Establishment of a New Orthotopic Perirenal-Space-Grafted Mouse Model of Retroperitoneal Sarcoma

Background: Retroperitoneal sarcoma is a group of tumour originating from mesenchymal tissue in the retroperitoneal space. Although its incidence is lower than other tumours, retroperitoneal sarcoma has attracted increasing attention due to their high degree of malignancy, high recurrence rate and younger tendency. At present, the available treatments for retroperitoneal sarcoma are limited due to a lack of basic research and appropriate animal models. Results: In this study, two human sarcoma cell lines (dedifferentiated liposarcoma cell line SW872 and brosarcoma cell line HT1080) were selected to establish an orthotopic xenotransplantation retroperitoneal sarcoma mouse model. To achieve this, sarcoma cells labelled with GFP-Luc-Puro tag were inoculated into the perirenal space of nude mice. We evaluated the model by live imaging, histopathology and transcriptome analysis. Doxorubicin and Cisplatin were used to test the response of this model to chemotherapy drugs. The abdominal bulge of mice was obvious after 15 days for HT1080 and 30 days for SW872. Obvious chemical signals could be detected by an imaging system (IVIS Lumina (cid:0) ), and a solid tumour was visible in the abdominal cavity of live mice by MRI in vivo. Solid tumours inoculated in the perirenal space were signicantly larger than subcutaneous tumours. Moreover, solid tumours treated with doxorubicin and cisplatin were signicantly smaller than those treated with PBS. Through transcriptome analysis, we found serval signal pathway associated with tumorigenesis including TNF and NF-κB signaling pathways. Notably, a signicant increase in copy number of MDM2 and a signicant reduction in TP53, consistent with a shift in the MDM2/TP53 axis, with the most decisive genomic event contributing to liposarcoma development. Conclusions: In summary, we successfully established an orthotopic xenotransplantation model of retroperitoneal sarcoma, which can be used to study the initiation and development of retroperitoneal sarcoma, drug screening, diagnostic methods and new surgical procedures.


Introduction
Soft tissue sarcomas (STS) are a heterogeneous group of rare malignancies with mesenchymal origin that arise in any anatomic site. They account for almost 1.5% of all human malignant tumours, with an increased incidence in younger adults [1,2]. The retroperitoneum is the primary site of 15-20% of STS [3,4]. Moreover, the retroperitoneal location is often associated with a worse prognosis than STS in other locations. Among the retroperitoneal sarcomas (RPS), malignant retroperitoneal sarcoma (MRPS) accounts for about 80% of cases. However, all sarcoma subtypes can arise at this location. The most common histotypes of RPS are well-differentiated/dedifferentiated liposarcoma, solitary brous tumours, leiomyosarcoma, undifferentiated pleomorphic sarcoma and malignant peripheral nerve sheath tumours. RPS originating in this anatomic location are usually hidden deep in the retroperitoneum. They tend to reach large sizes and often encase or in ltrate multiple adjacent organs before they become symptomatic [5]. It is di cult to complete radical resection for these tumours, and the recurrence rate is high [6]. Because MRPS are not sensitive to radiotherapy and chemotherapy [7],the treatment and prognosis of MRPS represents a common global challenge in the eld of oncology.
There are limited advancements in diagnosis and treatment options available to RPS patients compared to other cancers due to a lack of basic and clinical research, especially appropriate cell and animal models [8]. Current retroperitoneal sarcoma models include cell-derived xenograft (CDX), patient-derived xenograft (PDX) and transgenic animal models. SICD mice was used for patient-derived subcutaneous xenograft [9][10][11][12][13][14][15]. Athymic nude NMRI mice was used for subcutaneous xenograft [16]. Assi et al. [17] employed intramuscular injection of human SW872 cells. Jason M. Warram et al. [18] implanted HT1080 into the posterior of female nude athymic mice. Daniel Johannes Tilkorna et al. [19]  These previously reported PDX, CDX and transgenic animal models have played an important role in the study of retroperitoneal sarcomas, but they are limited in terms of sample acquisition, animal raising, short-term e ciency and the complexity of genetic mutations. The most common model, subcutaneous xenograft, cannot re ect the occurrence and development of tumours to the same extent as orthotopic xenografts. Orthotopic xenograft models provide a more appropriate microenvironment compared to ectopic xenograft models. A simple animal model that closely mimics the real situation with good shortterm e ciency is necessary for molecular biology, imaging, drug screening and testing [19].
Here, we performed orthotopic transplantation of the brosarcoma cell line HT1080 and liposarcoma cell line SW872 into the retroperitoneal perirenal space to establish a mouse model of retroperitoneal sarcoma with new in situ growth that is easily accessed and presents faster growth and high verisimilitude. This orthotopic xenograft model will serve as a powerful in vivo tool to further understand sarcomagenesis and improve drug screening, diagnostic procedures and the development of therapeutic modalities.

