Introduction

Soft tissue sarcomas are rare malignancies originating from the mesoderm. There exist more than 50 histological subtypes based on their tissues of origin. Undifferentiated pleomorphic sarcoma (UPS) comprises a subset of 5 to 15% of soft tissue sarcomas for which clinicians cannot determine the tissue of origin from analysis of the tumor’s molecular and histologic characteristics1,2,3. A subset of UPS, unclassified spindle cell sarcomas (USC), are rare high-grade malignancies with an unknown cell of origin4,5. The standard of treatment for UPS is surgical resection. For patients in whom surgery is not feasible, chemotherapy and radiation therapy may be used, but they have low-to-moderate response rates6,7. UPS is characterized by a high level of genomic instability, as indicated by its complex karyotype8,9,10, yet specific recurring genetic alterations are uncommon11. UPS has a 5-year overall survival rate of 60% and a 10-year overall survival rate of 48%12. Malignant peripheral nerve sheath tumors (MPNSTs), which comprise 2% of all soft tissue sarcomas1, are another rare and genomically unstable sarcoma13,14. Half of MPNSTs occur in patients with the autosomal dominant cancer predisposition syndrome neurofibromatosis type 1 (NF1; OMIM # 162200). Approximately 40% of MPNST cases are sporadic, and the remaining 10% are associated with previous radiation therapy15,16,17. Chemotherapy and radiation therapy are ineffective for treating MPNSTs2,18,19; therefore, surgical resection remains the standard treatment20,21,22,23. MPNSTs have a high rate of local recurrence24,25. The 5-year survival rate is 35 to 50%, and the 10-year survival rate is 7.5%, stressing the need for improved treatment options17,26. Limited treatment options are available for patients with UPS, USC and MPNST, largely because the mechanisms involved in the progression of these heterogeneous sarcomas require further investigation17,27,28,29.

UPS and MPNST both have complex karyotypes, and no specific genomic alterations are unique to the tumor type9,10,30,31. UPS has complex cytogenetic rearrangements in 30 to 35% of the genome32, but to date, no specific alterations have been useful for classifying UPS. Loss of chromosome 13q, which leads to inactivation of the retinoblastoma (RB) pathway, is frequently observed in UPS33,34,35, as are losses of genes in the TP53/ARF/MDM2 pathway36,37. A study of soft tissue sarcoma by The Cancer Genome Atlas Research Network10 reported several somatic copy number alterations, specifically deep deletions in RB1 in 16%, TP53 in 16%, and CDKN2A in 20% of UPS. The mutation burden was low in soft tissue sarcomas, an average of 1.06 per Mb in 206 tumors. The significantly mutated genes in UPS included RB1 and potential driver truncating mutations of NF110. MPNST’s complex genomic profile includes copy number aberrations, like microdeletions of NF1 in 5 to 20% of cases38,39, as well as CDKN2A inactivation40,41, and TP53 mutations. Recently, mutations in the genes of the polycomb repressive complex 2 (PRC2) components SUZ12 and EED and the loss of histone 3 lysine 27 trimethylation (H3K27me3) were proposed as markers for NF1-related and sporadic MPNSTs28,42.

Cell line models that can form tumors in mice are valuable tools to advance the understanding of soft tissue sarcomas with no known genetic drivers such as UPS and MPNST. UPS cell lines have been developed by our laboratory (UPS186 and RIS81929) and others43,44, some with the ability to form tumors when injected subcutaneously into mice45,46. Well-studied NF1-associated MPNST cell lines are also available within the scientific community, including SNF02.2 and SNF94.3, both derived from a metastatic site in the lung, SNF96.2, isolated from a recurrent MPNST; and S462, ST88-14, T265, and MPNST64247,48,49,50. Longo et al. recently developed and characterized a sporadic MPNST cell line, 2XSB51. The NF1-associated MPNST cell lines listed above can form tumors in mice, allowing for in vivo study of these difficult-to-treat sarcomas47,48,52,53,54. Multiple new NF1-MPNST cell lines and PDX models, with corresponding clinical and genomic annotations, are now available for preclinical studies55. However, further development of UPS and MPNST resources is warranted because cell lines from genomically complex sarcomas are underrepresented in cancer cell line databases such as the Cancer Cell Line Encyclopedia and Catalog Of Somatic Mutations In Cancer56,57. While some sarcoma cell lines are listed in large-scale systematic databases like The Cancer Dependency Map (depmap)58, they have not been studied as extensively as those from other lineages. Broad systematic approaches that include gene-dependency screens and small molecule sensitivity data (among other—omics data analyses)58,59, when applied to poorly understood sarcomas like those presented here could identify unique targets and avenues for drug discovery.

