Methods of characterizing, selecting a patient for treatment for, or treating liposarcoma

Abstract

Described herein is a method of characterizing a subject's liposarcoma (LPS). Methods of selecting a subject with LPS for chemotherapy and methods of treatment are also described.

Description

[0001] METHODS OF CHARACTERIZING, SELECTING A PATIENT FOR TREATMENT FOR, OR TREATING LIPOSARCOMA CROSS-REFERENCE TO RELATED APPLICATIONS

[0002] [1] This application claims benefit of U.S. Provisional Application No. 63 / 666,407 filed July 1, 2024 and U.S. Provisional Application No. 63 / 746,467 filed January 17, 2025, the entire contents of each is incorporated herein by reference.

[0003] SEQUENCE LISTING

[0004] [2] This application contains a sequence listing that has been submitted electronically as an .xml file with name “3451W01WO_Sequence Listing” on 1 July 2025 and incorporated herein by reference in its entirety. The .xml file was generated on June 16, 2025 and is 16,384 bytes in size. The entire contents of the Sequence Listing are herein incorporated by reference.

[0005] FIELD

[0006] [3] Identified herein are differences in gene expression between well-differentiated liposarcoma (WDLPS) and de-differentiated liposarcoma (DDLPS) tumors. In particular, IGF1 autocrine signaling loss is identified as a key feature of liposarcoma (LPS) dedifferentiation. These molecular insights provide the basis for methods that are better at distinguishing WDLPS from DDLPS, compared to relying on tumor morphology alone. The methods can also be used to optimize patient treatment strategies and provide therapeutic interventions that specifically target LPSs, for example DDLPS.

[0007] BACKGROUND

[0008] [4] Soft tissue sarcomas (STS) are a collection of mesenchymal neoplasms likely derived from aberrant stem cell differentiation. Almost half of patients with STS will develop metastatic disease, where the median overall survival is 12-18 months. Unlike many other cancer types, outcomes for patients with STS have not improved in over 40 years. Due to the rarity of each subtype and lack of representative model systems, the molecular drivers of sarcomagenesis remain poorly understood. [5] Liposarcoma (LPS) is the most common type of STS. There are different subtypes that are distinct in their pathology and clinical outcome: well-differentiated liposarcoma (WDLPS), de-differentiated liposarcoma (DDLPS), myxoid / round-cell liposarcoma (MRCLPS), and pleomorphic liposarcoma (PLS). LPS serves as an informative model to investigate dysregulated differentiation given its well and de-differentiated subtypes. The only curative therapy for localized LPS is complete surgical resection. Surgical resection, however, is not always a treatment option. Radiation and conventional cytotoxic chemotherapy are used to treat metastatic or unresectable LPS.

[0009] [6] WDLPS and DDLPS are the most common subtypes of LPS. WDLPS is typically low grade, resembles adipose tissue, and has little metastatic potential. Watchful waiting is often the recommended course of action for WDLPS patients, as surgical resection can be difficult. Unfortunately, in approximately 25% of diagnosed WDLPS cases, a DDLPS component emerges, resulting in a 6-fold increased risk of death for these patients. A DDLPS component is determined by the morphology of the cells, mitotic rate, pleomorphism, and cellular density. If a DDLPS component is present in an otherwise WDLPS tumor, the LPS is considered to be a mixed LPS. DDLPS resembles fibroblasts, is more aggressive, and has a high metastatic potential. DDLPS patients have a 15-20% risk of metastasis and a 5-year survival rate of only about 30%. Genomic analysis of paired WDLPS and DDLPS patient samples to date has not revealed any universal genomic event to explain the transition from WDLPS to DDLPS.

[0010] [7] Therefore, a need remains to identify markers that can be used to monitor the progression from low-grade to aggressive subtypes of LPS. Moreover, the identification of such markers may also uncover new and more effective treatment options, in particular for DDLPS.

[0011] SUMMARY

[0012] [8] Liposarcoma (LPS) derives from an adipocyte precursor cell. In normal adipose tissue, IGF1 signaling is important in adipogenesis and the maintenance of normal adipocyte function. It acts in an autocrine fashion to upregulate Peroxisome Proliferator Activated Receptor Gamma (PPARG) and increase expression of downstream markers of adipocyte terminal differentiation. Mature adipocytes show relatively high levels of Insulin-like Growth Factor 1 (IGF 1) and PPARG (e.g., PPARG1 and / or PPARG2) as well as downstream markers of differentiation such as FABP4 and LPL, and relatively low levels of Insulin-like Growth Factor 1 Receptor (IGF1R). It has now been discovered that well-differentiated liposarcoma (WDLPS) tumor cells show a pattern of expression of these markers similar to that of mature adipocytes. In contrast, de-differentiated liposarcoma (DDLPS) tumor cells show almost no relative expression of IGF1, and downstream markers Fatty Acid Binding Protein 4 (FABP4) or Lipoprotein Lipase (LPL). They retain reduced expression of PPARG and have high relative expression of IGF1R.

[0013] [9] Accordingly, a method of characterizing a subject’s LPS is provided, comprising determining expression levels of IGF1, IGF1R and / or PPARG (e.g., PPARG1 and / or PPARG2) in a sample of the LPS obtained from the subject.

[0014]

[0010] It has also been discovered that loss of IGF1 expression in LPS tumor cells causes de-differentiation of LPS and ultimately increases its metastatic potential. Accordingly, also provided is a method of identifying a subject suffering from a LPS with metastatic potential, the method comprising determining expression levels of IGF1, IGF1R and / or PPARG (e.g., PPARG1 and / or PPARG2) in a sample of the LPS obtained from the subject.

[0015]

[0011] In some instances, such methods further comprise determining the expression levels of LPL and / or FABP4. For example, expression levels of the following markers may be determined in the sample: (i) IGF1R, PPARG (e.g., PPARG1 and / or PPARG2), and LPL; and / or (ii) IGF1, FABP4, and LPL.

[0016]

[0012] The expression levels of these markers can also be used to guide therapeutic decisions. Accordingly, in some aspects, a method of treating a subject suffering from a LPS is provided. The method may comprise administering to the subject a therapeutically effective amount of a chemotherapy, wherein the subject is selected for chemotherapy if the LPS has: (a) a 2-fold or more increase in the expression of IGF1R in the sample obtained from the subject compared to a control; and / or (b) a 50% or more decrease in the expression of PPARG (e.g., PPARG1 and / or PPARG2) and / or IGF1 in the sample obtained from the subject compared to the control sample (e.g., normal adipose tissue).

[0017]

[0013] A method of selecting a subject with LPS for chemotherapy is also provided, comprising determining expression levels of IGF1, IGF1R and / or PPARG (e.g., PPARG1 and / or PPARG2) in a sample of the LPS obtained from the subject compared to a control sample; wherein the subject is selected for chemotherapy if the sample has: (a) a 2-fold or more increase in the expression of IGF1R in the sample obtained from the subject compared to the control; and / or (b) a 50% or more decrease in the expression of PPARG and / or IGF1 in the sample obtained from the subject compared to the control sample (e.g., normal adipose tissue).

[0018]

[0014] In some instances, a method of treating a subject suffering from a LPS with chemotherapy or selecting a subject with LPS for chemotherapy further comprises determining the expression levels of LPL and / or FABP4. For example, the subject may be selected for chemotherapy if the sample has a 50% or more decrease in the expression of FABP4 and / or LPL in the sample obtained from the subject compared to the control sample.

[0019]

[0015] As the transition from WDLPS to DDLPS is associated with a loss of IGF 1 expression and an increase in metastatic potential, it is believed that restoring IGF1 in de-differentiated tumor cells may result in a less aggressive form of LPS. Accordingly, methods of treating subjects with LPS are provided that are aimed at restoring IGF1 signaling in the LPS. Various therapeutic agents are available that can be used to restore IGF1 signaling. These include PPAR-y agonists and recombinant human IGF1 and IGF1 mimetics. Accordingly, in some aspects, a method of treating a subject with LPS is provided, comprising administering a therapeutically effective amount of a selective PPAR-y agonist. Also provided is a method of treating a subject with LPS comprising administering a therapeutically effective amount of recombinant human IGF1 or an IGF1 mimetic.

[0020]

[0016] IGF1R can be targeted to selectively kill LPS cells, e.g., using antibody-drug conjugates (ADCs). For example, a loss of the autocrine IGF1 signaling loop may result in the overexpression of IGF1R in DDLPS tumor cells. Accordingly, in further aspects, a method of treating a subject with LPS is provided, comprising administering a therapeutically effective amount of an ADC, wherein the antibody specifically binds to the extracellular region of IGF1R.

[0021]

[0017] Also provided is a method of treating a subject with LPS comprising administering a therapeutically effective amount of a selective IGF1R kinase inhibitor.

[0022]

[0018] The different treatment modalities set out in the above methods of treating a subject with LPS may be combined, e.g., to improve efficacy or enhance treatment responses. BRIEF DESCRIPTION OF THE DRAWINGS

[0023]

[0019] Drawings are for illustration purposes only.

[0024]

[0020] FIG. 1 shows a representative uniform manifold and approximation projection (UMAP) showing clustering of 164,229 cells, labelled by type, in a single nucleus RNA sequencing (snRNA-seq) analysis cohort. Unlabeled clusters are composed of tumor cells.

[0025]

[0021] FIG. 2A shows a representative UMAP showing clustering of 80,881 tumor cells obtained from WDLPS and DDLPS patients that are part of the snRNA-seq analysis cohort of FIG 1.

[0026]

[0022] FIG. 2B shows a representative UMAP showing the cells as mapped according to FIG. 2A that express usage program 1 “Early developmental” (see Example 2 and Table 2).

[0027]

[0023] FIG. 2C shows a representative UMAP showing the cells as mapped according to FIG. 2A that express usage program 3 “Insulin-mediated adipogenesis” (see Example 2 and Table 2).

[0028]

[0024] FIG. 2D shows representative stacked violin plots showing the level of each usage program score by pathology, i.e., WDLPS and DDLPS.

[0029]

[0025] FIG. 3 A shows a representative UMAP showing clustering of 15,995 cells, labelled by type, in a single nucleus ATAC sequencing (snATAC-seq) analysis cohort.

[0030]

[0026] FIG. 3B shows a representative bar chart obtained by performing Genomic Regions Enrichment of Annotations (GREAT) analysis, showing the enrichment of peaks associated with different signaling pathways in WDLPS tumor cells.

[0031]

[0027] FIG. 3C shows a representative bar chart obtained by GREAT analysis, showing the enrichment of peaks associated with different signaling pathways in DDLPS tumor cells.

[0032]

[0028] FIG. 3D shows a representative plot showing differentially accessible peaks in the IGF1 gene body in WDLPS tumor cells relative to DDLPS tumor cells. The identified peaks in the IGF1 gene body in WDLPS tumor cells are located at candidate response elements (CREs), as indicated.

[0033]

[0029] FIG. 4A shows a representative dot plot that compares the fraction of cells in DDLPS, WDLPS, and normal adipocytes that express IGF1, IGF1R1, PPARG, FABP4, and LPL, and the mean expression of each gene in each group.

[0030] FIG. 4B shows representative violin plots showing the level of an IGF1 signature score across normal adipose, WDLPS tumor cells, and DDLPS tumor cells, showing that DDLPS tumor cells comprise two populations, one with a higher IGF1 signature score that is comparable to that observed for the WDLPS tumor cells, and another with a lower IGF1 signature score.

[0034]

[0031] FIG. 4C shows a light microscopy image of a LPS tumor sample stained with haematoxylin and eosin. The transition zone between WDLPS and DDLPS is shown by a dashed line.

[0035]

[0032] FIG. 4D shows spatial transcriptomic analysis overlaid on the light microscopy image of FIG. 4C. Consensus non-negative matrix factorization (cNMF) programs are indicated. Light gray signifies a WDLPS -associated program, i.e., usage program 3 “Insulin-mediated adipogenesis”, and dark gray signifies a DDLPS -associated program, i.e., usage program 1 “Early developmental” (see Example 2 and Table 2).

[0036]

[0033] FIG. 5 shows a representative bar graph showing expression of adipocyte differentiation markers PPARG and adiponectin in LPS6 cancer cells after exogenous IGF1 treatment.

[0037]

[0034] FIG. 6 A shows a representative bar graph showing expression of PPARG 1 , PPARG2, and ADIPOQ in mesenchymal stem cells (MSCs) before (control, “C”, left bars) and after exposure to IGF 1+DM (“DM”, right bars), determined by qPCR. Gene expression levels are fold change as determined by the delta-delta CT method, normalized to beta-actin and relative to control for each cell line. Statistical significance was determined by Student’s t-test. The asterisk (*) indicates p < 0.05.

[0038]

[0035] FIG. 6B shows a representative bar plot showing expression of PPARG1, PPARG2, and ADIPOQ in LPS6 cells before (control, “C”, left bars) and after IGF 1+DM exposure (“DM”, right bars). Gene expression levels are fold change as determined by the delta-delta CT method, normalized to beta-actin and relative to control for each cell line. Statistical significance was determined by Student’s t-test. The asterisk (*) indicates p < 0.05.

