vivo.

Discussion
INPP4B is a lipid phosphatase known to regulate phosphoinositide 3-kinase (PI3K)/AKT signaling. Originally, INPP4B was described as a tumor suppressor gene in various cancers, but it is now controversially discussed as an oncogenic driver (for review see: [11,17]). However, increased INPP4B expression has been reported for several tumor entities, e.g., AML, melanoma, and colon cancers, suggesting the oncogenic potential of INPP4B (for review see: [17]).
Physiologically, INPP4B is highly expressed in human heart and skeletal muscle tissue [12]. Low INPP4B levels have been reported in various cancers and neoplasms [21,[24][25][26][27]30], suggesting a tumor suppressor function in these tumor entities. INPP4B levels are signifcantly reduced in human hepatocellular, gastric, and gallbladder carcinoma [10,28,29]. In a most recent study, decreased INPP4B expression levels were reported for multiple myeloma cell lines as well as bone marrow plasma of multiple myeloma patients, and lower INPP4B levels correlated with a poor outcome [31]. Fittingly, in the study presented, we demonstrated that compared to the healthy human retina INPP4B mRNA expression levels are signifcantly decreased in RB cell lines, indicating a tumor-suppressing role of INPP4B in retinoblastoma. By contrast, INPP4B levels were upregulated in gallbladder and pancreatic cancer compared with non-tumor tissues [29,48], suggesting an oncogenic role in these tumor entities. Increased INPP4B expression was likewise reported in acute myeloid leukemia, colon cancer, and some melanoma subtypes [32][33][34] supporting the notion of INPP4B as an oncogene [17,35]. Moreover, high INPP4B expression has been reported in gastric cancer patients with large tumors and low to undiferentiated metastasis, which is correlated with a poor prognosis. By contrast, INPP4B expression correlated with a good prognosis in patient with small tumors in a highly to moderately diferentiated metastasis stage [28]. Further along this line, primary nonmetastatic colorectal cancer stemlike cells (CR-CSLCs) display signifcantly reduced INPP4B levels, while they are increased in highly metastatic CR-CSLCs [49]. Based on these data, it has been hypothesized that INPP4B may have seemingly contradictory functions as an oncogenic driver or a tumor suppressor depending on the tumor entity, cancer grade, and clinical stage [29,49].     It has been demonstrated that overexpression of INPP4B induces chemosensitivity in human hepatocellular carcinoma and prostate cancer cells lines [10,50]. Wang et al. likewise demonstrated that INPP4B overexpressing multiple myeloma cells become more sensitive to bortezomib, while INPP4B knockdown cells became more resistant to bortezomib treatment strongly suggesting INPP4B as a key regulator of chemosensitivity [31]. Accordingly, INPP4B overexpression inhibited chemoresistance of primary nonmetastatic CR-CSLCs, but increased chemosensitivity in metastatic CR-CSLCs [49]. In our study, decreased INPP4B expression levels in etoposide resistant compared to chemosensitive RB cell lines and increased levels in chemotherapy-treated RB patient tumors compared to the nontreated likewise strongly suggested an impact of INPP4B on the etiology of etoposide chemoresistance in RB.
In the study presented, INPP4B overexpression reduced proliferation, viability, and growth of etoposide-resistant RB cell lines and concomitantly increased caspase-3/7 mediated apoptosis levels, supporting INPP4B`s role as a tumor suppressor in RB cells. Consistent with our data, it has been demonstrated that INPP4B overexpression inhibits cervical and human hepatocellular carcinoma (HCC) as well as multiple myeloma and acute myeloid leukemia, cell proliferation, and induces caspase-3-mediated apoptosis in HCC cell lines [10,21,31]. Similarly, INPP4B knockdown increased the proliferation of human basal-like breast cancer cells [26]. By contrast, INPP4B downregulation reduced proliferation and increased apoptosis of gastric, pancreatic, and gallbladder cancer cells, while INPP4B overexpression leads to opposing efects [28,29,48]. Further along this line, loss of INPP4B signifcantly inhibited proliferation of NPM1-mutated OCI-AML3 cells [51], and overexpression of INPP4B enhanced proliferation of melanoma cells and melanocytes as well as colon cancer cells, in which INPP4B acts as an oncogenic driver [33,34].