This study aimed to generate and characterize new cellular models of UPS and MPNST that will deepen our understanding of the aberrant cellular processes of these sarcomas known to lack clear genetic or molecular markers. The UPS and MPNST cell lines generated in this study will be available to the scientific community to facilitate the study of these debilitating tumors.

Materials and methods

UPS and MPNST primary culture and cell lines

We obtained approval from the Institutional Review Board of The University of Texas MD Anderson Cancer Center and patients’ informed written consent before establishing the cell lines described in this study. The cell lines were isolated from surgical specimens from patients at MD Anderson Cancer Center. A sarcoma pathologist confirmed the final histology of the tumors at MD Anderson Cancer Center. A portion of the original specimen was fixed in formalin and paraffin-embedded, followed by immunohistochemical (IHC) staining for hematoxylin and eosin to visualize the samples. Isolation of tumor cells for RIS620, UPS271.1, and UPS511 was described previously by May et al.29 USC060.1 and UPS0103 cells were isolated as described by Peng et al.60. Briefly, fresh sterile samples from resected tumors were minced in Dulbecco’s modified Eagle medium (DMEM; Corning, cat # 10013CV and incubated with collagenase type I (3%; Sigma-Aldrich, cat # C-0130), DNase I (0.02%; Sigma-Aldrich, cat # D-4527), and hyaluronidase (1.5 mg/ml; Sigma-Aldrich, cat # H-3506) at 37 °C for 2 to 4 h. After digestion, the sample was strained through a mesh screen (BD Biosciences, cat # 352360), and undigested tissue was discarded. The sample was centrifuged, washed, resuspended in phosphate-buffered saline (PBS; Corning, Corning, NY, cat # 21040CV), and gently transferred to tubes containing 10 ml of a 100% polysucrose/sodium diatrizoate solution (Histopaque-1077; Sigma-Aldrich, cat # H8889) overlaid with 15 ml of 75% Histopaque-1077. After centrifugation at 4 °C for 30 min at 1200 × g, the tumor cells in the top interface above the 75% Histopaque were collected and plated on cell culture plates. After establishing cells that grew in a monolayer, the morphology of the cell lines was evaluated by microscopy.

UPS cells were cultured in DMEM supplemented with 20% FBS, 1% penicillin/streptomycin (100 U/ml), and 1% L-alanyl-L-glutamine (glutagro; Corning, cat # 25-015-CI). The USC060.1 and UPS271.1 nomenclature denotes that these cell lines were established from a patient-derived xenograft tumor that arose from a subcutaneous injection of either USC060.1 cells or UPS271.1 cells into a mouse. MPNST3813E cells and MPNST007 cells were cultured in DMEM supplemented with 10% FBS and 1% penicillin/streptomycin (100 U/ml). MPNST4970 was cultured in DMEM/F12 (Corning, cat # 10090CV) supplemented with 20% FBS, 1% penicillin/streptomycin (100 U/ml), 1% glutagro, and 1% nonessential amino acids (Corning, cat # 25-025-CI), and Schwann cell growth supplement (ScienCell Research Laboratories, cat # 1752). Isolation of MPNST642 cells was described previously47, and this cell line was characterized further in this study by sequencing analysis. Cells were cultured in a humidified incubator at 37 °C with 5% CO2. Cells were tested for mycoplasma every 6 months.

Doubling time was measured by determining cell confluence using an Incucyte system (Essen Bioscience, Ann Arbor, MI). UPS or MPNST cells were plated in 24-well plates or 96-well plates at 10 to 30% confluence and imaged every 4 h until they reached confluency. The doubling time was calculated using a nonlinear regression fit for exponential growth in Prism 7 software (GraphPad Software, La Jolla, CA).