Inoculation of brosarcoma HT1080 and liposarcoma SW872 cells into the perirenal space
The anatomic location of the retroperitoneum is the space between the retroperitoneal peritoneum and the abdominal transverse fascia under the diaphragm but above the pelvic diaphragm. Retroperitoneal tumours usually originate in the latent space of the retroperitoneum, including primary retroperitoneal tumours and metastatic tumours from fat, loose connective tissue, muscle, fascia, blood vessels, nerves and lymphoid tissue in the retroperitoneal space. As shown in Fig. 1A,the retroperitoneal space is marked by the dotted line. We chose the perirenal space for transplantation of HT1080 and SW872 with GFP-Luc-Puro tagged cells, as shown in the position marked by an asterisk (Fig. 1A, B). We injected trypan blue into the perirenal space to highlight the location and post injection effects (Fig. 1B).
HT1080 and SW872 cells exhibit fast growth and tumorigenesis in the retroperitoneal perirenal space.
The tumorigenesis of HT1080 and SW872 cells could be observed after inoculation of the perirenal space with GFP-Luc-Puro tagged HT1080 and SW872 cells (Fig. 2). The luminescence signal was detected and it gradually enhanced with time ( Fig. 2A), suggesting that the tumours are established and grow. We also con rmed the tumorigenesis through MRI: a solid tumour with a clear boundary was formed at the site of the asterisk (Fig. 2B). HT1080 or SW872 cells could form obvious solid tumours in the perirenal space within 15 and 30 days, respectively. These mice were dissected after decapitation, and we observed that both brosarcoma HT1080 and liposarcoma SW872 cells could form visible solid tumours with a little adhesion to the surrounding organs (Fig. 2C).
Retroperitoneal sarcoma cells inoculated in the perirenal space have stronger tumorigenic ability than inoculated subcutaneously.
Previous studies have reported that it is di cult for retroperitoneal sarcoma cells to form subcutaneous xenograft tumours in nude mice. To examine the differences in tumorigenicity between perirenal space inoculation and subcutaneous inoculation, we injected same amount of HT1080 or SW872 cells in perirenal space and subcutaneous space of nude mice to establish xenograft tumours, respectively. Both HT1080 and SW872 cells inoculated in the perirenal space had stronger tumorigenic ability compared to cells inoculated subcutaneously (p < 0.001) (Fig. 3A, B). MDM2 overexpression and its inhibition of p53 activity in the sarcomatous tumour lead to overexpansion of the cells, which may be one of the pathogenic mechanisms underlying RLPS [12,24]. As a common tumour marker, Ki67 is commonly used as a marker of malignant proliferation. In the current study, extensive Ki67-and MDM2-positive cells were observed in these tumor tissues, suggest that the tumors in current models had basic characteristics of retroperitoneal tumours (Fig. 3C). Meanwhile, Ki67 staining was performed to examine tumor cell proliferation between perirenal space and subcutaneous grafted tumours. Since subcutaneous grafted HT1080 cells formed hydrous pus like masses and could not be dissected, further IHC staining was conducted only on SW872 tumour masses. The results showed that the positive rate of Ki67 staining had no signi cant difference between perirenal-space-grafted and subcutaneously-grafted tumours (Fig. 3D); however, the TUNEL-positive cells in perirenal-space-grafted tumours were signi cantly less than those in subcutaneously-grafted-tumours (Fig. 3E). These results suggest that perirenal space is more suitable for the survival but not proliferation of retroperitoneal sarcomas as compared to subcutaneous space.