We demonstrated that INPP4B overexpression reduces anchorage-independent growth of etoposide-resistant  [26]. Similarly, knockdown of INPP4B in thyroid, mammary epithelial cell and breast cancer cell lines provided an advantage for anchorageindependent growth [26,27,30]. By contrast, INPP4B overexpression in acute myeloid leukemia cells increased their colony formation potential [32], and INPP4B overexpression likewise led to enhanced anchorageindependent growth in cancer entities, in which INPP4B was identifed as an oncogene, e.g., gallbladder cancer, colon cancer, and acute myeloid leukemia cell lines [29,34,51].
In the study presented INPP4B overexpression in etoposide-resistant RB cells resulted in decreased tumor formation capacity or reduced size of CAM tumors in ovo. Besides, the migration potential was decreased at least in one RB cell line investigated, strengthening the anti-tumorigenic role of INPP4B in retinoblastoma. Fittingly, INPP4B overexpression in cervical cancer and ductal carcinoma cells decreased tumor growth in mice [21,27], and INPP4B knockdown in breast cancer cells increased the number and size of tumors in an athymic murine xenograft model [26]. Moreover, in a genetically-engineered triple-negative breast cancer mouse model INPP4B knockout mice displayed a signifcant, dose-dependent increase in tumor emergence, indicating a tumor suppressor function of INPP4B in these tumor entities [25]. By contrast, INPP4B depletion in melanocytes leads to a delay in tumor development in vivo, suggesting a tumorigenic capacity of INPP4B in this setting [33].
INPP4B has been reported as a negative regulator of PI3K/AKT signaling [13] and was anticipated to act as a tumor suppressor by inhibiting this pathway [27]. Most recently, Wang et al. demonstrated that INPP4B overexpression decreased the phosphorylation of AKT in multiple myeloma and hepatocellular carcinoma cells [31], whereas tumors derived from INPP4B knockout mice were found to be enriched for AKT [25]. It has, however, been shown that in cervical carcinoma cells INPP4B overexpression reduces the phosphorylation of both AKT and SGK3, sharing structural and functional similarities [21]. Terefore, we investigated the phosphorylation status of AKT and SGK3 following INPP4B overexpression in order to reveal the possible involvement of each pathway as potential INPP4B signaling mechanisms in etoposide resistant RB cells. No signifcant changes in AKT or p-AKT levels were discernible following INPP4B overexpression, whereas p-SGK3 levels increased, indicating a potential involvement of the SGK3 signaling pathway in RB etoposide-resistance. In accordance with our data, INPP4B overexpression promoted SGK3 phosphorylation but did not infuence p-AKT levels in AGS gastric cancer cells, while INPP4B reduction elevated AKT phosphorylation, but did not increase p-SGK3 levels in BGC823 gastric cancer cells [28]. Along this line, INPP4B knockdown leads to a reduction in p-SGK3 levels, but did not infuence AKT activation in NPM1-mutated OCI-AML3 acute myeloid leukemia cells [51] and INPP4B expression is not correlated with alterations in AKT phosphorylation in leukemia, suggesting an AKT-independent mechanism [9]. Indeed, INPP4B seems to alternatively signal via SGK3 in cells without canonical AKT signaling [52]. Studies indicated a correlation between high INPP4B expression and SGK3 phosphorylation levels in breast cancer and melanoma cells, in which INPP4B overexpression triggered phosphorylation and activation of SGK3, not AKT. In these cancer cell lines INPP4B signaling, however, increased proliferation and anchorage-independent growth [33,53]  whereas our functional studies indicate a tumor suppressive role of INPP4B in retinoblastoma. INPP4B overexpression in etoposide-resistant RB cells induced changes in gene regulation revealed by RNAseq analysis. GO-term analysis of the RGS revealed connections to apoptosis (GO-term "negative regulation of apoptotic signaling pathway") and gene expression regulation (GO-term "gene expression regulation") as well as cytokine activity and cell adhesion. Tese data are in line with the results of our in vitro experiments which showed induced caspase dependent apoptosis levels following INPP4B overexpression in both etoposide-resistant cell lines investigated. A main difference between the two RB cells lines investigated is their growth behavior and fttingly, the signifcant GO-term "cell adhesion" is connected to RB355 cells, growing as an adherent culture. Tumor cell adhesion to the extracellular matrix is an important facilitator of therapy resistance. It could be shown that the cell adhesion resistome is involved in the homeostasis of cancer cells and fundamentally contributes to adaptation mechanisms, including survival and growth, induced by molecular drugs [54]. Te observation that diferent DEGs have been identifed for the two RB cell lines investigated is most likely attributable to the fact that Y79 and RB355 retinoblastoma cells are diferent per se, one growing as suspension cells (Y79) and one, as mentioned above, as an adherent culture (RB355).