Short tandem repeat (STR) DNA profiling was performed using a Powerplex 16 HS kit (Promega, cat # DC2101). Genomic DNA was extracted from the tumor specimens and the cultured cell lines using a QIAmp mini DNA isolation kit (Qiagen, cat # 51304). Sixteen loci were screened for regions of microsatellite instability with defined trinucleotide or tetranucleotide repeats within the chromosomes. The STR results for each cell line were authenticated against the corresponding primary tumor’s STR profile and internal database of over 4000 public profiles and profiles unique to cell lines developed by MDACC investigators. To comply with the Health Insurance Portability and Accountability Act of 1996 and MD Anderson institutional policy, only a selection of human loci is shown in Table 2. The alleles from 8 human loci are reported in this study: AMEL, CSF1PO, D13S317, D16S539, D5S818, D7S820, TH01, and TPOX.

Normal Schwann cells (NSC) served as control cells. NSC were cultured in Schwann Cell Medium (ScienCell Research Laboratories, cat # 1700 and 1701). Human bone-marrow-derived mesenchymal stem cells (MSC) were cultured in mesenchymal stem cell medium (PromoCell cat # C12974 and C28009) and used as a reference for MDM2 amplification. DDLPS-224A cells were cultured in DMEM supplemented with 20% FBS and 1% penicillin/streptomycin (100 U/ml) and used as a positive control for MDM2 amplification61.

Clonogenicity and migration assays

Clonogenicity was assayed as described by May et al.29. Briefly, cells were seeded onto 6-well plates at low densities of 5000 cells for UPS cell lines and 500 to 2000 cells for MPNST cells (depending on the cell line) and were allowed to form colonies for 14 to 30 days. Images were obtained using a digital scanner (Konica Minolta, Tokyo, Japan) or a BZ-X800 microscope (Keyence Corporation, Itasca, Illinois). Migration assays were performed using a modified Boyden chamber assay with Transwell inserts containing polyethylene terephthalate membranes with 8-µm pores (Corning, cat # 354578). For clonogenicity and migration assays, cells were fixed with 5% glutaraldehyde and stained with 0.1% crystal violet in 20% methanol. For migration assays, Images at ×20 magnification were taken with a Nikon microscope (Tokyo, Japan) or a BZ-X800 microscope (Keyence Corporation, Itasca, Illinois).

Protein expression assays

Sodium dodecyl sulfate-polyacrylamide gel electrophoresis and Western blotting were performed as described by May et al.29. Briefly, proteins were extracted from cultured cells in radioimmunoprecipitation buffer (RIPA) containing protease inhibitors (Sigma-Aldrich, cat # 11697498001). Thirty micrograms of protein were loaded on to 4 to 15% precast polyacrylamide gels (Bio-Rad, cat# 4561085) gels for PRC2 components and p53. Acid extraction following the standard protocol (Abcam) was performed on cultured cells to isolate histones, and 3 μg of protein was separated on 4 to 20% precast polyacrylamide gels (Bio-Rad, cat # 4561095) for detection of histone H3 and H3K27me3. Proteins were transferred onto low-fluorescence polyvinylidene fluoride membranes using a semidry transfer method according to the manufacturer’s instructions (Bio-Rad, cat # 1704274). Membranes were blocked in 5% nonfat dry milk before probing with the specific antibodies listed in Supplementary Table 1. Horseradish peroxidase-conjugated secondary antibodies were detected by enhanced chemiluminescence (GE Healthcare, cat # RPN2236) and imaged using a Chemidoc imager (Bio-Rad).

