Devices & Hardware

A fluorescent orthotopic model of metastatic cervical carcinoma

Description
Currently, there is no mouse model of cervical cancer that allows for the study of the later stages of the disease, including metastasis. We report here the development of an orthotopic model of human cervical carcinoma in which tumor fragments are
Published
of 7
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Related Documents
Share
Transcript
  Clinical & Experimental Metastasis  21:  275–281, 2004.© 2004  Kluwer Academic Publishers. Printed in the Netherlands.  275 A fluorescent orthotopic model of metastatic cervical carcinoma Rob A. Cairns 1 , 2 & Richard P. Hill 1 , 2 , 3 1  Experimental Therapeutics Division, Ontario Cancer Institute/Princess Margaret Hospital, Toronto, Ontario, Canada; Departments of   2  Medical Biophysics and   3  Radiation Oncology, University of Toronto, Toronto, Ontario, Canada Received 8 November 2003; accepted in revised form 6 May 2004 Key words: cervicalcarcinoma,DsRed, enhancedgreenfluorescentprotein, fluorescentmicroscopy,lymph-nodemetastasis,orthotopic model Abstract Currently, there is no mouse model of cervical cancer that allows for the study of the later stages of the disease, includingmetastasis. We report here the development of an orthotopic model of human cervical carcinoma in which tumor frag-ments are surgically implanted into the cervix of SCID mice. The human cervical carcinoma cell lines used in this study(CaSki, ME-180, and SiHa) have been engineered to stably express the fluorescent proteins enhanced green fluorescentprotein (EGFP) or DsRed2, allowing for  in vivo  optical monitoring of tumor growth and metastasis. The cervical implantsdevelopinto large intraperitonealmasses involvingthe entire reproductivetract, with little local invasionof other abdominalstructures. These tumors metastasize initially to local lymph nodes and later to lung, a pattern consistent with the clinicalcourse of the disease. It was found that the use of the DsRed2 protein as a fluorescent marker has distinct advantages overEGFP due to the wavelength of its emission spectrum (575–625 nm vs 500–550 nm). Tissue penetration of light at thiswavelength is greater, and the auto-fluorescence of mouse tissues is less intense, resulting in an enhanced signal to noiseratio compared to results obtained with EGFP. This model should allow for a more relevant investigation of the factors thataffect the metastasis of cervical carcinoma and presents an opportunity to evaluate potential therapeutic strategies designedto prevent the spread of disease.  Abbreviations:  SCID – severe combined immunodeficient; EGFP – enhanced green fluorescent protein; VEGF – vascularendothelial growth factor; PD-ECGF – platelet-derived endothelial cell growth factor; MMP – matrix metalloproteinase;HPV – human papilloma virus; ATCC – American type culture collection; α -MEM – alpha minimal essential medium; FBS– fetal bovine serum; CMV – cytomegalovirus;FACS – fluorescence activated cell sorting Introduction In the United Sates, it is estimated that there will be 12,200new cases of cervical carcinoma diagnosed in the year 2003,and that there will be 4,100 deaths due to the disease [1].In the developing world, where screening programs are farless widespread, cervical carcinoma remains the third mostcommon cancer among women with approximately 370,000new cases arising each year [2].The spread of cervical carcinoma to the pelvic and aor-tic lymph nodes is a common occurrence, and is one of theprimary determinants of outcome for patients [3, 4]. There- fore, understandingthe process of lymph-node metastasis inthis disease is essential forimprovingprognosticcapabilitiesand for designing rational therapeutic strategies. In recentyears, clinical studies have identified a variety of molecularand biochemical factors as being potentially associated withmetastasis and/or poor outcome. These include angiogenic Correspondence to : Dr R.P. Hill, Princess Margaret Hospital, 610 Univer-sity Avenue, Rm 10-113, Toronto, Ontario, Canada M5G 2M9. Tel:  + 1-416-946-2979; Fax: + 1-416-946-2984; E-mail: hill@uhnres.utoronto.ca factors such as VEGF, VEGF-C, and PD-ECGF, the mat-rix metalloprotease MMP-2, the cell surface glycoproteinCD44v6, and the growth factor receptor c-erb-B2 [5–10]. In addition, the level of tumor hypoxia has been shown tohave a negative impact on outcome, where more hypoxictumors are more likely to metastasize and have a poorerpro-gnosis [11–13]. These studies raise a number of interesting questions that require an appropriate experimental systemwith which to conduct