Interestingly, some of the verifed upregulated genes of the identifed RGS seem to trigger the oncogenic role of INPP4B, which is not refected by our functional data. However, mirroring previously described efects of an increased phosphorylation of SGK3 seen after INPP4B overexpression in both RB cell lines. In this context, INPP4B has already been described as an oncogenic driver through phosphorylation and activation of SGK3 in a subset of melanoma and colon carcinoma [33,34].
One of the upregulated genes potentially driving the oncogenic role of INPP4B in RB is the macrophage-capping protein (CAPG) which raised the expression of CAPG was likewise shown in diferent metastatic cancers, supporting its involvement in tumor cell invasion and metastatic processes [55,56]. Increased CAPG expression has been correlated with elevated invasiveness and migration in several human tumor entities like, e.g., glioblastoma [57]. Besides, increased CAPG expression strongly correlates with the resistance to paclitaxel chemotherapy [58], and knockdown of the circular RNA circ_0055412 promotes the cisplatin sensitivity of glioma cells through modulation of the CAPG signaling pathway [59]. Finally, elevated CAPG expression is correlated with unfavorable clinical parameters and poor patients`outcomes in diferent cancers, suggesting a potential role as a biomarker for prognosis and prediction of therapy outcome [60][61][62].
Te hematopoietic cell signal transducer HCST, also upregulated in INPP4B overexpressing etoposide resistant Y79 RB cells, has been suggested as a potential biomarker for renal cell carcinoma and lung cancer diagnosis and prognosis [63,64]. High HCST expression leads to signifcant enrichment in cell adhesion, tumor formation, and immune and infammatory responses in a renal cell carcinoma specimen [63].
Additionally, we identifed an upregulation of the tumor necrosis factor superfamily member 4 (TNFSF4) after INPP4B overexpression in etoposide-resistant RB355 cells. Higher levels of TNFSF4 were likewise detected in breast and bladder carcinoma as well as in serum and tumor tissues of lung adenocarcinoma patients, and it has been shown that stressinduced induction of TNFSF4 in cancer-associated fbroblasts alleviates the resistance of lung adenocarcinomas against chemotherapeutics by inhibiting tumor cell apoptosis [65][66][67].
High levels of the calcium-binding protein 1 (CABP1), as detected after INPP4B overexpression in etoposide-resistant RB355 cells, were likewise revealed by Kaplan-Meier analysis of glioblastoma hub genes and was negatively associated with relapse-free survival of glioblastoma patients [68]. Besides, CABP1 was identifed as one of 5 key prognostic genes for predicting the survival of invasive lobular breast cancer survival [69]. Additionally, CABP1 can adjust the activity of inositol 1,4,5-triphosphate receptors in a calciumdependent manner [70]. Regardless of the described functions, up to now CABP1 has seldom been investigated in the context of oncological research.
Functions mediated by the verifed upregulated interferon-induced protein with tetratricopeptide repeats (IFIT) and the downregulated epithelial membrane protein 1 (EMP1) are in line with the tumor suppressive function of INPP4B seen in RB cells. Members of the IFIT genes have been shown to promote drug resistance after depletion and are negatively associated with tumor malignancy due to proapoptotic efects and the activation of caspase-3 after overexpression (for review see [71]). Besides, their potential use as cancer biomarkers and prognostic factors as well as novel therapeutic targets for cancer therapy has been discussed (for review see [71,72]. EMP1 is a transmembrane glycoprotein involved in oncogenic processes like proliferation, migration, invasion, metastasis, and malignant progression [73][74][75]. In a previous study by our group, we demonstrated that EMP1 knockdown in RB cells signifcantly reduces cell viability and proliferation and increases apoptosis [76]. Real-time PCR verifcation of our RNAseq data, revealing a downregulation of EMP1 in INPP4B overexpressing, etoposide-resistant Y79 cells, did, however, not reach signifcance.
Nevertheless, further functional experiments will be required to unravel the interplay between INPP4B and the identifed diferentially expressed genes in terms of tumor suppressive as well as oncogene-like impacts.

Data Availability
Te RNAseq data used to support the fndings of this study have been deposited in the Gene Expression Omnibus (GEO) database (GSE226408).

Disclosure
Nicole Dünker and Maike Anna Busch shared last authorship.

Conflicts of Interest
Te authors declare that there are no conficts of interest.