Compare gene expression between perirenal-space-grafted and subcutaneously-grafted tumours To further understand why the tumorigenic ability of cells inoculated in the perirenal space was stronger than that in cells inoculated subcutanueously, three solid SW872 tumours formed in the perirenal space and three solid SW872 tumours formed in subcutaneous space were selected to construct transcriptome sequencing libraries. Whole transcriptome pro ling was performed (Fig. 4). The transcriptome data indicated that a total of 18,983 genes were detected in these tumours (see Supplementary table 1). The volcanic map shows gene distribution and gene expression differences (Fig. 4A). 11,061 genes were found to be differentially expressed, of which 5876 genes were upregulated and 5185 were downregulated. We selected 421 genes related to tumour formation or tumorigenesis. There were a total of 217 genes whose variation range reached 1.3 times (Fig. 4B; Supplementary table 2). KEGG pathway classi cation and enrichment analysis for DEGs were also carried out. A total of 216 genes were allocated to six main KEGG metabolic pathways (Fig. 4C): 22 genes were classi ed as cellular processes, 30 genes as environment information, 8 genes as genetic information processing, 83 genes as human disease, 32 genes as metabolism, and 55 genes as organismal system pathways. The main secondary pathways were cell growth and death (9), signal transduction (26), folding, sorting and degradation (4), infection diseases-viral(9), global and overviw maps(6), and immune system (20). The main altered pathways are TNF signaling pathway, NF-kappa B signaling pathway, Proteasome, NOD-like receptor signaling pathway, IL-17 signaling pathway (Fig. 4D). These changed molecules are mainly involved in regulating cell growth, death and the immune system. Compared with subcutaneously-grafted tumours, the TNF signaling pathway is inhibited in perirenal-space-grafed tumours, which may result in decreased apoptosis. In the NF-kappa B signaling pathway, uPA and P65 are enhanced in perirenal-space-grafted tumours, which may promote cell survival. Genes that signi cantly up-regulate more than 4 times include PRF1, LAMC2, MAGEA1, AIM2, EMP2, PTPN22, CHRFAM7A, KLF2, KRT15, KRT8, EDA2R, SPHK1, F2RL1, CLU, PTGER4, PLAU, LAMB3, SMAGP, Genes down-regulated by more than 4 times include CXCL1, MMP1, TNFSF11, G0S2, THY1, HSPB1, MBP, IL6, TNFRSF11B, ICAM1, SELENBP1, PYCARD, CD40, UBD1, IRAK3, CLIP3, ADAMTS12, CYBA, VCAM1, etc., see Supplementary table 2 Notably, when perirenal-space-grafted tumours were compared to subcutaneously-grafted tumours, MDM2 expression was upregulated 3.14 times (p < 0.001), while TP53 expression was downregulated 1.5 times (0.01 < p < 0.05; see Supplementary table 1). MDM2/TP53 was identi ed as the most decisive pathway associated with liposarcoma development in clinical samples.
Use perirenal-space-grafted tumorigenesis model to evaluate drug therapies for retroperitoneal sarcoma.
Although surgery represents the mainstay for the treatment of localized RPS, doxorubicin is used as the rst-line drug in soft tissue sarcomas (STS). Cisplatin is a commonly used chemotherapy drug. We evaluated the response of ps-CDX models to doxorubicin and cisplatin. The therapeutic effect was assessed by examining changes in bioluminescence imaging signaling intensity, tumour weight, and body weight of tumour-bearing mice. We injected 100− µl DMEM containing 1 × 10 7 HT1080 brosarcoma cells into the perirenal space of each nude mouse, which were divided into three groups: PBS, cisplatin (5 mg/kg), and doxorubicin (4 mg/kg). Mice received intraperitoneal injection of their respective treatment starting from the third day after the initial cell injection, then received subsequent treatments every 3 days. In addition, on day 3, 6, and 15, luciferin signal acquisition was performed using the IVIS Lumina II system. The uorescence signal intensity of the PBS group gradually increased with time, and the weight of solid tumours was signi cantly higher than those of mice in the cisplatin (p < 0.01) and doxorubicin (p < 0.01) groups (Fig. 5A, B). It is noted that long-term use of cisplatin and doxorubicin resulted in a signi cant decrease in body weight. (Fig. 5C).