[0039]

[0036] FIG. 7A shows a representative plot showing expression of PPARG1 in human adipose, WDLPS, and DDLPS tissue samples sequenced in multiome cohort, augmented by 7 additional patient samples. Each dot represents a sample from an individual patient. Statistical significance was defined as p < 0.05 by Kruskal Wallis, indicated with as asterisk (*) in the figure; ns = not significant.

[0040]

[0037] FIG. 7B shows a representative expression of PPARG2 (bottom) in human adipose, WDLPS, and DDLPS. Each dot represents a tissue sample from an individual patient. PPARG2 transcript was not detected in one WDLPS and two DDLPS samples. Statistical significance was defined as p < 0.05 by Kruskal Wallis, indicated with as asterisk (*) in the figure; ns = not significant.

[0041]

[0038] FIG. 8 shows a representative coverage plot over selected region of PPARG gene body containing differentially accessible peaks (DAPS, tall light gray bars shown in DDLPS panel) in the snATAC-seq multiome data. Top panel shows normalized signal for WDLPS cells, DDLPS shows normalized signal for DDLPS cells. Peaks deemed significantly differentially accessible between WDLPS and DDLPS cells by logistic regression, adjusted for sequencing depth. Statistical significance was defined as adjusted p value < 0.05. The promoter-like signature region (‘Promoter’) is denoted by the block shown in the “Promoter” panel. PPARG2 transcription start site (TSS) is indicated by a black arrow.

[0042]

[0039] FIG. 9 A shows a representative bar graph showing expression of PPARG 1 , PPARG2, and ADIPOQ in mesenchymal stem cells (MSCs) when exposed to IGF 1+DM alone (left bars) and when exposed to IGF 1+DM and rosiglitazone (right bars), as determined by qPCR. Gene expression levels are fold changes as determined by the delta-delta CT method, normalized to beta-actin and relative to control for each cell line. Statistical significance was determined by Student’s t-test. The asterisk (*) indicates p < 0.05.

[0043]

[0040] FIG. 9B shows a representative bar plot showing expression of PPARG1, PPARG2, and ADIPOQ in LPS6 cells when exposed to IGF 1+DM alone (left bars) and when exposed to IGF1+DM and rosiglitazone (right bars), determined by qPCR. Gene expression levels are fold change as determined by the delta-delta CT method, normalized to beta-actin and relative to control for each cell line. Statistical significance was determined by Student’s t-test. The asterisk (*) indicates p < 0.05.

[0044]

[0041] FIG. 10A shows a representative bar graph showing expression of PPARG1, PPARG2, and ADIPOQ in mesenchymal stem cells (MSCs) when exposed to IGF1+DM alone (left bars) and when exposed to IGF 1+DM and liraglutide (right bars), determined by qPCR. Gene expression levels are fold change as determined by the delta-delta CT method, normalized to beta-actin and relative to control for each cell line. Statistical significance was determined by Student’s t-test. The asterisk (*) indicates p < 0.05.

[0045]

[0042] FIG. 10B shows a representative bar plot showing expression of PPARG1, PPARG2, and ADIPOQ in LPS6 cells when exposed to IGF 1+DM alone (left bars) and when exposed to IGF 1+DM and liraglutide (right bars), determined by qPCR. Gene expression levels are fold change as determined by the delta-delta CT method, normalized to beta-actin and relative to control for each cell line. Statistical significance was determined by Student’s t-test. The asterisk (*) indicates p < 0.05.

[0046]

[0043] FIG. 11A shows a representative plot showing the percentage viability of 93T449 cells when treated with various concentrations of an IGF 1R- ADC (gray circles) or an isotype control ADC (black squares). Cell viability was determined by XTT assay, setting 0 pg / ml isotype control as 100% viability. Points on the graph and error bars show means + / - standard deviation.

[0047]

[0044] FIG. 1 IB shows a representative plot showing the percentage viability of a control cell line, i.e., human aortic endothelial cells (HAECs), for comparison to FIG. 9A. The experimental set-up was the same as for FIG. 9A, except for the cells treated. Points on the graph and error bars show mean + / - standard deviation.

[0048]

[0045] FIG. 12 shows representative Kaplan-Meier analysis curves showing overall survival of LPS patients when delineating the data based on IGF1 expression. Patient data was obtained from TCGA bulk RNA-seq data comprising 58 patients with DDLPS and one patient with WDLPS. “High” indicates patients having tumor samples with IGF1 expression greater than the median expression value for the cohort. “Low” indicates patients having tumor samples with IGF1 expression less than the median expression value for the cohort. Number at risk is shown in the panel beneath the Kaplan-Meier analysis curves. Statistical significance was determined by log-rank test.

[0049] DETAILED DESCRIPTION

[0050]

[0046] In order for the following description to be more readily understood, certain terms are first defined below. Additional definitions may be set forth throughout the specification.

[0051]

[0047] Unless otherwise required by context, singular terms shall include pluralities and plural terms shall include the singular. Thus, as used in this specification and the appended claims, the singular forms “a,” “an” and “the” include plural referents unless the context clearly dictates otherwise.

[0052]

[0048] Unless specifically stated or obvious from context, as used herein, the term “or” is understood to be inclusive and covers both “or” and “and.” Furthermore, “and / or” where used herein is to be taken as specific disclosure of each of the two specified features or components with or without the other. Thus, the term “and / or” as used in a phrase such as “A and / or B” herein is intended to include “A and B”, “A or B”, “A” (alone), and “B” (alone). Likewise, the term “and / or” as used in a phrase such as “A, B, and / or C” is intended to include “A and / or B and / or C” and to thus encompass each of the following aspects: A, B, and C; A, B, or C; A or C; A or B; B or C; A and C; A and B; B and C; A (alone); B (alone); and C (alone).

[0053]

[0049] It is understood that wherever aspects are described herein with the language “comprising,” otherwise analogous aspects described in terms of “consisting of’ and / or “consisting essentially of’ are also provided. In other words, if an aspect is described as comprising A, B and C, aspects consisting essentially of A, B and C are also contemplated, as are aspects consisting of A, B and C.

[0054]

[0050] The term “about” refers to an interval of accuracy that a person skilled in the art will understand to still ensure the technical effect of the feature in question. The term indicates a deviation from the indicated numerical value of ±10%, ±5%, or ±1% of the indicated numerical value.

[0055]

[0051] The term “determining expression levels" or grammatical variants thereof can refer to the quantification of the amount of the mRNA or protein of a marker described herein in a LPS sample relative to a suitable control sample. For example, the level may be indicated as “low” or “high” relative to the control sample. In some instances, the term may be used to refer to determining the presence (+) or absence (-) of expression of a marker described herein in a LPS sample relative to a control without further quantification. For example, the expression of IGF 1 or PPARG (e.g., PPARG2) may be substantially reduced in DDLPS such that, depending on the method used for detection of the mRNA or protein, levels of expression may be below the limit of detection in a DDLPS sample relative to a control sample such adipose tissue, such that expression is indicated as absent (-). Thus, in some instances, the term “determining expression levels” may be used interchangeably with the term “determining expression”.

[0052] The term “in vitro’’’ refers to events that occur in an artificial environment, e.g., in a test tube or reaction vessel, or in a cell in cell culture, etc., rather than within a multi-cellular organism.

[0056]

[0053] The term “in vivo” refers to events that occur within a multi-cellular organism, such as a human or a non-human animal.

[0057]

[0054] The term “marker” refers to a differentially expressed gene that can be used to distinguish between well-differentiated subtypes of LPS with low metastatic potential (e.g., WDLPS) and de-differentiated subtypes with high metastatic potential (e.g., DDLPS tumor cells). As explained in more detail below, levels of expression can be determined either by detecting mRNA expressed from the gene or by detecting the encoded protein. In some instances, it may also be useful to determine the activation state of a gene regulatory region to determine expression of the marker in sample. For example, if the gene regulatory region is in an inactive state, it can be inferred that no expression of the marker whose expression is controlled by the gene regulatory region can be detected.

[0058]

[0055] Unless otherwise defined herein, technical and scientific terms used herein have the same meaning as commonly used and / or understood by one of ordinary skill in the art to which this application belongs. In case of conflict, the present specification, including definitions, will control.

[0059]

[0056] Generally, techniques of cell and tissue culture, molecular biology, virology, immunology, microbiology, genetics, analytical chemistry, synthetic organic chemistry, medicinal and pharmaceutical chemistry, and protein and nucleic acid chemistry and hybridization described herein are those well-known and commonly used in the art. Enzymatic reactions and purification techniques are performed according to manufacturer’s specifications, as commonly accomplished in the art or as described herein.

[0060] Diagnostic methods

[0061]

[0057] Though both WDLPS and DDLPS usually exhibit the same amplification of chromosome 12ql3-l including the MDM2 gene, their histopathologies and clinical behaviors are distinct. Further, to date, genomic analysis of paired WDLPS and DDLPS patient samples has not revealed any universal genomic event to explain the transition from WDLPS to DDLPS.

[0058] Differences have now been discovered in gene expression between well-differentiated liposarcoma (WDLPS) and de-differentiated liposarcoma (DDLPS) tumors. These differences can be used to characterize a liposarcoma (LPS) based on a defined set of markers. This represents an improvement over histopathological characterization, which can be subjective, whereas assessing increased or reduced expression of a set of markers can provide an objective measurement. Moreover, as shown herein, the identified markers may provide a more reliable measurement to determine the transition of a LPS to a subtype with greater metastatic potential. Specifically, the disclosed markers can be used to determine if a phenotypic WDLPS tumor may be progressing to a DDLPS tumor, which is associated with worsened patient outcomes. Accordingly, methods are provided for diagnosing a subject’s LPS that determine the expression levels of markers associated with a DDLPS phenotype.

[0062]

[0059] In particular, two panels of differentially expressed genes are identified herein that can be used to distinguish between WDLPS and DDLPS tumor cells. The first panel of genes (referred to as “Program 1” in Table 2) is characteristic of WDLPS tumor cells. The genes in the first panel express markers associated with early developmental pathways, including R0B01, R0B02, and Insulin-like Growth Factor 1 Receptor (IGF1R). The second panel of genes (referred to as “Program 3” in Table 2) is characteristic of DDLPS tumor cells. The genes of the second panel express markers associated with adipocyte differentiation and insulin / RTK signaling including Peroxisome Proliferator Activated Receptor Gamma (PPARG) and Insulin-like Growth Factor 1 (IGF1). Markers characteristic of the first panel and / or the second panel can be used to characterize a subject’s LPS.

[0063]

[0060] Thus, in one aspect, provided herein is a method of characterizing a subject’s LPS comprising determining expression levels of one or more markers that are associated with the first panel of genes and / or the second panel of genes in a sample of the LPS obtained from the subject. In another aspect, provided herein is a method of identifying a subject suffering from a LPS with metastatic potential, the method comprising determining expression levels of one or more markers that are associated with the first panel of genes and / or the second panel of genes in a sample of the LPS obtained from the subject.

[0061] Specific markers selected from the first and second panels of genes include one or more of IGF1R, PPARG, Lipoprotein Lipase (LPL), IGF1, and Fatty Acid Binding Protein 4 (FABP4).

[0064]

[0062] There are two PPARG isoforms: PPARG1 and PPARG2. These may also be interchangeably referred to herein as PPAR-yl and PPAR-y2, respectively.

[0065]

[0063] The expression level of one or more of the markers may be determined. For example, expression levels of IGF1, IGF1R, or PPARG (e g., PPARG1 and / or PPARG2) may be determined. More typically, the expression level of two or more (e.g., three, four, or five or more) of the markers is determined. For example, the expression level of at least three of the markers may be determined. The markers may comprise (i) IGF1R, PPARG (e.g., PPARG1 and / or PPARG2), and LPL; and / or (ii) IGF1, FABP4, and LPL. For instance, expression levels of IGF1, IGF1R, and PPARG (e.g., PPARG1 and / or PPARG2) may be determined.

[0066]

[0064] In particular, provided herein is a method of characterizing a subject’s LPS (e.g., to assess its metastatic potential) that comprises determining expression levels of IGF 1, IGF1R and / or (e.g., PPARG1 and / or PPARG2) in a sample of the LPS obtained from the subject.

[0067]

[0065] In some aspects, a method of characterizing a subject’s LPS is provided that comprises determining expression levels of the following markers in a sample of the LPS obtained from the subject: (i) IGF1R, PPARG, and LPL) and / or (ii) IGF1, FABP4, and LPL.

[0068]

[0066] As shown herein, a PPARG2 gene regulatory region (i.e., 12351171-12351492 on chr3) is an inactive conformation (i.e., is inaccessible chromatin) in DDLPS tumors. Conversely, the same PPARG2 gene regulatory region is in an active conformation (i.e., is accessible chromatin) in WDLPS tumors. Accordingly, also provided herein is a method of characterizing a subject’s LPS that comprises determining the activation state of a PPARG2 gene regulatory region (e.g., nucleotides 12351171-12351492 on chromosome 3) in a LPS sample obtained from the subject, wherein (i) an inactive state of the gene regulatory region indicates that the LPS is dedifferentiated, i.e., DDLPS, and (ii) an active state of the gene regulatory region indicates that the LPS is well-differentiated, i.e., WDLPS.