laboratory experiments capable of determining which factors play important mechanistic roles.Unfortunately, there is no mouse model of cervical car-cinoma that allows for the study of lymph node or distantmetastasis. The transgenic K-14 HPV mouse, in which cer-vical carcinoma develops upon exposure to 17 β -estradiol,has been useful in examining the early stages of epithelialtransformation, however,spread of disease in this model hasnot been reported [14].An alternativetoa transgenicmouseis axenograftmodelin which human tumor cells are implanted into immune de-prived mice. There is a large and growing body of work supporting the use of orthotopic implantation where, in  276  R.A. Cairns & R.P. Hill general, the natural history of the disease resembles moreclosely that seen in the clinic [reviewed in 15, 16]. This isespecially true for studies of metastasis, in that orthotopic-ally transplanted tumors tend to be more aggressive and todemonstrate patterns of spread similar to those observed inpatients.Another advantage of transplant models is that the tu-mor cells can be genetically engineered  in vitro  to expressreporter genes that allow for their detection after implanta-tion. Recently, the use of the fluorescent protein EGFP hasbeen used to generate orthotopic models in which tumorgrowth and metastasis can be monitored optically in realtime [17–19]. By combining external imaging techniqueswith intravital microscopy, this strategy allows for a quant-itative analysis of the full course of disease from a few cellsthrough to large tumors. Therefore, even the early stages of metastasis are accessible for study. Furthermore,throughtheuse of multiple fluorescent proteins, specific populations of cells can be tracked in individual animals, which may helpto reduce the inter-animal variation often encountered whenstudying metastasis.We report here the development of an orthotopic xeno-graft model of human cervical carcinoma using cell linesengineered to express the fluorescent marker proteins EGFPand DsRed2. We also show that these tumors are metastaticand spread initially to the local pelvic lymph nodes andsubsequently to other distant sites. Materials and methods  Mice and tumor cell lines CaSki, ME-180, and SiHa human cervical carcinoma celllines were obtained from the ATCC. Cells were maintainedby alternate  in vitro  and  in vivo  passage in order to minim-ize the selection of cells adapted to grow in either situation.  In vitro , cells were grown as monolayers in plastic flasksusing  α -minimal essential medium ( α -MEM) (Gibco BRL,Burlington, Canada) supplemented with 10% fetal bovineserum (FBS, Wisent, Canada) at 37  ◦ C in 5% CO 2 . For in vivo  growth,cells between the 2–4th in vitro  passage wereremoved from the flasks during exponential growth using0.05%trypsin for 10 min at 37 ◦ C and transplantedinto 8- to12-week-old female SCID mice obtained from an in-housebreeding program. Each tumor was initiated by injecting2 . 5 × 10 5 cells in 50  µ l of media into the left gastrocnemiusmuscle. Tumor growth was monitored by an external meas-urementoflegdiameter. Animalswerehousedat theOntarioCancer Institute animal colony and had access to food andwater  ad libitum . All experimentswere performedaccordingto the regulations of the Canadian Council on Animal Care. Construction and analysis of fluorescent cell lines CaSki, ME-180, and SiHa human cervical carcinoma cellswere transfected, using Clonfectin (Clontech, Palo Alto,California), with plasmids containing either the EGFP orDsRed2 gene under the control of the CMV promoter, andthe neomycin resistance gene under the control of the SV40early promoter(Clontech, Palo Alto, California). After 48 h,cells were exposed to 800  µ g/ml G418 in order to selectfor stable transfectants. After growth under selection for 2weeks, cells were sorted by fluorescence activated cell sort-ing (FACS) in order to isolate the brightest 5 to 10% of thepopulation. These cells were then cloned and the brightestclones were selected and analysed by flow cytometry. Thestability of the fluorescent marker was assessed by platingcells at colony forming density on 10-cm dishes in triplicateand determining the number of fluorescent colonies arisingafter a period of 10 to 14 days. This assay was performedusing cells growing  in vitro  as well as cells isolated frommetastasesgrowinginthemouselung.Individuallungswereenzymatically and mechanically digested to obtain a singlecell suspension as described previously [20]. Orthotopic implantation Intramuscular (i.m.) tumors 0.6–0.8 g in size were excisedand dissected under sterile conditions. The tumors were cutinto2–3mm 3 fragmentsin α -MEMmediaandplacedonice.Female SCID mice were anaesthetized by isofluorane inhal-ation and the uterus was exposed by an abdominal midlineincision. A small incision was made in the uterus at the levelof the cervix and a 2–3 mm 3 tumor fragment was suturedin place using a single 8-0 silk suture. The abdomen wasclosed in two layers using 4-0 silk sutures and stainless steelwound clips. Tumors were imaged/removedas a function of time, weighed, and fixed in formalin for histological exam-ination. Mice were further dissected in order to locate andimage metastases.  