Discussion
Tumours located in the retroperitoneal region often associate with a worse prognosis compared with the tumors located in trunk and limbs. In fact, in this anatomic location, tumours have a tendency to grow to a larger size, associated with a comparatively lower chance of surgical excision. The purpose of this study was to establish a xenograft tumours at the site of origin of retroperitoneal tumours in nude mice to provide a mouse model to study retroperitoneal sarcoma tumorigenesis.
In this study, we successfully established a retroperitoneal orthotopic xenograft model by inoculating liposarcoma and brosarcoma cells into primary origin sites of retroperitoneal sarcoma in nude mice. Furthermore, we found that retroperitoneal sarcoma cells inoculated in the perirenal space have stronger tumorigenic ability than inoculated subcutaneously, which is due to better survival in the perirenal space. Through RNA sequencing, we identi ed a number of signaling pathways that promote the formation of retroperitoneal sarcoma in perirenal-space-grafted tumours altered compared to those in subcutaneouslygrafted tumours. Among them, TNF/NF-kappa B and MDM2/TP53 signaling pathways attracted our attention. TNF, as a critical cytokine, can induce a wide range of intracellular signal pathways including apoptosis and cell survival as well as in ammation and immunity [25]. TNFR1 induce activation of many genes, primarily related to two distinct pathways: survival pathway including NF-kappa B and MAPK cascade, and death pathway including apoptosis and necroptosis. TNFR2 signaling activates NF-kappa B pathway including PI3K-dependent NF-kappa B pathway and JNK pathway leading to survival [26]. NFkappa B is the generic name of a family of transcription factors that function as dimers and regulate genes involved in immunity, in ammation and cell survival [27]. Notably, the MDM2/TP53 axis, which plays a reference role in the study of the mechanism underlying the formation of human retroperitoneal sarcoma. Moreover, in order to verify the practicability of this model, we tested the effects of doxorubicin and cisplatin treatment on the retroperitoneal orthotopic xenograft, verifying that the animal model can be used for evaluating pharmacological effect.
In clinical practice, retroperitoneal tumours are usually large and can crush or invade adjacent organs and blood vessels, such as the kidneys, adrenal gland, spleen and colon [28,29]. Complete surgical resection is currently recognised as the most effective treatment, but the complete resection rate of primary tumours is still low and the recurrence rate remains high [6]. The primary site of STS has an important impact on the prognosis of patients, as the prognosis of RPS is much worse than tumours located in the limbs or trunk. At present, the scope of resection of RPS remains uncertain. The current debate is focussed on the need to dissect adjacent tissues that cannot be observed by the naked eyes [30]. Although the recurrence rate can be greatly reduced by large-scale whole or extended resection of uninvolved organs, the removal of adjacent organs is also a challenge, associated with increased complications after organ resection. The predictive ability of surgical resection margin evaluation on post operative complications still needs to be explored [31].
There is no evidence to show that retroperitoneal tumours are sensitive to radiotherapy, chemotherapy or immunotherapy; therefore, postoperative adjuvant therapy is not commonly used [32]. Due to a lack of effective drugs, systemic therapy has failed to play an important role in the treatment of retroperitoneal liposarcoma. In recent years, due to continuing basic and clinical research, some promising clinical exploration prospects have emerged. Molecular targeting and cell therapy have gained attention and become the focus of investigations [33]. The mechanism of retroperitoneal development, accurate preoperative diagnosis, precise intraoperative surgery and precise non-surgical targeted therapy are also under investigation. All the research requires an animal model that closely resembles the real retroperitoneal tumour for preliminary exploration and advancements in technology.
In this study, we report the establishment of an orthotopic xenotransplantation model of retroperitoneal sarcoma. This model has several characteristics. Firstly, it is inoculated in situ, which mimics the environment of retroperitoneal sarcoma occurrence. The model also allows the visualisation of tumour cells due to GFP-Luc labelling, which can re ect the occurrence and development of the tumour more accurately and comprehensively, and can also more accurately re ect the therapeutic effect of drugs.
From the anatomical map, it can be seen that the model is closely related to clinical retroperitoneal tumours. Therefore, it can be used to study new surgical procedures and outcomes. It is worth noting that retroperitoneal sarcoma is not sensitive to any of the current chemotherapeutic drugs. This model can be used to verify the effects of newer treatments including cell therapy and immunotherapy, which have emerged in recent years. Secondly, the observed enhancement of the MDM2/TP53 signalling pathway is similar to that observed in the clinic, so this model can be used for the development and exploration of new diagnostic markers and methods. Finally, the mice used in this model are ordinary nude mice, and the cells are commercial cell lines, which provides great advantages in terms of operation, feeding conditions and price compared to severely defective nude mice like SCID mice. Unlike the previous PDX model, tissues do not need to be obtained from patients, which requires extended waiting time for appropriate surgery and strict ethical review. The commercial cell lines used in the current model are easier to obtain.