In vivo xenograft studies

The MD Anderson Institutional Animal Care and Use Committee approved all animal procedures. We proceeded with patient-derived xenograft (PDX) implant generation when we received <1 g of residual tumor; if the residual tumor weighed >2 g, we generated a PDX implant and established a cell line. For PDX implantation experiments, 6-week-old male or female NOD.Cg-Prkdcscid Il2rgtm1Wjl/SzJ (NSG) mice (Jackson Laboratory, Bar Harbor, ME) were implanted with pieces of MPNST007, MPNST3813E, RIS620, or UPS0103 tumor (~3–5 mm in diameter) in the flank. For injections of USC060.1 cells, 3 × 106 cells in 1:1 PBS:Matrigel (Corning, cat #354248) (0.1 ml/mouse) Trypan blue viable cells were injected subcutaneously into the flank of 6-week-old male or female NSG mice. For the UPS511 cell line, 5 × 106 Trypan blue viable cells in 1:1 PBS:Matrigel (Corning, cat #354248) (0.1 ml/mouse) were injected subcutaneously into the flank of 6-week-old male or female NSG mice. Mice were monitored for tumor size and well-being and were euthanized when tumors reached an approximate volume of 1000–2000 mm3. Tumor growth was measured once per week using a caliper. The mice were not assessed for metastases. Figures show data from at least five representative animals. Resected tumors were weighed, fixed in formalin and paraffin-embedded for IHC studies. IHC studies including hematoxylin and eosin were performed to compare the xenograft specimens to the original patient tumors. The percentage uptake was calculated by dividing the number of mice that grew tumors by the total number of mice implanted or injected with cells. The latency period was estimated from the tumor growth data.

Sequencing

Targeted sequencing was performed for a panel of 4 genes frequently mutated in UPS and MPNST according to TCGA: TP53, ATRX, RB1 and CDKN2A, using genomic DNA isolated from cultured cells with a Qiagen QIAmp mini DNA isolation kit10,62. DNA was fragmented and bait-captured according to the manufacturer’s guidelines63. Libraries were sequenced on a HiSeq2000 or HiSeq4000 sequencer (Illumina, San Diego, CA) with 76-bp paired-end reads for MPNST007, MPNST3813E, UPS0103, and RIS620; and 76-bp single reads for USC060.1, UPS271.1, and UPS511. The raw data were converted to fastq format and aligned to the reference genome (hg19) with the Burroughs–Wheeler Aligner64. Median target coverage was 256 × to 763× for MPNST007, MPNST3813E, MPNST4970; and UPS0103 cells and 207× to 230× for USC060.1, UPS271.1, and UPS511 cells. Mutations were called against a common normal lung sample using the MuTect and Pindel tools65,66. Sanger sequencing of TP53 was performed with genomic DNA from the original MPNST specimen using either an ABI 3730XL or 3730 DNA sequencer (Thermo Fisher Scientific, Waltham, MA) with a Big Dye Terminator cycle sequencing chemistry kit (Thermo Fisher Scientific, cat # 4337455) according to the manufacturer’s guidelines.

Copy number alteration analysis

MDM2 copy number analysis was performed using the MDM2 TaqMan Copy Number Assay Hs02970282_cn (Thermo Fisher Scientific cat # 4400291) and the reference assay for TERT (Thermo Fisher Scientific cat # 4403316) according to manufacturer’s guidelines. Results were analyzed using CopyCaller Software v2.1 (Thermo Fisher Scientific, Waltham, MA).

Results

Establishment and morphology of UPS and MPNST cell lines

All cell lines were established from cells isolated from residual resected tumors. The UPS cell lines were generated from recurrent or metastatic tumors, and the MPNST tumors (all three NF1-associated) were generated from primary or recurrent tumors (Table 1). A sarcoma pathologist at MD Anderson Cancer Center confirmed the final histology of the tumors (AJL). The STR profiles of the established cell lines matched those of the original patient tumors, indicating that each of the cell lines was derived from the tumor of origin, and there was no cross-contamination with other cell lines. The hematoxylin and eosin staining of the original specimens showed a mixture of tumor cells and stromal cells in UPS and MPNST tumors (Fig. 1a, lower panels). The cells were all mononuclear. USC060.1 and UPS511 cell lines had an elongated spindle-like shape, while UPS271.1 and UPS0103 cell lines were polygonal and RIS620 cells exhibited both polygonal and spindle shapes (Fig. 1a, upper panels). MPNST007 cell lines had an elongated spindle shape, whereas MPNST3813E cells showed more polygonal morphology in comparison, as did MPNST4970 cells (Fig. 1a, upper panels). Table 2 shows STR profiles from the resected tumor and corresponding cell line.