[0069] Sample

[0070]

[0067] Typically, the sample for use with the diagnostic methods described herein is a resected tumor or a tissue biopsy. mRNAs can be isolated from the sample using standard protocols to determine expression levels of the markers described herein. Alternatively, the sample may be processed, e.g., for immunohistological analysis to detect the encoded protein, or for RNA in situ hybridization. For example, the sample obtained from the subject may be formalin-fixed and, if desired, paraffin-embedded prior to immunohistological analysis using a set of antibodies that specifically bind to the markers described herein. Alternatively, the formalin-fixed, paraffin-embedded sample may be analyzed using one or more nucleic acid probes that specifically hybridize to the markers described herein. In some instances, the sample may be divided for analysis of expression levels using both standard histological analysis, i.e., nucleic acid analysis, and immunohistological assays.

[0071] Control

[0072]

[0068] The described diagnostic methods typically compare a subject’s LPS against one or more control samples. The one or more control samples can comprise a positive control or a negative control. The one or more control samples may comprise at least one sample obtained from a subject suffering from WDLPS. The control sample may be a pooled sample obtained by combining two or more samples of the same type, e.g., two or more normal adipose tissue samples, or two or more WDLPS tumor biopsies. The expression level of a marker in the control sample may be the mean of the expression level of the marker calculated in each of the two or more samples making up the control sample.

[0073]

[0069] The one or more control samples may comprise normal tissue, i.e., non-cancerous tissue. For example, the control sample may be normal adipose tissue. Normal adipose tissue expresses IGF1, PPARG (e.g., PPARG1 and PPARG2), FABP4, and LPL, and thus can be used as a positive control when determining expression levels of IGF1, PPARG (e.g., PPARG1 and / or PPARG2), FABP4, and / or LPL in an LPS tumor sample.

[0074]

[0070] Various other tissues express PPARG including brain, gallbladder and tonsil tissue, and one or more of such tissues can also serve as (a) positive control(s) when determining PPARG expression. Similarly, FABP4 is also expressed in lung and breast tissue, which may serve as positive control when determining FABP4 expression. LPL is also expressed in cardiac muscle tissue, which hence can serve as positive control when determining LPL expression.

[0075]

[0071] Normal pancreas, thymus, or retina do not express IGF1 and therefore can be used as negative control when determining IGF1 expression. Normal retina, pancreas, and smooth muscle do not express PPARG and thus can be used as negative control when determining PPARG expression. Normal liver and gallbladder do not express FABP4 and hence can be used as negative control for determining FABP4 expression. Normal retina, liver, and gallbladder do not express LPL and therefore may be used as negative control when determining LPL expression.

[0076]

[0072] IGF1R is expressed in the brain, gallbladder, and cardiac muscle, and thus one or more of these tissues can serve as a positive control when determining IGF1R expression. Normal retina and liver tissue do not express IGF1R and thus can be used as negative control when determining IGF1R expression.

[0077]

[0073] In some instances, it may be desirable to use multiple tissues as positive and negative controls when determining expression levels of the markers disclosed herein.

[0078]

[0074] In some instances, the control sample may be an LPS sample obtained from the same subject at an earlier timepoint. A sample obtained from the same subject at an earlier timepoint may be a useful control to monitor disease progression. If no change in expression in the one or markers described herein is observed in a second LPS sample obtained from a subject at a later timepoint relative to a first LPS sample obtained from the subject at an earlier timepoint, the LPS has not progressed.

[0079] Marker detection

[0080] Nucleic acid probes

[0081]

[0075] The expression level of a marker disclosed herein may be determined using a nucleic acid probe that is capable of specifically hybridizing to an mRNA encoding the marker. Various methods that employ nucleic acid probes for determining expression levels are known in the art. For example, the expression level of an mRNA may be determined using quantitative polymerase chain reaction (qPCR) and pairs of forward and reverse primers that specifically hybridize to the mRNA. Alternatively, one or more TAQMAN® probes that specifically hybridize to an mRNA may be used.

[0082]

[0076] In some instances, one or more nucleic acid probes (e.g., a set of qPCR primers or a TAQMAN® probe) that specifically hybridize to mRNAs encoding IGF1, IGF1R and / or PPARG (e.g., PPARG1 and / or PPARG2) are used to determine the expression levels of one or more markers in a LPS sample obtained from a subject. For example, a set of qPCR primers or a TAQMAN® probe may be used to determine the expression of PPARG2.

[0083]

[0077] Typically, a panel of at least two (e.g., three, four, five or more) nucleic acid probes is used to determine the expression levels of two or more markers in a LPS sample obtained from a subject. A suitable panel may comprise three nucleic acid probes (e.g., three sets of qPCR primers or three TAQMAN® probes) that specifically hybridize to mRNAs encoding IGF1R, PPARG (e.g., PPARG1 and / or PPARG2), and LPL, respectively. Alternatively, a suitable panel may comprise three nucleic acid probes (e.g., three sets of qPCR primers or three TAQMAN® probes) that specifically hybridize to mRNAs encoding IGF1, FABP4, and LPL, respectively.

[0084]

[0078] Also contemplated are kits that comprise one or more nucleic acid probes (e.g., a panel of nucleic acid probes, such as one or more sets of qPCR primers) and, e.g., suitable reagents to perform a qPCR reaction. In addition to a panel of nucleic acid probes, kits may further comprise multiple tissue samples as described above that can serve as positive and negative controls to determine expression levels of the marker(s).

[0085]

[0079] In some instances, RNA in situ hybridization (RNA ISH) methods (e.g., RNAscope) may be used to determine the expression levels of the markers described herein. Alternatively, the expression levels of markers described herein may be determined using spatial transcriptomics e.g., using RNA sequencing, which may be bulk RNA sequencing or single-cell RNA sequencing. Suitable kits for performing RNA ISH or spatial transcriptomics are described in the Examples and / or are readily available commercially.

[0086]

[0080] Exemplary nucleic acid sequences of nucleic acid primers are shown in Table 1.

[0087] Table 1. Exemplary nucleic acid sequences of nucleic acid primers.

[0088] Antibodies

[0089]

[0081] More typically, the expression level of a marker described herein is determined, e.g., using an antibody that specifically binds to the marker. Advantageously, immunohistology is readily combinable with standard histological analysis. Antibodies that are capable of specifically binding to IGF1, IGF1R, PPAR-y (e.g., PPAR-yl and / or PPAR-y2), LPL and FABP4 are readily available commercially and can be used with standard immunohistology protocols. Such antibodies may be linked to a detectable moiety (such as a fluorescent moiety or an enzyme). Accordingly, expression levels of one or more markers may be determined with one or more antibodies, e.g., an antibody that is capable of specifically binding to IGF1, an antibody that is capable of specifically binding to IGF1R, an antibody that is capable of specifically binding to PPAR-y, an antibody that is capable of specifically binding to LPL, and / or an antibody that is capable of specifically binding to FABP4.

[0090]

[0082] An anti-PPAR-y antibody may specifically bind to PPAR-yl or PPAR-y2. Alternatively, an anti-PPAR-y antibody may specifically bind to both isoforms (i.e., PPAR- yl and PPAR-y2). Examples of anti-PPAR-y antibodies are sc-166731 (Santa Cruz, CA), PAI -824 (Thermo-Fisher) and PP-K8450B-00 (Novus).

[0091]

[0083] Typically, a panel of at least two (e.g., three, four, five or more) antibodies is used to determine the expression levels of two or more markers disclosed herein in a LPS sample obtained from a subject. A suitable panel may comprise three antibodies that specifically bind to IGF1R, PPAR-y, and LPL, respectively. Alternatively, a suitable panel may comprise three antibodies that specifically bind to IGF1, FABP4, and LPL, respectively.

[0092]

[0084] An antibody for use in the disclosed methods may be a polyclonal or a monoclonal antibody. More typically, the antibody is a monoclonal antibody. The antibody may be a single-domain antibody (e.g., a nanobody), or a single-chain domain antibody (e.g., a singlechain Fv). The antibody may be conjugated to a detectable moiety (e.g., a brightfield dye or a fluorophore). For example, the detectable moiety can be a dye detectable via brightfield microscopy, or a fluorophore detectable by fluorescence microscopy.

[0093]

[0085] Any method known in the art for determining protein expression in a sample may be used, e.g., immunohistochemistry (IHC), Western blot, immunofluorescence (IF), flow cytometry. IHC is commonly used to analyze tissue samples obtained from a subject suffering from LPS, e.g., tumor biopsies or resected tumors. Accordingly, IHC may be particularly useful to determine levels of expression of the markers described herein in a tissue sample, i.e., a LPS sample.

[0094]

[0086] Determining protein expression in a LPS sample (e.g., using IHC) may be ancillary to a standard analysis performed by a pathologist. For example, a LPS sample may also be analyzed to determine the presence of DDLPS tumor cells by determining the morphology of the cells, mitotic rate, pleomorphism, and cellular density in the LPS sample.

[0095]

[0087] Also contemplated are kits that comprise an antibody panel and multiple tissue samples as described above that can serve as positive and negative controls to determine expression levels.

[0096] Expression level

[0097]

[0088] As shown herein, WDLPS tumor cells show high expression levels of IGF1, PPARG, FABP4, and LPL and low expression levels of IGF1R, an expression pattern similar to adipocytes. Conversely, DDLPS tumor cells show very low / almost no expression of IGF1, FABP4, or LPL, and a high expression of IGF1R. Similarly, expression of PPARG (e.g., PPARG1 and / or PPARG2) is markedly decreased in DDLPS compared to WDLPS or normal adipose tissue. Expression of PPARG (e.g., PPARG1) was retained only in a minority of DDLPS tumor cells.

[0098]

[0089] Accordingly, a LPS may be characterized as having metastatic potential, when a tissue biopsy is IGFlRHlgh, PPARGLow(e.g., PPARG1LOWand / or PPARG2Low), and LPLLOW, as identified, e.g., by immunohistology. Similarly, a LPS may be characterized as having metastatic potential when a tissue biopsy is IGF1LOW, FABP4Low, and LPLLOW. Determining the expression of PPARG2 may be particularly useful owing to the especially low levels of expression of PPARG2 (e.g., absence of expression) in DDLPS compared to normal adipose tissue and WDLPS.

[0099]

[0090] When histological analysis has identified a tissue sample as DDLPS, IGFlRHlghstaining alone is a good indicator that the subject suffers from a LPS with high metastatic potential. Similarly, PPARG2Lowstaining, including absence of PPARG2 staining, alone may be a suitable indicator that the subject suffers from a LPS with high metastatic potential (e.g., DDLPS). In contrast, when histological analysis has identified a tissue sample as DDLPS, confirmation that the DDLPS is IGF1 positive is a good indicator that the subject suffers from a LPS with low metastatic potential.

[0100]

[0091] The diagnostic methods described herein may rely on threshold values to determine if a subject’s LPS has increased metastatic potential relative to the threshold value, e.g., is progressing from WDLPS to DDLPS, or is DDLPS.

[0101]

[0092] The expression of a single marker may be assessed. For example, a ten-fold or more decrease (e.g., twenty-fold or more, fifty -fold or more) in the expression of PPARG2 in the sample obtained from the subject compared to a control sample is indicative of a LPS with metastatic potential. For instance, the decrease in the expression of PPARG2 in the sample obtained from the subject may be one hundred-fold or more compared to a control sample. In other instances, a 5-fold or more decrease in the expression of PPARG1 in the sample obtained from the subject compared to a control sample may be indicative of a LPS with metastatic potential. In some instances, the expression of PPARG1 may decrease by six-fold or more, seven-fold or more, eight-fold or more compared to a control sample. The control sample may be normal adipose tissue or a WDLPS sample.

[0102]

[0093] Other markers may have similar utility in indicating a LPS’s metastatic potential. For example, a 2-fold or more (e.g., 3-fold, 4-fold, or 5-fold) increase in the expression of IGF1R in a LPS sample obtained from a subject compared to a control sample (e.g., adipose tissue obtained from the subject) is indicative of the LPS having metastatic potential. Similarly, a 50% or more (e.g., 60%, 70%, or 80% or more) decrease in the expression of PPARG (e.g., PPARG1 and / or PPARG2), LPL, FABP4 or IGF1 in a LPS sample obtained from a subject compared to a control sample is indicative of the LPS having metastatic potential.

[0103]

[0094] For enhanced accuracy and / or sensitivity, more than a single marker is assessed. In some instances, a 2-fold or more (e.g., 3-fold, 4-fold, or 5-fold) increase in the expression of IGF1R in the sample obtained from the subject compared to the control, a 50% or more decrease in the expression of PPARG (e.g., PPARG1 and / or PPARG2) and a 70% or more (e.g., 80% or more, or 90% or more) decrease in the expression of LPL in the sample obtained from the subject compared to the control sample is indicative of a LPS with metastatic potential. In some instances, a 70% or more (e.g., 80% or more, or 90% or more) decrease in the expression of IGF 1, FABP4, and LPL in the sample obtained from the subject compared to the control sample is indicative of a LPS with metastatic potential.