Conclusion
The present study has demonstrated that sarcoma cells, including brosarcoma and liposarcoma cells, can be successfully grafted into the retroperitoneal perirenal space of nude mice. The consistently high engraftment rate and major preservation of immunophenotypes observed indicates that such xenografts can be used as tools in studies aimed at improving our understanding of retroperitoneal sarcoma tumorigenesis. Potential applications of the xenografts include preclinical testing of responses to various chemotherapeutic regimens, evaluation of novel therapeutic agents and diagnostic methods, analysis of tumour progression at the cellular and molecular levels, and identi cation of new therapeutic targets. In vivo animal studies

Methods
The RPS orthotopic xenograft mouse model was established by injecting 0.1 ml of tumour cell suspension (DMEM) containing1 × 10 7 cells for SW872 and 1 × 10 5 cells for HT1080 into the perirenal space and subcutaneous space. When the tumour volume reached a palpable size, the mouse was used for further studies.

Immunohistochemistry staining
Tumour sections were cut at a thickness of 5 μm using a microtome and mounted on glass microscope slides. For immunohistochemical staining, sections were dewaxed in xylene and hydrated in graded alcohol solutions and distilled water. Routine haematoxylin and eosin (H&E) staining was carried out. Immunohistochemical staining was performed using antibodies against the proto-oncogene MDM2 (bs23748R, Bioss, China) and proliferation marker Ki67 (15580, Abcam) according to the manufacturers' recommendations.

Bioluminescence imaging
Nude mice bearing SW872 and HT1080 with GFP-Luc-Puro tagged cell orthotopic xenografts were injected with 100 μl of luciferin solution intraperitoneally. The mice were left for 10 min and then anesthetised before imaging. Bioluminescence of live mice was monitored using an in vivo imaging system (IVIS Lumina ; Caliper life Science, Hopkinton, MA, USA). The luminescence intensity was quanti ed as photon counts per second using Living Image software.

MRI imaging
Nude mice bearing SW872 and HT1080 with GFP-Luc-Puro tagged cell xenografts were evaluated by in vivo MRI, as described previously [23].

TUNEL Staining
Brie y, slides (4 mm) were stained with TdT-mediated dUTP nick end labeling (TUNEL) probes (Roche, 11684817910, United States) according to the product instructions. DAPI was used to stain the nuclei.
Images were captured using a uorescence microscope.

RNA sequencing
Transcriptome sequencing was conducted by Wuhan BGI Technology Co., Ltd. Brie y, total RNA was extracted from xenografts in the perirenal space and subcutaneous locations, then the Illumina HiSeq4000 platform RSEM was used for RNA quanti cation. Based on raw count data, differential expression analysis between samples was performed using the PossionDis algorithm.

Statistical Analysis
All data were collected from more than three independent experiments, expressed as the mean ± standard error of the mean. All statistical signi cance was determined by unpaired two-sided t-tests using GraphPad Prism 5. For all gures, differences were considered signi cant at *p < 0.05, **p < 0.01 and ***p

Declarations
Ethics approval and consent to participate All mice were raised in the Laboratory Animal Center of Xiamen University. This research study was approved by the Ethics Committee of the Xiang'an Hospital of Xiamen University. All animal experiments were performed in strict accordance with the relevant laws and regulations approved by the Institutional Animal Care and Use Committee of Xiamen University. All efforts are to minimize the number of animals, minimize the manipulation, and minimize the pain of the animals.

Consent for publication
Not applicable.

Availability of data and materials
All data generated or analyzed during this study are included in this article.

Competing interests
The author declare that he have no competing interests.