Table 1 Characteristics of UPS and MPNST patients from whom tumors were resected and cell lines derived.
Full size table
Fig. 1: Characterization of cell lines and original tumors.
figure 1

a Bright-field images of UPS and MPNST cell lines (upper panels) and images of hematoxylin and eosin (H&E)-stained sections of the original patient tumors (lower panels), ×20 magnification. Scale bar represents 100 µm. b MDM2 copy number analysis in UPS cell lines. MSC mesenchymal stem cells, were used as a reference. DDLPS-224A cells were used as a positive control for MDM2 amplification. TERT was used as the reference gene for the MDM2 copy number assay. N = 2, graph shows calculated copy number; error bars indicate copy number range. c Immunoblots showing expression of EZH2, SUZ12, EED, RbAP46/48, and H3K27me3 in normal Schwann cells (NSC), MPNST007, MPNST3813E, MPNST642, and MPNST4970 cells. EZH2, SUZ12, EED, and RbAP46/48 levels were normalized to actin (middle panels). H3K27me3 expression was normalized to H3 levels (lower two panels).

Full size image
Table 2 Short tandem repeat (STR) fingerprinting of UPS and MPNST cell lines.
Full size table

As of November 2021, USC060.1 has been carried to passage 50, UPS271.1 to passage 36, UPS511 to passage 35, and UPS0103 to passage 17. MPNST007 and MPNST3813E have been taken over passage 100, MPNST4970 has been carried to passage 41. Our laboratory will continue to passage the lower passage UPS and MPNST cell lines to determine whether they can be carried to over 90 passages.

Misclassification of the UPS tumors as dedifferentiated liposarcoma (DDLPS) was ruled out by MDM2 copy number analysis using genomic DNA isolated from the established cell lines. Bone marrow-derived mesenchymal stem cells (MSC) served as the reference for MDM2 copy number, as they represented a potential cell of origin for UPS. DDLPS-224A, a DDLPS cell line with known amplification of MDM2, served as a positive control61. No amplification of MDM2 was detected in the UPS cell lines, providing evidence that the specimens from which these cell lines were derived were indeed UPS and not misclassified DDLPS.

Protein expression

Western blotting was used to assess the expression of histone marks and proteins known to be associated with MPNST: H3K27me3, and the PRC2 members SUZ12, EED, RbAp46/48, and EZH2. NSCs, considered the cell of origin for MPNST, served as control cells. The MPNST007 and MPNST4970 cell lines expressed higher levels of EZH2 than did the control NSCs. The MPNST cells were also assayed for the absence of H3K27me3, which is considered a potential diagnostic marker of NF1-associated MPNSTs28. The H3K27me3 mark was present in MPNST007 cells and MPNST4970 but not in MPNST642 or MPNST3813E cells (Fig. 1c, lower panels). These findings suggested that MPNST007 and MPNST4970 are PRC2-competent, and MPNST642 and MPNST3813E are PRC2-incompetent.

Cell proliferation rate

The proliferation rate for the UPS and MPNST cell lines was determined by measuring relative confluence (Fig. 2a, b). For the UPS cell lines, the doubling time ranged from approximately 33.3 to 279.8 h. USC060.1 cells had the shortest doubling time (33.3–53.5 h), and the doubling time of UPS271.1 cells was similar to that of USC060.1 cells (48.8–56.6 h). UPS511 cells and RIS620 cells had intermediate doubling times (90.6–92.0 h and 151.5–174.8 h, respectively). The UPS0103 cell line had the slowest doubling time of the UPS cells in this study, at 279.8 h. Of the MPNST cell lines, MPNST3813E cells doubled in 36.3 h (range 34.6 to 39.7 h). MPNST007 cells doubled in 49.4 to 58.9 h, and MPNST4970 doubled in ~84.6 h. These results showed that the cell lines established in the current study have a wide range of doubling times, reflecting these tumors’ heterogeneous nature.

Fig. 2: In vitro growth studies of UPS and MPNST cells.
figure 2

Doubling time for UPS (a) and MPNST (b) cell lines. Cells were plated and allowed to grow until they reached a growth plateau. Relative confluence was measured using an Incucyte device (Essen Bioscience). c Clonogenicity assay for UPS and MPNST cells. For UPS cells, 5000 cells were seeded and allowed to grow for the indicated number of days. For MPNST cells, 500 to 2000 cells were seeded and allowed to grow for 16 days. Migration of UPS cells and MPNST cells, ×20 magnification. d Targeted sequencing results for selected sarcoma genes. Refer to the results section for details on the mutations identified.