[0104]

[0095] In other instances, a 2-fold or more (e.g., 3-fold, 4-fold, or 5-fold) increase in the expression of IGF1R in the sample obtained from the subject compared to the control and a 50% or more decrease in the expression of PPARG (e.g., PPARG1 and / or PPARG2), and a 70% or more (e.g., 80% or more, or 90% or more) decrease in the expression of LPL, IGF1, and FABP4 in the sample obtained from the subject compared to the control sample is indicative of a LPS with metastatic potential.

[0105] Therapeutic applications

[0106]

[0096] The diagnostic methods disclosed herein can be used to assess the metastatic potential of a subject’s LPS and thus can be used to identify subjects that could benefit from a therapeutic intervention or particular types of therapies. In particular, differences in gene expression between WDLPS and DDLPS tumors have been identified including, but not limited to, decreases in the expression levels of IGF1, PPARG (e.g., PPARG1 and / or PPARG2), LPL, and FABP4 and an increase in the expression level of IGF1R in DDLPS compared to WDLPS or normal adipocytes.

[0107]

[0097] The LPS of the subject described herein may be a WDLPS, a DDLPS, or a mixed LPS which comprises WDLPS and DDLPS tumor cells. Even a small focus of hyper-dense, highly mitotic cells may indicate the presence of DDLPS tumor cells in a LPS sample and results in a pathologist indicating that the sample is a mixed LPS.

[0108]

[0098] In some instances, the described therapies are particularly beneficial for subjects with de-differentiated LPS (e.g., DDLPS), or mixed LPSs.

[0109] Recombinant human IGF1 and IGF1 mimetics

[0110]

[0099] It has now been discovered that subjecting LPS tumor cells to exogenous IGF1 increases adipocyte differentiation markers, for example PPARG (e.g., PPARG1 and / or PPARG2) and adiponectin (ADIPOQ). Stimulating differentiation of LPS tumor cells by upregulating IGF1 signaling, particularly in LPS cells that exhibit evidence of de-differentiation markers as identified herein, may thus result in such cells having a reduced metastatic phenotype and an increased adipocyte phenotype, resulting in a decreased metastatic potential.

[0100] Accordingly, a method of treating a subject with LPS is provided and comprises administering a therapeutically effective amount of recombinant human IGF1 or an IGF1 mimetic to the subject. Also provided is a recombinant human IGF1 or an IGF1 mimetic for use in a method of treating a subject with LPS. The use of a recombinant human IGF1 or an IGF1 mimetic for the manufacture of a medicament for LPS is also provided.

[0111]

[0101] Exemplary IGF1 mimetics for use with the treatment methods described herein include, but are not limited to, BVS857 or Nap-FFG-GYGSSSRRAPQT (Nap is naphthalene acetic acid, the amino acid sequence FFG-GYGSSSRRAPQT is shown in SEQ ID NO: 9). Nap-FFG-GYGSSSRRAPQT is described in Wang et al. (2023) Nanotoday, doi.org / 10.1016 / j.nantod.2023.101894.

[0112] PPAR-y agonist

[0113]

[0102] It has now been discovered that subjecting LPS to exogenous IGF1 increases the adipocyte differentiation marker PPARG (e.g., PPARG1 and / or PPARG2). Thus, directly agonizing PPAR-y in LPS tumor cells may also result in upregulation of adipocyte differentiation markers, thereby decreasing the metastatic potential of the LPS tumor cells. An increase in differentiation induced by the PPAR-y agonist alone or in combination with another therapy described herein (e.g., recombinant IGF1 or an IGF1 mimetic, or a GLP-1 agonist) may reduce the metastatic potential of an LPS.

[0114]

[0103] Accordingly, a method of treating a subject with LPS is provided and comprises administering a therapeutically effective amount of a selective PPAR-y agonist. Also provided is a selective PPAR-y agonist for use in a method of treating a subject with LPS. The use of a selective PPAR-y agonist for the manufacture of a medicament for LPS is also provided.

[0115]

[0104] Exemplary PPAR-y agonists suitable for use with the treatment methods described herein include, but are not limited to, AMG-131, Muraglitazar, Lobeglitazone, Naveglitazar, Rosiglitazone, Pioglitazone, and Troglitazone.

[0116] GLP-1 agonist

[0117]

[0105] It has now been discovered that subjecting LPS to a combination of a selective glucagon-like peptide-1 (GLP-1) agonist and IGF1 increases adipocyte differentiation markers (e.g., ADIPOQ) more than subjecting an LPS to IGF1 alone. Thus, an increase in differentiation induced by the selective GLP-1 agonist alone or in combination with another therapy described herein may reduce the metastatic potential of an LPS (e.g., recombinant IGF1 or an IGF1 mimetic, or a PPAR-y agonist). Therefore, a selective GLP-1 agonist may be useful in a method of treating LPS.

[0118]

[0106] Accordingly, a method of treating a subject with LPS is provided and comprises administering a therapeutically effective amount of a selective GLP-1 agonist.

[0119]

[0107] Exemplary selective GLP-1 agonists suitable for use with the treatment methods described herein include, but are not limited to, liraglutide, exenatide, albiglutide, dulaglutide, lixisenatide, and semaglutide.

[0120] IGF1R kinase inhibitors

[0121]

[0108] It has now been discovered that expression of IGF1R in DDLPS is increased compared to WDLPS or normal adipocytes. This increased expression, in turn, may be associated with increased or aberrant signaling of the IGF1R kinase. Therefore, a method of treating a subject with LPS is provided, comprising administering a therapeutically effective amount of a selective IGF1R kinase inhibitor. Also provided is an IGF1R kinase inhibitor for use in a method of treating a subject with LPS. The use of an IGF1R kinase inhibitor for the manufacture of a medicament for LPS is also provided.

[0122]

[0109] Exemplary selective IGF1R kinase inhibitors for the treatment methods disclosed herein include, but are not limited to, Linsitinib or cyclolignan PPP.

[0123] IGF1R antibody-drug conjugates (ADCs)

[0124] [HO] It has now been discovered that IGF1R is overexpressed in DDLPS tumor cells compared to normal adipocytes and WDLPS tumor cells. This may allow for a chemotherapeutic intervention that specifically targets LPS tumor cells that overexpress IGF1R using an antibody that specifically binds to the extracellular region of IGF1R. Furthermore, as shown herein, an IGF1R antibody-drug conjugate (ADC) can also be used to specifically target and reduce the viability of WDLPS-derived cells. Accordingly, a method of treating a subject with LPS is provided, comprising administering a therapeutically effective amount of an antibody-drug conjugate (ADC), wherein the antibody specifically binds to the extracellular region of IGF1R. Also provided is such an ADC for use in a method of treating a subject with LPS. The use of such an ADC for the manufacture of a medicament for LPS is also provided.

[0125]

[0111] Exemplary antibodies for use in the ADC include, but are not limited to, lonigtumab (hz208F2-4), ganitumab, teprotumumab, figitumumab, dalotuzumab, robatumumab, and veligrotug (or a half-life extended variant thereof such as VRDN-003). The ADC may be lonigutamab ugodotin (W0101).

[0126]

[0112] The drug used in the ADC may be a cytotoxic agent such as a tubulin inhibitor or a DNA damaging agent. The tubulin inhibitor can be a dolastatin or an auristatin (e.g., MMAE), a maytansinoid (e.g., maytansine or DM-1), or a tubulysin. In some instances, a dolastatin / auristatin derivative such as ugodotin may be used. The DNA damaging agent can be an agent that induces DNA double-strand breaks, a DNA alkylating agent, a topoisomerase inhibitor, or a DNA crosslinking agent. The agent that induces DNA double-strand breaks may be a calicheamicin. The DNA alkylating agent may be a duocarmycin. The topoisomerase inhibitor may be an exatecan (e.g., deruxtecan). The DNA crosslinking agent may be a pyrrolobenzodiazepine.

[0127] Combination therapy

[0128]

[0113] It may be advantageous to combine the therapies disclosed herein. Accordingly, in some aspects, provided herein is a method of treating a subject suffering from a LPS with chemotherapy, comprising administering to the subject a therapeutically effective amount of the chemotherapy comprising a first therapy and a second therapy.

[0129]

[0114] The first therapy may comprise a selective IGF1R kinase inhibitor (e.g., Linsitinib or cyclolignan PPP), and the second therapy may comprise a recombinant human IGF1 or an IGF1 mimetic (e.g, BVS857 or Nap-FFG-GYGSSSRRAPQT).

[0130]

[0115] Alternatively, the first therapy may comprise a selective IGF1R kinase inhibitor (e.g, Linsitinib or cyclolignan PPP), and the second therapy may comprise a selective PPAR-y agonist (e.g, AMG-131, Muraglitazar, Lobeglitazone, Naveglitazar, Rosiglitazone, Pioglitazone, or Troglitazone).

[0131]

[0116] Or, the first therapy may comprise an antibody-drug conjugate (ADC) wherein the antibody (e.g, lonigutamab (hz208F2-4), ganitumab, teprotumumab, figitumumab, dalotuzumab, robatumumab, or veligrotug) specifically binds to the extracellular region of IGF1R, and the second therapy may comprise a recombinant human IGF1 or an IGF1 mimetic (e.g., BVS857 or Nap-FFG-GYGSSSRRAPQT). The ADC may be lonigutamab ugodotin (WO 101).

[0132]

[0117] Or, the first therapy may comprise an antibody-drug conjugate (ADC) wherein the antibody (e.g., lonigutamab (hz208F2-4), ganitumab, teprotumumab, figitumumab, dalotuzumab, robatumumab, or veligrotug) specifically binds to the extracellular region of IGF1R, and the second therapy may comprise a selective PPAR-y agonist (e.g., AMG-131, Muraglitazar, Lobeglitazone, Naveglitazar, Rosiglitazone, Pioglitazone, or Troglitazone). The ADC may be lonigutamab ugodotin (W0101).

[0133]

[0118] Or, the first therapy may comprise a recombinant human IGF1 or an IGF1 mimetic (e.g., BVS857 or Nap-FFG-GYGSSSRRAPQT) and the second therapy may comprise a selective GLP-1 agonist (e.g., liraglutide, exenatide, albiglutide, dulaglutide, lixisenatide, or semaglutide).

[0134]

[0119] In some aspects, the first and second therapies may be suitable for increasing the differentiation of the LPS. For example, the first therapy may be a recombinant human IGF1 or an IGF1 mimetic (e.g., BVS857 or Nap-FFG-GYGSSSRRAPQT), and the second therapy may be a selective PPAR-y agonist (e.g., AMG-131, Muraglitazar, Lobeglitazone, Naveglitazar, Rosiglitazone, Pioglitazone, or Troglitazone). Alternatively, the second therapy may be a selective GLP-1 agonist (e.g., liraglutide, exenatide, albiglutide, dulaglutide, lixisenatide, or semaglutide). In some instances, the first therapy may be a selective PPAR-y agonist (e.g., AMG-131, Muraglitazar, Lobeglitazone, Naveglitazar, Rosiglitazone, Pioglitazone, or Troglitazone), and the second therapy may be a selective GLP-1 agonist (e.g., liraglutide, exenatide, albiglutide, dulaglutide, lixisenatide, or semaglutide).

[0135]

[0120] In some aspects, the first therapy may be suitable for increasing the differentiation of the LPS (e.g., IGF1 or an IGF1 mimetic, a selective PPAR-y agonist, a selective GLP-1 agonist, or a combination thereof) and the second therapy may selectively target IGF1R- overexpressing LPS cells. For example, the second therapy may be an antibody-drug conjugate (ADC) wherein the antibody (e.g., lonigutamab (hz208F2-4), ganitumab, teprotumumab, figitumumab, dalotuzumab, robatumumab, or veligrotug) specifically binds to the extracellular region of IGF1R, or the second therapy may be a selective IGF1R kinase inhibitor (e.g., Linsitinib or cyclolignan PPP).

[0136]

[0121] The first therapy and second therapy may be administered concurrently. Alternatively, the first therapy and second therapy may be administered sequentially, i.e., one treatment course begins after the other treatment course ends. For example, in some instances, he first therapy is administered before the second therapy. In other instances, the second therapy is administered before the first therapy.

[0137]

[0122] A chemotherapy described herein may be combined with a conventional chemotherapy for LPS. Conventional chemotherapy includes treatment with doxorubicin, ifosfamide, gemcitabine + docetaxel, gemcitabine + dacarbazine, or combinations thereof.

[0138]

[0123] Alternatively, a chemotherapy described herein may be combined with an MDM2 inhibitor (e.g., RG7112) or a CDK4 inhibitor (e.g., palbociclib).

[0139] Patient stratification selection

[0140]

[0124] As shown herein, subjects having an IGF1 positive dedifferentiated LPS (e.g., IGF1 positive DDLPS) are less likely to develop recurrent and / or metastatic disease. This finding supports the therapeutic strategies set out herein to restore IGF1 signaling in subject with an IGF lRHlghdedifferentiated LPS (e.g., IGF lRHlghDDLPS). As demonstrated here, such LPS can be characterized as IGFlRHlgh, PPARGLow(e.g., PPARG1LOWand / or PPARG2Low), and LPLLO", or IGF1LOW, FABP4LOW, and LPLLOWas determined, e.g., by immunohistology or gene expression analysis.

[0141]

[0125] Some of the experiments described herein indicate that IGF1 expression is a useful marker for assessing the survival probability in subjects with LPS. The level of IGF1 expression in a sample obtained from a subject’s LPS relative to the mean expression level of IGF1 in a control population of subjects with LPS may therefore be used to stratify a subject for therapeutic intervention, e.g., chemotherapy.

[0142]

[0126] Other experiments described herein indicate that levels of expression of PPARG2 and optionally PPARG1 can be used to determine whether an LPS is WDLPS or DDLPS. These experiments further indicate that levels of expression of PPARG2 and optionally PPARG1 may be used to monitor whether WDLPS is progressing or has progressed to a DDLPS.

[0127] Thus, in some aspects, a method of identifying a subject suffering from a LPS with metastatic potential is provided that comprises determining expression levels of IGF 1 , IGF 1R and / or PPARG (e.g., PPARG1 and / or PPARG2) in a sample of the LPS obtained from the subject. In some instances, the method further comprises determining the expression levels of LPL and / or FABP4.

[0143]

[0128] For example, in some instances, a method of identifying a subject suffering from a LPS with metastatic potential is provided that comprises determining expression levels of the following markers in a sample of the LPS obtained from the subject: (i) IGF1R, PPARG, and LPL; and / or (ii) IGF1, FABP4, and LPL.

[0144]

[0129] Also provided is a method of identifying a subject suffering from a LPS with metastatic potential, comprising determining the activation state of a PPARG2 gene regulatory region (e.g., nucleotides 12351171-12351492 on chromosome 3) in a LPS sample obtained from the subject.

[0145]

[0130] In some instances, a method of selecting a subject with LPS for chemotherapy may comprise assessing by histological analysis whether the subject’s LPS is lGFlRPositiveor IGFlRHlgh, wherein the subject is (i) selected for chemotherapy if the LPS is lGFlRp0Sltlveor IGFlRHlgh, and (ii) not selected for chemotherapy if the LPS is !GFlRNegativeor IGF1RLOW. For example, the LPS may be a mixed LPS that comprises IGFlRPositivetumor cells. In other instances, a method of selecting a subject with LPS for chemotherapy may comprise assessing by histological analysis whether the subject’s LPS is DDLPS and IGFlRHlgh, wherein the subject is (i) selected for chemotherapy if the DDLPS is IGFlRHlgh, and (ii) not selected for chemotherapy if the DDLPS is IGF1 positive. The chemotherapy may be an IGF1R ADC.

[0146]

[0131] In further aspects, a method of selecting a subject with LPS for chemotherapy is provided, comprising determining expression levels of IGF 1R and / or PPARG (e.g., PPARG1 and / or PPARG2) in a sample of the LPS obtained from the subject compared to a control sample; wherein the subject is selected for chemotherapy if the sample has: (a) a 2-fold or more increase in the expression of IGF1R in the sample obtained from the subject compared to the control; and / or (b) a 50% or more decrease in the expression of PPARG and / or IGF1 in the sample obtained from the subject compared to the control sample (e.g., normal adipose tissue).

[0132] In some instances, a method of selecting a subject with LPS for chemotherapy further comprises determining the expression levels of LPL and / or FABP4. For example, the subject may be selected for chemotherapy if the sample has a 50% or more decrease in the expression of FABP4 and / or LPL in the sample obtained from the subject compared to the control sample.

[0147]

[0133] In particular, a method of selecting a subject with LPS for chemotherapy is provided that comprises determining expression levels of the following markers in a sample of the LPS obtained from the subject compared to a control sample: (i) IGF1R, PPARG, and LPL; and / or (ii) IGF1, FABP4, and LPL; wherein the subject is selected for chemotherapy if (a) the sample has a 2-fold or more increase in the expression of IGF1R in the sample obtained from the subject compared to the control and a 50% or more decrease in the expression of PPARG and LPL in the sample obtained from the subject compared to the control; and / or (b) a 50% or more decrease in the expression of IGF 1, FABP4, and LPL in the sample obtained from the subject compared to the control sample. A subject with LPS that does not meet these thresholds may be continued to be monitored for tumor progression, e.g., using the diagnostic methods described herein.

[0148]

[0134] In some aspects, a method of selecting a subject with LPS for chemotherapy is provided that comprises determining the activation state of a PPARG2 gene regulatory region (i.e., nucleotides 12351171-12351492 on chromosome 3) in a LPS sample obtained from the subject, wherein the subject is selected for chemotherapy if the PPARG2 gene regulatory region is in an inactive state.

[0149]

[0135] In the above methods, selecting a subject for chemotherapy may mean that the subject receives a recommendation for a therapy, e.g., treatment with one or more of the compounds discussed above. Such a recommendation may form a part of a diagnostic report that includes the analysis results for the sample obtained from the subject. Alternatively, the treating physician may write a prescription for the therapy based on the analysis results. Typically, the selection does not include a treatment step in which the subject is administered a therapeutically effective amount of the therapy for which the subject is selected.

[0150]

[0136] Accordingly, in further aspects, a method of treating a subject suffering from LPS with chemotherapy is provided that comprises administering to the subject a therapeutically effective amount of the chemotherapy, wherein the subject is selected for chemotherapy if the LPS has: (a) a 2-fold or more increase in the expression of IGF1R in the sample obtained from the subject compared to a control; and / or (b) a 50% or more decrease in the expression of PPARG (e.g., PPARG1 and / or PPARG2) and / or IGF1 in the sample obtained from the subject compared to the control sample (e.g., normal adipose tissue).

[0151]

[0137] In some instances, such a method further comprises determining the expression levels of LPL and / or FABP4 in the sample. For example, the subject may be selected for chemotherapy if the sample has a 50% or more decrease in the expression of FABP4 and / or LPL in the sample obtained from the subject compared to the control sample.

[0152]

[0138] Accordingly, also provided are methods of treating a subject suffering from LPS with chemotherapy that comprise administering to the subject a therapeutically effective amount of the chemotherapy, wherein the subject is selected for chemotherapy if the LPS has (a) a 2-fold or more increase in the expression of IGF1R in the sample obtained from the subject compared to the control and a 50% or more decrease in the expression of PPARG and LPL in the sample obtained from the subject compared to the control; and / or (b) a 50% or more decrease in the expression of IGF1, FABP4, and LPL in the sample obtained from the subject compared to the control sample.

[0153]

[0139] The chemotherapy may be selected for its potential to regulate the dysregulated IGF1-IGF1R signaling axis in de-differentiated LPS tumor cells. For example, the chemotherapy may stimulate IGF1 signaling (which is reduced in DDLPS tumor cells). Suitable chemotherapies may include administering a therapeutically effective amount of a recombinant human IGF1, IGF1 mimetic, PPAR-y agonist to the subject selected to receive chemotherapy. Alternatively, the chemotherapy may be aimed at reducing aberrant IGF1R signaling (which may be increased in DDLPS tumor cells due to the high expression of IGF1R). Suitable chemotherapies may therefore include (in addition or alternatively) administering a therapeutically effective amount of an IGF1R kinase inhibitor to the subject selected to receive chemotherapy.

[0154]

[0140] In some instances, the chemotherapy may be selected for its capability to specifically target and selectively kill DDLPS tumor cells. Such chemotherapies may be based on the high expression of IGF1R in DDLPS tumor cells. Suitable chemotherapies may therefore include administering a therapeutically effective amount of an IGF IR-targ eting ADC to the subject selected to receive chemotherapy. EXAMPLES

[0155]

[0141] The following examples are included for illustrative purposes only and are not intended to limit the scope of the invention. Those of ordinary skill in the art may be aware of materials and methods similar or equivalent to those described below and any of these can be used to practice or test the provided methods and compositions.

[0156] Example 1. Multiome experiment

[0157]

[0142] This example describes a multiome experiment conducted to assess the differences between liposarcoma (LPS) subtypes and control tissue (adipose tissue). Multiome analysis allows for the simultaneous analysis of chromatin accessibility (by an assay for transposase- accessible chromatin with sequencing (ATAC-seq)) and gene expression (by single nucleus RNA sequencing (snRNA-seq)) of the captured cells.

[0158] Control and patient samples

[0159]

[0143] Nine well-differentiated liposarcoma (WDLPS), ten de-differentiated liposarcoma (DDLPS), and four adipose control liquid nitrogen snap-frozen patient samples were obtained.

[0160]

[0144] Part of each sample was retained for pathology analysis. Pathology of the samples was confirmed by review of surgical pathology reports as well as independent pathology review. All LPS samples had evidence of chromosome 12q amplification. Patient age and sex were consistent with clinical observations of LPS epidemiology. Patient tumor or control tissue samples were derived from the visceral abdominopelvic region except for two samples that were obtained from metastatic sites - one from the supraclavicular soft tissue region and the other from the retro-orbital space (only this tumor sample had exposure to chemotherapy at time of resection). Three samples were exposed to preoperative radiation. All other samples were obtained from patients who were treatment-naive.

[0161] Sample preparation and sequencing

[0162]

[0145] The remaining part of each sample was utilized for multiome analysis. Nuclei were isolated by mincing with spring scissors and mechanical dissociation by pipetting for a total of 10 minutes in Tris-Salt-Tween (0.03% Tween) extraction buffer on ice and filtered through a 30 pm strainer. Washed nuclei were centrifuged (500 x g, 5 minutes, 4°C). The pellet was resuspended and permeabilized using 100 pl of commercially available lysis buffer on ice for 2 minutes. Nuclei were then washed and resuspended in a commercially available diluted nuclei buffer. For visualization, nuclei were stained with 4',6-diamidino-2-phenylindole (DAPI) and counted on a hemocytometer. Nuclei were loaded with a target range from 1,118 to 10,000, depending on sample size. Single nucleus RNA (snRNA) and single nucleus AT AC (snATAC) libraries were prepared following a 10X Genomics protocol (Chromium Next GEM Single Cell Multiome ATAC+ Gene Expression User Guide Rev E). Each sample went through the following specific procedures: Transposition, GEM Generation and Barcoding, Post-GEM Incubation and Cleanup, Pre-Amplification PCR, ATAC library construction, cDNA Amplification, and Gene Expression Library Construction. Library metrics were analyzed with the Agilent Bioanalyzer and sequenced per the specifications in the Agilent’s Multiome User Guide. All libraries were sequenced using an Illumina NextSeq 2000 sequencer. snRNA-seq cohort generation

[0163]

[0146] For snRNA-seq data, 10X Cell Ranger ARC v2.0.0 (Cell Ranger v6.1.1) was used to demultiplex sequencing data, process barcodes, align sequencing reads to genome GRCh38-2020_arc_v2.0.0 and perform Unique Molecular Identifier (UMI) counting. Non-duplicate reads that mapped confidently to the genome were counted to form a genebarcode matrix. Cellranger output raw feature matrix (rather than the filtered matrix) was carried forward for downstream analysis, to allow for maximum capture of as many high- quality cells from each assay individually.

[0164]

[0147] Barcodes representing empty droplets, droplets containing two nuclei, nuclei with fewer than 500 UMI, 200 genes, or with more than 5% of reads from the mitochondrial genome, and outlier nuclei for total gene counts were removed from the analysis.

[0165]

[0148] Gene expression values were log-normalized, filtered for highly variable genes, and scaled. Principal component analysis (PCA) was performed, regressing total counts, before dimensionality reduction and visualization with Uniform Manifold Approximation and Projection (UMAP). Leiden unsupervised clustering using 40 principal components (PCs) at a low resolution on each sample was performed. Differential gene expression analysis comparing each Leiden cluster to all others with a Wilcoxon rank-sum test and Bonferroni FDR correction was used to broadly annotate cell types (adipocyte, endothelial, lymphoid, myeloid, tumor, unknown) in each tissue sample. Malignant cells were identified by marker gene differential expression as well as the presence of inferred CNV at chromosome 12q. Inferred CNV analysis was performed using both the InferCNV and SCEVAN algorithms, using lymphoid and myeloid cells as the reference cell type and all other cell types as the observation / experimental groups.

[0166]

[0149] Raw counts from each sample with broad cell type annotations were merged, filtered as above (with the addition of removing genes detected in less than 3 cells and barcodes with more than 8000 genes), log-normalized, and scaled. Since each sample was processed individually, Harmony _py v0.0.6 was used to mitigate potential batch effects (e.g., due to individual sample processing or radiation exposure) when performing dimension reduction and unsupervised clustering. PCA was performed on the adjusted 50 PCs followed by Leiden clustering at a resolution of 0.4 and UMAP visualization. ‘Unknown’ cells were reintegrated, clustered, log-normalized, scaled, and had differential gene expression performed to further identify more granular cell types. The complete snRNA-seq cohort consisted of 164,229 cells. Further downstream analysis was performed on original log-normalized expression values.