Full size image

In vitro tumorigenic potential of the new UPS and MPNST cell lines

The UPS and MPNST cell lines were assessed for clonogenic and migratory potential. For clonogenic assays, cells were seeded at a low density and allowed to form colonies for up to 16 days. All the UPS cell lines formed colonies. USC060.1, UPS271.1, and RIS620 formed small colonies, UPS511 formed large distinct colonies, and UPS0103 formed diffuse colonies (Fig. 2c, upper panels). MPNST007, and MPNST3813E formed distinct colonies, while MPNST4970 cells formed more diffuse colonies. Further analysis of tumorigenic potential was performed by measuring the migration of the established cell lines in a modified Boyden chamber assay. All the UPS and MPNST cell lines exhibited the ability to migrate toward a chemoattractant (Fig. 2c, lower panels). These results showed that the newly established cell lines could form colonies and migrate.

DNA mutation analysis of selected cancer genes

The mutation status of genes frequently mutated in UPS according to the TCGA, including ATRX, CDKN2A, RB1, and TP53, was assessed using high-coverage targeted sequencing with a minimum median target coverage of ~200×10,62,67,68. Figure 2d summarizes the mutations found in the cell lines. UPS271.1, UPS511, UPS0103 cell lines had no mutations in the cancer genes listed above. The USC060.1 cell line harbored a mutation in RB1: c.C1574G; p.A525G. This mutation has been reported in hemangioblastomas and acute myeloid leukemia69,70. The RIS620 cell line carried a TP53 mutation at c.G818A; p.R273H, that was documented in lung cancer, esophageal cancer, pancreatic carcinoma, hematological malignancies, ovarian and endometrial cancers, and gastric cancer33,71,72,73,74,75,76. Mutation status for EED, EZH2, NF1, and SUZ12 in the MPNST cell lines was reported by Kochat et al.68.

In vivo tumorigenicity studies

The in vivo tumor-forming capacity of the established cell lines was determined by either subcutaneous injection or tumor implantation. For USC060.1, and UPS511 cell lines, 3–5 million cells were injected subcutaneously into the flank of NSG mice, and tumor growth was measured weekly (Fig. 3a, b). For the implantation experiments, UPS0103, RIS620, MPNST007, and MPNST3813E, tumor pieces of ~3–5-mm3 were embedded into the flank of NSG mice, which were then monitored for tumor growth (Fig. 3c–f). Mice were sacrificed when the tumors reached a volume of 1000 to 2000 mm3. Hematoxylin and eosin staining of the xenograft tumors was performed on slides cut from formalin-fixed, paraffin-embedded blocks (Fig. 3a–f, right panels). Cells from the xenografts had similar morphology as cells from the parental tumors (H&E insets). The percentage of uptake of tumor cells and latency period are shown in Table 3. Overall, the cell line injection and implant xenograft models had uptake ranging from 70 to 100%. The latency period was the longest for UPS511 (50 days) cells and the shortest for RIS620 cells (23 days) (Table 3). The UPS271.1 and MPNST4970 cell lines show a heterologous growth pattern for the generation of xenografts that is currently under investigation.

Fig. 3: In vivo xenograft models of UPS and MPNST.
figure 3

a–d Tumor growth after UPS cell line injection or implantation in five representative animals (left panels). Images of hematoxylin and eosin (H&E)-stained sections of xenograft tumors at ×20 magnification (right panels). Scale bar indicates 100 µm. Inset shows original patient tumor image at ×20 magnification. e, f Tumor growth after MPNST cell line implantation in five representative animals (left panels). Images of H&E-stained sections of xenograft tumors at ×20 magnification (right panels). Scale bar indicates 100 µm. Inset shows original patient tumor image at ×20 magnification.

Full size image
Table 3 Latency and uptake for UPS and MPNST xenograft models.
Full size table

Discussion

Primary tumor cell lines established from patient tumors can recapitulate the original tumor for further study in a preclinical setting. The goal of the present study was to generate and characterize cellular and xenograft models for the study of 2 genomically complex sarcomas, UPS and MPNST. Including models like these in the latest systematic approaches58 or other large-scale methods will reveal previously unknown targetable vulnerabilities in sarcomas. We successfully established 4 UPS cell lines and one USC cell line. UPS was previously classified as malignant fibrous histiocytoma (MFH); the disadvantage of using any cell lines generated before the renaming of MFH as UPS is that they could be other sarcoma subtypes1,77,78. UPS cell lines with pathological confirmation are needed to overcome this complication. The UPS cell lines described in this study were derived from patient tumors classified as UPS or USC by a sarcoma pathologist at MD Anderson Cancer Center.