[0167]

[0150] The snRNA-seq dataset was used to perform differential gene expression analysis and consensus non-negative matrix factorization (cNMF) analysis. snATAC-seq cohort generation

[0168]

[0151] As above, Cell Ranger ARC v2.0.0 was used to create the barcode peak matrix using the Cell Ranger GRCh38 reference genome. The Cell Ranger output fragments file was used as the input to Signac vl .8.0, without filtering for barcodes deemed ‘cells’ by Cell Ranger to maximize the number of high quality cells analyzed from AT AC data (and not limit to only cells with the highest quality of both RNA and ATAC libraries). Doublets were removed by scDblFinder vl .12.0 with default parameters. Each sample had low quality barcodes filtered by: TSS > 2, fragments in peaks > 1000, nucleosome and blacklisted fraction were dynamically filtered by including barcodes that were less than 3 times the mean absolute deviation from the median prior to normalization, dimensionality reduction (using latent semantic indexing, LSI), and clustering. Gene activity was determined with FindAllMarkers to evaluate open chromatin regions near marker genes for broad cell type annotation on a per-sample basis. Tumor cell identity was confirmed with CopyscAT v0.40, where amplification of chromosome 12q was inferred.

[0152] Samples were merged, and reclustered (dims = 2:8; resolution = 0.1). Peaks were recalled by cell type (WDLPS tumor, DDLPS tumor, adipose, lymphoid, myeloid. . . etc). QC parameters for the entire cohort-level object were based on the distribution of the following metrics: peak region fragments < 10,000 and > 1,000, FRiP > 0, blacklist fraction < 0.02, nucleosome signal < 1.03, and TSS enrichment > 2. Cell annotations were confirmed with Find AllMarkers (Supplement). After filtering, 15,984 high quality nuclei remained for downstream analysis.

[0169]

[0153] The snATAC-seq dataset was used to perform differential peak accessibility analysis, which was used to analyze genomic region enrichments by pathology and to identify open peaks over IGF1 promoter regions in WDLPS.

[0170] Example 2. LPS subtypes have distinct transcriptomic cellular programs

[0171]

[0154] This example demonstrates the distinct transcriptomic profiles of two LPS subtypes. In particular, compared to cells obtained from DDLPS tumors, cells obtained from WDLPS tumors were enriched for genes involved in adipogenesis, insulin response, and lipid metabolism pathways.

[0172]

[0155] The 164,229 cells identified in the snRNA-seq cohort (see Example 1) were annotated for cell type based on marker gene expression, showing that the identified cells were consistent with cells found in visceral adipose tissue. A uniform manifold and approximation projection (UMAP) of the 164,229 cells, labeled by cell type, is shown in FIG. 1.

[0173]

[0156] Cells with amplification of chromosome 12q, inferred by high expression of marker genes (i.e., MDM2, CPM, FRS2, CDK4), were annotated as tumor cells. Specifically, a total of 80,881 WDLPS and DDLPS tumor cells were identified in the single-nucleus RNA sequencing (snRNA-seq) data. A UMAP of the 80,881 WDLPS and DDLPS tumor cells, labeled by tumor type (dark grey=DDLPS;light grey=WDLPS), is shown in FIG. 2A.

[0174]

[0157] Differences in the transcriptome between WDLPS and DDLPS tumor cells were analyzed. In particular, differential gene expression and consensus non-negative matrix factorization (cNMF) analysis were performed.

[0175] Differential gene expression analysis between WDLPS and DDLPS tumor cells

[0176]

[0158] This analysis revealed that insulin-like growth factor (IGF) binding proteins, IGF1, and Peroxisome Proliferator Activated Receptor Gamma (PPARG) were among the top differentially expressed genes in WDLPS and collagens (C0L1A2, C0L1A1) and insulinlike growth factor 1 receptor 1 (IGF1R) were among the top differentially expressed genes in DDLPS cells.

[0177]

[0159] Enrichment of known biological pathways were analyzed using BEANIE, a software package that identifies robustly enriched gene signatures that are not driven by a single patient or an imbalance in cell number per sample. The use of BEANIE sets a high threshold for a robust gene set enrichment result. This analysis revealed that adipogenesis, insulin response, and lipid metabolism pathways were enriched in WDLPS tumor cells.

[0178] Consensus non-negative matrix factorization (cNMF) analysis

[0179]

[0160] cNMF is an unbiased method to uncover co-regulated genes, constituting usage programs, also referred to as gene panels herein, that exist in a given dataset.

[0180]

[0161] Application of cNMF analysis to the present snRNA-seq cohort revealed seven total usage programs present within the LPS tumor cells (see Table 2). Programs that spanned multiple patients and were statistically significantly enriched for pathology were selected, narrowing the list to two usage programs. Usage program 1 was significantly associated with DDLPS, and usage program 3 was significantly associated with WDLPS.

[0181]

[0162] The overlap of the cNMF-derived usage program genes with Hallmark and Gene Ontology Biologic Processes gene sets was calculated. This showed that usage program 1 overlapped with largely early developmental pathways, including early mesenchymal pathways in large part due to genes ROBO1, ROBO2, and IGF1R. This program was therefore labelled “Early developmental”. Usage program 3 overlapped with adipocyte differentiation and insulin / RTK signaling, due to genes PPARG and IGF1 and was labelled “Insulin-mediated adipogenesis”. The cells high for usage program 1 are shown in the UMAP of FIG. 2B. The cells high for usage program 3 are shown in the UMAP of FIG. 2C. Comparing these UMAPs against FIG. 2A shows the overlap with DDLPS and WDLPS tumor cells, respectively. This was further validated by plotting each usage program score by pathology, as shown in FIG. 2D.

[0182] Table 2: Usage programs present within the LPS tumor cells.

[0183]

[0163] Taken together, differential gene expression and cNMF analysis both show that insulin-mediated adipogenesis pathways are enriched in WDLPS tumor cells, whereas early developmental pathways, including early mesenchymal pathways, are enriched in DDLPS tumor cells.

[0184]

[0164] This example demonstrates that a subject’s LPS can be characterized by providing a sample of the LPS obtained from the subject and determining the expression levels of genes involved in insulin-mediated adipogenesis (program 1, e.g., IGF1R) or the early development pathway (program 3, e.g., PPARG).

[0185] Example 3. Epigenetic landscape of LPS is characterized by IGF1 signaling differences

[0186]

[0165] This example demonstrates that IGF1 chromatin is less accessible in DDLPS compared to WDLPS and normal adipocytes.

[0166] ATAC-seq quality control metrics and filtering yielded 15,995 cells for downstream analysis. Similar cell types were recovered as seen in GEX data and tumor cells were again identified based on inferred amplifications in chromosome 12q. Without the need for integration, cells clustered by cell type except for tumor cells, which generally yielded patient-specific clusters. A UMAP of the 15,995 cells, labeled by type, is shown in FIG. 3 A.

[0187]

[0167] For analysis, 8,873 cells composed of adipocytes, WDLPS, and DDLPS tumor cells were selected. Adipocytes were included in this analysis to serve as a positive control for the biological relevance of ATAC data given the lower number of cells for this analysis.

[0188]

[0168] Differentially accessible peaks for each pathology group were calculated and Genomic Regions Enrichment of Annotations (GREAT) analysis was run on the identified peaks. Peaks were selected as differentially accessible if they met the threshold of an adjusted p value of < 0.05 and a log2 fold change of 1 or greater.

[0189]

[0169] As expected, adipocyte specific peaks were enriched for adipocytokine signaling, regulation of adipocyte differentiation, lipid metabolism, and insulin and IGF1 signaling. Relative to adipocytes, both DDLPS and WDLPS tumor cells were enriched for genomic regions associated with RXR and RAR heterodimerization and TLR3 signaling. WDLPS tumor cell peaks were also enriched for IGF1 signaling, driven by peaks found within the IGF1 and PPARG gene bodies, as shown in FIG. 3B. GREAT analysis of DDLPS tumor cells is shown in FIG. 3C. Unlike WDLPS tumor cells, DDLPS tumor cells were not enriched for IGF1 signaling. Therefore, of the most significantly enriched epigenetic enrichments, WDLPS and DDLPS differed in IGF1 signaling. This analysis corresponds with the snRNA- seq analysis performed in Example 2.

[0190]

[0170] Potential epigenetic silencing of IGF1 was analyzed in DDLPS tumor cells. The differential accessibility of peaks found within the IGF1 gene body was calculated and showed that five peaks were differentially accessible in WDLPS relative to DDLPS, as shown in FIG. 3D. This suggests that chromatin is open at these three regions in WDLPS but not in DDLPS. Four of the identified peaks are located at candidate response elements (CREs) associated with distal enhancer or promoter regions of the IGF1 gene. IGF1 accessibility in WDLPS but not DDLPS correlates with the snRNA-seq expression data, where IGF1 is nearly ubiquitously expressed by all WDLPS tumor cells and only expressed in approximately 10% of DDLPS tumor cells. Two of the patients from the multi ome cohort contributed DDLPS cells with retained IGF1 expression. As of June 2024, neither patient experienced a recurrence of their LPS in the > 6 years since resection. DDLPS otherwise has a recurrence rate of 80%. Epigenetic enrichment for IGF1 signaling in WDLPS coupled with inaccessible IGF1 CREs in DDLPS suggest that IGF1 signaling may be important for maintaining a WDLPS cell state. Furthermore, patients in the multiome cohort with DDLPS tumor cells retaining IGF1 expression had better outcomes than historical controls, suggesting that IGF1 signaling may predict more indolent disease.

[0191]

[0171] This example demonstrates that a subject’s LPS can be characterized by providing a sample of the LPS obtained from the subject and determining the expression levels of IGF 1 and PPARG.

[0192]

[0172] This example also demonstrates that a subset of DDLPS tumor cells share characteristics with WDLPS tumor cells, for example IGF1 expression, which indicates that determining gene expression, particularly expression of genes associated with IGF1 signaling, can be useful to stratify LPS patients for appropriate treatment regimens. In short, determining expression levels of markers including PPARG and IGF1 in a sample of a LPS obtained from a subject can be used to characterize the subject’s LPS, e.g., to determine its metastatic potential. Moreover, these markers can be used to select a subject with LPS for particular therapeutic interventions (see Examples 4 and 5).

[0193] Example 4. IGF1 / IGF1R / PPARG signaling is absent in DDLPS

[0194]

[0173] This example illustrates that IGF1R is upregulated in DDLPS, whilst IGF1 is downregulated.

[0195]

[0174] The genes within the multiome data generated in Example 1 which drive the enrichment of insulin-mediated adipogenesis pathways in WDLPS are IGF1 and PPARG. IGF1 is a regulator of adipogenesis, acting in an autocrine fashion to upregulate PPARG and increase expression of downstream markers of adipocyte terminal differentiation.

[0196]

[0175] The expression of IGF1, IGF1R, PPARG, and downstream markers of mature adipocyte differentiation LPL and FABP4 were assessed in the snRNA-seq cohort obtained in Example 1. As shown in FIG. 4A, mature adipocytes (“adipose” in FIG. 4A) show relatively high levels of IGF 1, PPARG, FABP4 and LPL with low levels of IGF1R. WDLPS tumor cells showed similar patterns of expression. Conversely, DDLPS tumor cells show almost no relative expression of IGF1, FABP4, or LPL. Some PPARG was retained in a minority of the cells. Surprisingly, DDLPS tumor cells showed a high relative expression of IGF1R. These findings suggest that autocrine IGF1 signaling is absent in DDLPS tumor cells. Moreover, these findings indicate that IGF1R, PPARG, IGF1, FABP4, and LPL are suitable markers to characterize a subject’s LPS, e.g., to determine its metastatic potential.

[0197]

[0176] An “IGF1 signaling score” was derived from the summed expression of positively regulated genes (IGF1, PPARG, FABP4, LPL) minus IGF1R expression and applied to the multi ome transcriptomic data with the Scanpy package, categorized by normal adipose tissue, or WDLPS or DDLPS tumor cells. Statistically significant differences were observed among adipose, WDLPS, and DDLPS cells, where adipocytes have the highest score and DDLPS have the lowest, as illustrated in FIG. 4B. Unexpectedly and surprisingly, as also shown by FIG. 4B, the DDLPS tumor cells separated into two populations. The first population had an IGF1 signaling score comparable to that of WDLPS tumor cells. The second population had a much lower IGF1 signaling score.

[0198]

[0177] An additional Formalin-Fixed Paraffin-Embedded (FFPE) patient sample that showed an LPS with a transition zone, i.e., that comprised a WDLPS region and a DDLPS region (as confirmed by independent pathology review), was analyzed using spatial transcriptomics. Specifically, consensus non-negative matrix factorization (cNMF) programs as described in Example 2 were used to analyze the sample.

[0199]

[0178] The WDLPS region of the sample was enriched for the insulin-mediated adipogenesis cNMF program (usage program 3) and had a higher IGF1 signaling signature score. Furthermore, a minority of cells that were associated with the WDLPS-associated program, i.e., usage program 3 “Insulin-mediated adipogenesis”, were observable in the histological DDLPS region and a minority of the cells that were associated with the DDLPS -associated program, i.e., usage program 1 “Early developmental” were observable in the histological WDLPS region, as shown in FIG. 4C.

[0200]

[0179] This example demonstrates that IGF1 signaling can be interrogated in addition to histological analysis to better understand the identity of an LPS tumor. This example also demonstrates that IGF1R is uniquely upregulated in DDLPS tumor cells, which are more aggressive than WDLPS tumor cells, indicating that IGF1R may be a potential target for anti-cancer therapy (e.g., in DDLPS). This indicates that an antibody-drug conjugate (ADC) that specifically binds to the extracellular region of IGF1R may be useful for treating subject suffering from an LPS exhibiting markers of de-differentiation (e.g., DDLPS). Similarly, a selective insulin-like growth factor-I receptor kinase inhibitor may be useful in treating subtypes of LPS that exhibit markers of de-differentiation.