The establishment of new cancer cell line models presents some difficulties; most primary cultures fail to grow due to senescence after several months in culture. A recent study suggested the success rate is 15% in establishing sarcoma cell lines from resected tumors46. The soft tissue sarcoma cell lines described here have been in culture for ~1–2 years at this writing, showing no indication of senescence. Cell line contamination can be an issue for slow-growing cell lines (doubling times over 24 h) like those presented here. We confirmed no contamination in the cell lines based on the matching STR profiles for each cell line and the corresponding parental tumor. For the cell lines cultured from the initial patient-derived xenograft (USC060.1 and UPS271.1), targeted sequencing of the tumor samples showed over 95% of reads mapped to the human reference genome, ruling out the possibility of potential mouse cell contamination. Moreover, six of the eight established cell lines successfully formed xenograft tumors when administered to immunocompromised mice.

Targeted sequencing was performed on selected known cancer-related genes to identify genetic alterations in the established UPS and MPNST cell lines and provide an initial genomic characterization of the cell lines. Our finding that the genomes of the UPS and MPNST cell lines were not highly mutated agreed with the recent TCGA study of sarcomas10. In that study, the authors reported copy number losses in ATRX, CDKN2A, NF1, RB1, and TP53. PRC2 mutations, specifically inactivation of EED and SUZ12, have been documented in up to 90% of MPNSTs31,68,79,80. Of the 3 NF1-associated MPNST cell lines characterized in this study, MPNST642 carried a frameshift mutation (p.D675fs) in SUZ1268. Previous studies showed EZH2 is significantly upregulated in MPNST specimens compared to normal nerve cells, but not mutated68,81,82. The absence of EZH2 inactivating mutations in the NF1-MPNST cell lines described here (MPNST007 has a heterozygous gain of EZH2) supports the notion that functional EZH2 is required for uncontrolled tumor growth in MPNSTs68.

One drawback of the currently available large-scale analyses for these sarcoma subtypes is the relatively small number of samples in each study, which is expected due to their rarity (UPS, n = 44; MPNST, n = 5 in TCGA10). Creating more patient-derived UPS and MPNST resources gives researchers and clinicians much-needed tools to overcome the limited case numbers and available tumor materials. Patient-derived models that represent molecular subtypes of MPNST have multiple research and clinical applications in drug sensitivity studies, development of drug-resistant models, genetic/epigenetic manipulation, and large-scale—omics-based approaches. The UPS and MPNST PDX models established here can be used to interrogate tumor microenvironment, drug response, and mechanisms of drug resistance in vivo. The MPNST cell lines represent two subtypes of NF1-MPNST: those with intact PRC2 (MPNST007 and MPNST4970) and those lacking PRC2 activity (MPNST3813E). While an estimated 70–90% of MPNSTs feature loss of H3K27me3, the remaining 10–30% of MPNSTs retain functional PRC231,79. These subtypes can be leveraged in preclinical studies to elucidate the contribution of inactive PRC2 to MPNST progression, which remains unclear. From a personalized medicine standpoint, drug sensitivity data sets that include multiple MPNST subtypes (sporadic, NF1-associated, PRC2-wild type, PRC2 mutant, and potentially others) could help predict which treatments will work best for individuals based on the genetic makeup of their tumors. Similarly, insights into UPS’s genetic or epigenetic drivers and drug sensitivity could come from applying large-scale comprehensive -omics approaches to an extended pool of UPS models that represent the heterogeneous nature of this tumor type.

In summary, we successfully established several new cellular models for the study of UPS and MPNST. Furthermore, the initial characterization of the cell lines showed that they could be used as in vitro and in vivo models for further study of sarcoma etiology. Analysis of mutations in genes implicated in UPS and MPNST progression showed that the cell lines recapitulate known genetic alterations in UPS and MPNST. We hope that establishing these cellular models will allow researchers to take the next steps in investigating these debilitating soft tissue sarcomas in greater detail.