[0201] Example 5. IGF1 stimulation of DDLPS cells in vitro induces PPARG expression and adipocytic differentiation

[0202]

[0180] This example demonstrates that DDLPS cells stimulated with IGF1 undergo adipocytic differentiation.

[0203]

[0181] Examples 2-4 show that a subset of DDLPS cells lacks IGF1 autocrine signaling that is usually in place to maintain adipocyte differentiation. It was therefore investigated whether exogenous IGF1 could modulate the differentiation status of these cells.

[0204]

[0182] First, in vitro models that recapitulate patterns of expression seen in the LPS human transcriptomic data were selected. Human LPS cell lines were analyzed using bulk RNA sequencing data. The LPS6 cell line, which was established from a DDLPS tumor, had gene expression profiles in the IGF1 signaling axis that mirrored the multi ome transcriptomic data described in Examples 2-4.

[0205]

[0183] LPS6 cells were therefore utilized for in vitro experiments. LPS6 cells were maintained in RPMI- 1640 medium, supplemented with 10% FBS, l% glutaMaX, 1,000 U / ml penicillin and 1 mg / ml streptomycin. In the experiment, LPS6 cells were seeded at 30% confluency and treated with RPMI- 1640 medium with 10% charcoal stripped FBS with a cocktail of ‘differentiation media’ (DM) comprising dexamethasone (Dex 1 mM), indomethacin (IDM, 50 mM), and 3-isobutyl-l-methylxanthine (IBMX, 0.5 mM). Human recombinant IGF-1 (100 or 200 nM) was added to the DM where indicated. Media was replenished every 3 days and harvested after 18 days in culture. The control condition was LPS6 cells supplemented with RPMI-1640 medium with 10% charcoal stripped FBS.

[0206]

[0184] Quantitative PCR (qPCR) was used to measure expression of adipocyte differentiation markers PPARG (namely PPARG1) and Adiponectin (ADIPOQ) using the respective primer sets in Table 3. Surprisingly, expression of PPARG 1 was increased by 2-fold with IGF1 alone and over 5-fold with DM and IGF1 relative to control as shown in FIG. 5. Furthermore, expression of ADIPOQ was also induced after differentiation treatment.

[0185] These results indicate that recombinant human IGF1 (or an IGF1 mimetic such BVS857 and Nap-FFG-GYGSSSRRAPQT) may be used to initiate differentiation of dedifferentiated LPS tumor cells (e.g., DDLPS tumor cells) toward an adipocyte phenotype, thus treating tumors comprising these cells. Recombinant human IGF1 or an IGF1 mimetic may therefore be useful for treating an LPS exhibiting markers of de-differentiation.

[0207]

[0186] Furthermore, expression of PPARG is reduced in DDLPS, as shown in Examples 2 and 4. In normal adipocytes, PPARG is downstream of IGF1 signaling. These considerations, taken together with the data presented in this example, indicate that PPAR-y could also be acted upon agonistically in de-differentiated LPS tumor cells (e.g., DDLPS tumor cells) to promote their differentiation toward an adipocyte phenotype, thereby reducing their metastatic potential. A selective PPAR-y agonist such as AMG-131, Muraglitazar, Lobeglitazone, Naveglitazar, Rosiglitazone, Pioglitazone, and Troglitazone may therefore be useful in treating an LPS exhibiting markers of de-differentiation.

[0208] Example 6. PPARG isoform expression in DDLPS compared to WDLPS and normal adipose

[0209]

[0187] This example demonstrates that the PPARG isoform PPARG2 is transcriptionally silenced in DDLPS cells relative to WDLPS cells.

[0210]

[0188] Examples 2 and 4 show that expression of PPARG is reduced in DDLPS. In this example, further experiments were conducted to assess the expression of each of the two human isoforms of PPARG, i.e., PPARG1 and PPARG2, in human normal adipose cells, WDLPS cells, and DDLPS cells.

[0211]

[0189] Expression of PPARG1, PPARG2, and the terminal differentiation marker ADIPOQ was assessed using qPCR for the DDLPS-derived LPS6 cell line and for non-transformed human adipose-derived mesenchymal stem cells (MSCs) (ATCC: PCS-500-011) at “baseline” (i.e., cultured in maintenance medium before exposure to IGF1 differentiation medium (“IGF 1 -DM”)), or after exposure to IGF 1 -DM. Non-transformed human adipose- derived mesenchymal stem cells (MSCs) served as a positive control for the expected transcriptional response to IGF 1 -DM. The IGF-1 DM was 10% charcoal stripped FBS with a cocktail of dexamethasone (Dex 1 mM), indomethacin (IDM, 50 mM), 3-isobutyl-l- methylxanthine (IB MX, 0.5 mM), and human recombinant IGF-1 (Miltenyi Biotec, 200 nM). For the differentiation protocol, cells were seeded at 30-50% confluency and treated with IGF 1 -DM. IGF 1 -DM was replaced every three days. Differentiation induction was measured at day 3.

[0212]

[0190] The nucleic acid sequences of the primers used to measure expression of each gene / isoform are shown in Table 3. Expression was normalized to ?-actin expression.

[0213] Table 3: Primer sequences.

[0214]

[0191] The qPCR results for the MSCs, i.e., the positive control cells, are shown in FIG. 6A. The qPCR results for DDLPS-derived LPS6 cells are shown in FIG. 6B. At baseline, LPS6 cells exhibited lower levels of PPARG1 and PPARG2 relative to the undifferentiated MSCs. Small increases in PPARG1, PPARG2, and ADIPOQ were observed in LPS6 cells after exposure to IGF 1 -DM. LPS6 cells exhibited smaller increases in PPARG1, PPARG2, and ADIPOQ upon exposure to IGF 1 -DM than MSC control cells.

[0215]

[0192] To ensure that the observed deficit of PPARG2 in the LPS6 cell line was relevant to human LPS tissue, qPCR analysis to assess the expression of PPARG1 and PPARG2 was also performed in human normal adipose, WDLPS, and DDLPS tissue samples. The status of the samples was determined by independent pathologist review. Human normal adipose tissue samples were obtained from patients who had surgery for non-sarcoma reasons. All tissue samples taken for the qPCR analysis described in this Example were immediately adjacent to the samples taken for sequencing described in Example 1.

[0216]

[0193] Expression of PPARG1 was observed to be about fifteen-fold less in DDLPS tissue (“DDLPS”) compared to normal adipose tissue (“adipose”) and about eight-fold less in DDLPS tissue compared to WDLPS tissue (“WDLPS”) (FIG. 7A). The expression of PPARG2 was nearly 200-fold less in DDLPS tissue compared to adipose and WDLPS tissue (FIG. 7B). These data confirm that loss of PPARG2 expression is a feature of human DDLPS.

[0217]

[0194] The loss of PPARG2 expression in DDLPS may be due to epigenetic silencing of the PPARG isoform 2 transcription start site (TSS). Analysis of the snATAC-seq data generated in Example 1 illustrated that there are two differentially accessible peaks (DAPs) within a 3000 base pair region of the PPARG gene body between WDLPS and DDLPS (FIG. 8). The position of these peaks is shown in FIG. 8 as the two tall pale grey bars in the DDLPS panel. The observed DAPs indicate that WDLPS cells had accessible chromatin over a promoterlike sequence spanning nucleotides 12351171-12351492 on chromosome 3, which is adjacent to the TSS of PPARG2. The region of the promoter-like sequence was inaccessible in DDLPS cells. The position of the TSS is shown in FIG. 8 by the black arrow. The position of the promoter-like sequence adjacent to the TSS is shown in FIG. 8 by the grey block to the left side of the arrow showing the TSS.

[0218]

[0195] These data suggest that chromatin compaction over a promoter region proximal to the PPARG2 TSS may contribute to the lack of PPARG2 expression in DDLPS and a diminished capacity to respond to pro-adipogenic signals. Furthermore, the presence or absence of PPARG2 may serve as a major switch between the well and de-differentiated subtypes in human LPS.

[0219]

[0196] This example demonstrates the significant reduction in expression of PPARG1 and PPARG2 in DDLPS compared to WDLPS and normal adipose tissue. These data therefore indicate that expression of PPARG1 and / or PPARG2 can be used to determine whether an LPS is WDLPS or DDLPS. Furthermore, these data also indicate that expression of PPARG1 and / or PPARG2 may be used to monitor the status of an LPS tumor, e.g., to monitor if the tumor is WDLPS is progressing or has progressed to a DDLPS. Since reduction is PPRAG2 in DDLPS compared to WDLPS and normal adipose tissue was particularly prominent, likely due to epigenetic changes in the PRRAG2 promoter, this isoform may be a particularly useful marker to differentiate DDLPS from WDLPS.

[0220] Example 7. Combination of IGF1 and a PPAR-y agonist promotes LPS tumor cell adipocytic differentiation

[0221]

[0197] This example demonstrates that treating LPS tumor cells with a combination of recombinant IGF1 and a PPARG agonist increases adipocytic differentiation.

[0198] In Example 5, it was shown that applying recombinant human IGF1 to LPS6 cells increased expression of PPARG and adiponectin. This led to the hypothesis that recombinant human IGF1 (or an IGF1 mimetic such BVS857 and Nap-FFG-GYGSSSRRAPQT) and / or a selective PPAR-y agonist may be used to induce differentiation of de-differentiated LPS tumor cells (e.g., DDLPS tumor cells) toward a less malignant adipocyte phenotype. To test this hypothesis, LPS6 cells, and MSCs as positive control, were treated with IGF 1 -DM as described in Example 6 alone or in combination with the PPAR-y agonist rosiglitazone. Specifically, the IGF 1 -DM was supplemented with 1 pM rosiglitazone. Subsequently, expression of PPARG1, PPARG2 and ADIPOQ was assessed using qPCR.

[0222]

[0199] The qPCR results for the MSCs, i.e., the positive control cells, are shown in FIG. 9A. The qPCR results for LPS6 cells are shown in FIG. 9B. Treating LPS6 cells with a combination of the PPAR-y agonist rosiglitazone and recombinant IGF1 increased the expression of adipocyte differentiation marker ADIPOQ by 10-fold on average compared to treating LPS6 cells with recombinant IGF1 alone (FIG. 9B).

[0223]

[0200] The data present in this example suggest that administering a subject having LPS with a therapeutically effective amount of a selective PPAR-y agonist and a therapeutically effective amount of recombinant human IGF 1 or an IGF 1 mimetic may initiate differentiation of DDLPS tumor cells toward an adipocyte phenotype, or maintain differentiation of WDLPS tumor cells and thus prevent the emergence of a DDLPS phenotype.

[0224] Example 8. Combination of IGF1 and a GLP-1 agonist promotes LPS tumor cell adipocytic differentiation

[0225]

[0201] This example demonstrates that treating LPS tumor cells with a combination of recombinant IGF1 and a glucagon-like peptide-1 (GLP-1) agonist increases adipocytic differentiation.

[0226]

[0202] An epigenetic enrichment of GLP-1 -induced regulation of insulin signaling was observed in DDLPS (data not shown). Therefore, it was investigated whether a GLP-1 agonist such as liraglutide could promote tumor cell differentiation in DDLPS cultured in IGF1-DM. Liraglutide is a long-acting GLP-1 agonist considered to promote adipocyte differentiation by propagating insulin signaling (Challa et al. J. Biol. Chem. (2011) 287, 6421 and Zhang et al. iScience (2021) 24, 103382).

[0203] LPS6 cells, and MSCs as control, were treated with IGF 1 -DM as described in Example 6 alone or in combination with liraglutide, and the expression of PPARG1, PPARG2 and ADIPOQ was assessed using qPCR as described in that example. Specifically, the IGF 1 -DM was supplemented with 100 nM of liraglutide. The qPCR results for the MSCs, i.e., the positive control cells, are shown in FIG. 10A. The qPCR results for LPS6 cells are shown in FIG. 10B. Treating LPS6 cells with a combination of the GLP-1 agonist liraglutide and recombinant IGF1 increased the expression of adipocyte differentiation marker ADIPOQ by 50-fold on average compared to treating LPS6 cells with recombinant IGF1 alone (FIG. 10B).

[0227]

[0204] Similar to Example 7, the data in this example suggest that administering a subject having LPS with a therapeutically effective amount of a selective GLP-1 agonist and a therapeutically effective amount of recombinant human IGF1 or an IGF1 mimetic may initiate differentiation of DDLPS tumor cells toward an adipocyte phenotype, or maintain differentiation of WDLPS tumor cells and thus prevent the emergence of a DDLPS phenotype.

[0228] Example 9. IGF1R-ADC specifically targets LPS cells

[0229]

[0205] This example demonstrates that an IGF1R antibody-drug conjugate (ADC) can be used to specifically target and reduce the viability of LPS cells.

[0230]

[0206] As shown in FIG. 4A, DDLPS and WDLPS tumors comprise a greater fraction of cells that express IGF1R than normal adipose tissue. Furthermore, the mean IGF1R expression in the IGFIR-expressing DDLPS and WDLPS cells is higher than in normal adipose tissue. This was most prevalent in the DDLPS cohort.

[0231]

[0207] To assess whether IGFR1 overexpression could be exploited therapeutically, an IGFIR-targeted antibody drug conjugate (IGF 1R- ADC) and an IgGl isotype control ADC were constructed as described in Akla et al. Mol Cancer Ther (2020) 19 (1): 168-177. The antibodies were purchased from Creative Biolabs (Shirley, NY). Specifically, the anti-IGFIR antibody lonigutamab was linked to the anti-mitotic chemotherapeutic agent monomethyl auristain E (MMAE) with a drug-to-antibody ratio (DAR) of 4, to form the ADC W0101 (“IGF 1R- ADC”). Human IgGl isotype control antibody was also linked to MMAE with a DAR of 4 (“isotype control ADC”).

[0208] Each of these ADCs was tested on either WDLPS-derived 93T449 LPS cells, or human aortic endothelial cells (HAECs) as negative control. Cells were seeded in 96-well plates and allowed to adhere 16-24 hours prior to treatment. The cells then were exposed to 0, 2, 4, 6, 8 or 10 pg / ml of either IGF 1R- ADC or the isotype control ADC in complete cell culture medium for 72 hours. After 72 hours of treatment, cell viability was assessed using CyQUANT XTT Cell Viability Assay™ (Invitrogen) following the manufacturer's instructions. Specifically, cells were incubated with XTT working solution for 4 hours, and absorbance was measured at 450 nm and 660 nm using the CLARIOstar Microplate reader. 0 pg / ml isotype control ADC was set as 100% viability.

[0232]

[0209] The results for 93T449 cells are shown in FIG. 11 A, and the results for HAECs are shown in FIG. 1 IB. As expected, the isotype control did not consistently reduce the viability of either 93T449 LPS cells (FIG. 11 A, black squares) or HAECs (FIG. 11B, black squares). IGF1R ADC did not reduce the viability of HAECs (FIG. 11B, gray circles). Conversely, 2 pg / ml of the IGF1R ADC reduced the viability of 93T449 LPS cells to about 60%, an effect that was maintained with concentrations of 4, 6, 8 or 10 pg / ml of IGF1R ADC.

[0233]

[0210] This example demonstrates that an IGF1R antibody-drug conjugate (ADC) can be used to specifically target and reduce the viability of LPS cells.

[0234] Example 10. IGF1 is predictive of LPS patient overall survival

[0235]

[0211] This example demonstrates that patients having LPS tumors having a low IGF1 expression level (e.g., LPS tumors with an IGF1 expression level below that of the median of a sample of LPS tumors) have a decreased survival probability compared to LPS patients having LPS tumors having a high IGF1 expression level (e.g., LPS tumors with an IGF1 expression level above that of the median of a sample of LPS tumors).

[0236]

[0212] As explained in Example 4 and shown in FIG. 4B, when cells from the multiome snRNA-seq cohort were plotted for their “IGF1 signaling score”, unexpectedly and surprisingly, DDLPS tumor cells separated into two populations.

[0237]

[0213] Given the presence of IGF1 expressing cells in the DDLPS cohort, it was hypothesized that IGF 1 -expressing DDLPS tumors may exhibit clinical behavior more consistent with WDLPS and therefore confer a better patient prognosis.

[0214] Patient outcomes and IGF1 expression were interrogated for the multi ome cohort established in Example 1. WDLPS and DDLPS patient outcome data were coupled with IGF 1 expression as provided by the snRNA-seq data. Specifically, for each patient’s tumor sampled in the snRNA-seq data, relative mean IGF1 expression was calculated. This analysis indicated that patients with a DDLPS retaining the highest IGF 1 -expression levels remained recurrence free for more than 5 years, whereas patients with DDLPS who had lower IGF1 expression experienced recurrence and / or disease-related death. A similar pattern was seen in WDLPS patients.

[0238]

[0215] To validate these initial observations in a larger LPS cohort, bulk RNA-seq data and patient outcomes data obtained from The Cancer Genome Atlas (TCGA) for LPS patients were interrogated. This dataset comprised data obtained from tumors sampled from 58 patients with DDLPS and 1 patient with WDLPS. It was observed that patients with IGF1 expression above the median (i.e., “high”) had significantly longer overall survival than patients with IGF1 expression below the median (i.e., “low”) (FIG. 12).

[0239]

[0216] Thus, IGF1 might orchestrate a more differentiated tumor cell transcriptional program, leading to better prognosis in patients with WDLPS and DDLPS tumors expressing high levels of IGF 1.

[0240]

[0217] This example confirms that IGF1 expression is a useful marker to assess survival probability in patients suffering from LPS. The data in this example further indicate that the level of IGF1 expression in a sample obtained from a subject’s LPS relative to the mean expression level of IGF1 in a control LPS patient population may be used to stratify the subject for therapeutic intervention, e.g., chemotherapy.

[0241]

[0218] It should be understood that the details provided herein are given by way of illustration only, not limitation. Other features, objects, and advantages are apparent from the above detailed description, drawings and examples. Various changes and modifications will be apparent to those skilled in the art.

[0242]

[0219] All patents, patent publications, and non-patent publications referenced herein are indicative of the level of skill of those skilled in the art to which this invention pertains. All these publications are herein incorporated by reference to the same extent as if each individual publication were specifically and individually indicated as being incorporated by reference.

Claims

CLAIMS1. A method of characterizing a subject’s liposarcoma (LPS) comprising determining expression levels of Insulin-like Growth Factor 1 (IGF1), Insulin-like Growth Factor 1 Receptor (IGF1R), Peroxisome Proliferator Activated Receptor Gamma 1 (PPARG1) and / or Peroxisome Proliferator Activated Receptor Gamma 2 (PPARG2) in a sample of the LPS obtained from the subject.

2. A method of identifying a subject suffering from a liposarcoma (LPS) with metastatic potential, the method comprising determining expression levels of Insulin-like Growth Factor 1 (IGF1), Insulin-like Growth Factor 1 Receptor (IGF1R), Peroxisome Proliferator Activated Receptor Gamma 1 (PPARG1) and / or Peroxisome Proliferator Activated Receptor Gamma 2 (PPARG2) in a sample of the LPS obtained from the subject.

3. The method of claim 1 or 2, wherein the sample is a resected tumor or a tissue biopsy.

4. The method of any one of claims 1-3, wherein expression levels are determined using antibodies, wherein each antibody specifically binds to one of the markers.

5. The method of any one of claims 1-3, wherein expression levels are determined using nucleic acid probes, wherein each nucleic acid probe specifically hybridizes to one of the mRNAs encoding the markers.

6. The method of any one of the preceding claims, further comprising comparing the expression levels with a control sample.

7. The method of claim 6, wherein the control sample is at least one sample obtained from a subject suffering from well-differentiated liposarcoma (WDLPS) or normal tissue.

8. The method of claim 6, wherein the control sample is a reference data set of samples obtained from patients suffering from WDLPS or normal tissue.

9. The method of any one of the preceding claims, wherein: a. a 2-fold or more increase in the expression of IGF1R in the sample obtained from the subject compared to the control; and / or b. a 50% or more decrease in the expression of PPARG1, PPARG2 and / or IGF1 in the sample obtained from the subject compared to the control sample is indicative of a LPS with metastatic potential.

10. The method of any one of the preceding claims, wherein the method further comprises determining the expression levels of Lipoprotein Lipase (LPL) and / or Fatty Acid Binding Protein 4 (FABP4).

11. The method of claim 10, wherein a 50% or more decrease in the expression of FABP4 and / or LPL in the sample obtained from the subject compared to the control sample is indicative of a LPS with metastatic potential.

12. A method of treating a subject suffering from a liposarcoma (LPS) with chemotherapy, comprising administering to the subject a therapeutically effective amount of the chemotherapy, wherein the subject is selected for chemotherapy if the LPS has: a. a 2-fold or more increase in the expression of Insulin-like Growth Factor 1 Receptor (IGF1R) in the sample obtained from the subject compared to a control; and / or b. a 50% or more decrease in the expression of Peroxisome Proliferator Activated Receptor Gamma 1 (PPARG1), Peroxisome Proliferator Activated Receptor Gamma 2 (PPARG2) and / or Insulin-like Growth Factor 1 (IGF1) in the sample obtained from the subject compared to the control sample.

13. A method of selecting a subject with liposarcoma (LPS) for chemotherapy, comprising determining expression levels of Insulin-like Growth Factor 1 (IGF1), Insulin-like Growth Factor 1 Receptor (IGF1R), Peroxisome Proliferator Activated Receptor Gamma 1 (PPARG1) and / or Peroxisome Proliferator Activated Receptor Gamma 2 (PPARG2) in a sample of the LPS obtained from the subject compared to a control sample; wherein the subject is selected for chemotherapy if the sample has:a. a 2-fold or more increase in the expression of IGF1R in the sample obtained from the subject compared to the control; and / or b. a 50% or more decrease in the expression of PPARG1, PPARG2 and / or IGF1 in the sample obtained from the subject compared to the control sample.

14. The method of claim 12 or 13, wherein the method further comprises determining the expression levels of Lipoprotein Lipase (LPL) and / or Fatty Acid Binding Protein 4 (FABP4).

15. The method of claim 14, wherein the subject is selected for chemotherapy if the sample has a 50% or more decrease in the expression of FABP4 and / or LPL in the sample obtained from the subject compared to the control sample.

16. The method of any one of claims 12-15, wherein the chemotherapy comprises administering to the subject a therapeutically effective amount of one or more of: a. a selective insulin-like growth factor-I receptor kinase inhibitor, optionally Linsitnib or cyclolignan PPP; b. a selective PPAR-y agonist, optionally selected from AMG-131, Muraglitazar, Lobeglitazone, Naveglitazar, Rosiglitazone, Pioglitazone, and Troglitazone; c. recombinant human IGF1 or an IGF1 mimetic, optionally wherein the IGF1 mimetic is BVS857 or Nap-FFG-GYGSSSRRAPQT; and d. an antibody-drug conjugate (ADC) that specifically binds to the extracellular region of IGF1R, optionally comprising lonigutamab (hz208F2-4), ganitumab, teprotumumab, figitumumab, dalotuzumab, robatumumab, or veligrotug.

17. A method of treating a subject with liposarcoma (LPS) comprising administering a therapeutically effective amount of a selective insulin-like growth factor-I receptor kinase inhibitor.

18. The method of claim 17, wherein the selective insulin-like growth factor-I receptor kinase inhibitor is Linsitinib or cyclolignan PPP.

19. A method of treating a subject with liposarcoma (LPS) comprising administering a therapeutically effective amount of a selective PPAR-y agonist.

20. The method of claim 19, wherein the selective PPAR-y agonist is selected from the group consisting of AMG-131, Muraglitazar, Lobeglitazone, Naveglitazar, Rosiglitazone, Pioglitazone, and Troglitazone.

21. The method of claim 19 or 20, further comprising administering to the subject a therapeutically effective amount of recombinant human IGF1 or an IGF1 mimetic.

22. A method of treating a subject with liposarcoma (LPS) comprising administering a therapeutically effective amount of recombinant human IGF1 or an IGF1 mimetic.

23. The method of claim 22, further comprising administering to the subject a therapeutically effective amount of: a. a selective PPAR-y agonist, optionally selected from AMG-131, Muraglitazar, Lobeglitazone, Naveglitazar, Rosiglitazone, Pioglitazone, and Troglitazone; and / or b. a selective GLP-1 agonist, optionally liraglutide, exenatide, albiglutide, dulaglutide, lixisenatide, and semaglutide.

24. The method of any one of claims 21-23, wherein the IGF1 mimetic is selected from the group consisting of BVS857 and Nap-FFG-GYGSSSRRAPQT.

25. A method of treating a subject with liposarcoma (LPS) comprising administering a therapeutically effective amount of an antibody-drug conjugate (ADC), wherein the antibody specifically binds to the extracellular region of IGF1R.

26. The method of claim 25, wherein the ADC is lonigutamab ugodotin (W0101).

27. The method of claim 25 or 26, wherein the antibody is selected from the group consisting of lonigutamab (hz208F2-4), ganitumab, teprotumumab, figitumumab, dalotuzumab,robatumumab, and veligrotug.

28. The method of any one of claims 25-27, comprising administering a therapeutically effective amount of: a. a selective PPAR-y agonist, optionally selected from AMG-131, Muraglitazar, Lobeglitazone, Naveglitazar, Rosiglitazone, Pioglitazone, and Troglitazone; b. a selective GLP-1 agonist, optionally selected from liraglutide, exenatide, albiglutide, dulaglutide, lixisenatide, and semaglutide; and / or c. a recombinant human IGF1 or an IGF1 mimetic, optionally wherein the IGF1 mimetic is BVS857 or Nap-FFG-GYGSSSRRAPQT.

29. The method of any one of claims 17-28, wherein the LPS is characterized by: a. a 2-fold or more increase in the expression of IGF1R compared to a control; and / or b. a 50% or more decrease in the expression of PPARG1, PPARG2 and / or IGF1 compared to the control.

30. The method of claim 29, wherein the LPS is characterized by a 50% or more decrease in the expression of FABP4 and / or LPL compared to the control.

Images