Imaging fluorescent tumor cells A Leica MZ FLIII fluorescent stereomicroscope with a 100-W mercury lamp was used to observe fluorescent tumorsand metastases. Primary tumors were imaged externallyover time, after removing hair, and again after dissection.Metastases were imaged after dissecting the animals. EGFPand DsRed cells were observed together using a 480/40 ex-citation filter and a 510 long-pass emission filter. DsRedcells were observed alone using a 560/40 excitation filterand a 610 long-pass emission filter. Images were acquiredusing a Leica DC350 digital camera and analysed usingNorthern Eclipse software (Empix Imaging, Mississauga,Canada). The images presented here have been adjustedfor contrast and colour balance in order to increase easeof viewing. However, all quantification was performed onunmanipulated images. Results Construction and characterization of fluorescent cell lines The cervix cell lines CaSki, ME-180, and SiHa were en-gineered to stably, and constitutively express the fluorescentproteins DsRed and EGFP. Flow cytometry showed that the   An orthotopic model of cervical carcinoma  277cell lines could be distinguished from the untransfected par-ental lines and that the expression was maintained when thecells were grown as tumors  in vivo  (data not shown). Tofurther quantify the stability of the fluorescent phenotype,CaSkiandME-180cellswerecultured invitro  intheabsenceof selection for 20 days and then plated at colony-formingdensity, also in the absence of selection. While the majorityof the cells plated gave rise to fluorescent colonies, therewas some loss of fluorescent protein expression, especiallyin the CaSki EGFP cell line, where approximately 25% of the colonies were non-fluorescent (Table 1). To examine thestability of the fluorescent phenotype  in vivo , 5 × 10 5 CaSkicells were injected intravenously (i.v.) into SCID mice. Inthis situation, the cells arrest in the lung and develop intopulmonarymetastases. After14days, thelungswereexcisedand digested into single cell suspensions, which were thenplated at colony-forming density. In this situation a signi-ficant number of the tumor cell colonies recovered from thelungs were non-fluorescent indicating that there is a loss of the fluorescent phenotype in some tumor cells after growth in vivo . There was also considerable variation between micein the proportionoffluorescentcolonies recovered(Table 1).This may have been due in part to heterogeneity in the sizeof the lung nodules that contributed to the pool of cellsrecovered after lung digestion.In order to determine whether expression of the fluores-cent proteins, or the subcloning of the cell lines altered thegrowthcharacteristicsofthe newlyderivedcells,  in vitro  and in vivo  growth curves were determined for the EGFP andDsRed variants of the CaSki and ME-180 cell lines. Therewere no significant differences between the fluorescent vari-ants and the parental controls in terms of the  in vitro  growthrate (data not shown) or the growth rate of i.m. tumors. TheCaSki cell line demonstrated an  in vitro  doubling time of 12.5 h while the ME-180 cells showed a doubling time of 29.5h. Similarly, CaSki tumors grewmore rapidlythanME-180 tumors when injected i.m., where the doubling timeswere five days and nine days, respectively (Figure 1A). Optical imaging of tumor growth To assess our ability to detect tumorcells  in vivo  on the basisof their fluorescent signal, a group of CaSki DsRed i.m.tumors was imaged externally as a function of time, afterinjection of 2 . 5 × 10 5 cells. These images were segmentedbased on a fluorescent intensity threshold and two perpen- ◭ Figure 1.  Growth of fluorescent tumors. CaSki M1 and ME-180 M1 in-dicate the untransfected parental control lines. (A) growth rates of i.m.tumors initiated by injecting 2 . 5  ×  10 5 cells into the hindlimb of mice.Tumor weight was calculated by measuring leg diameter in a single groupof animals over time. Error bars represent 1 SD ( n  =  5–7 per group).(B) fluorescent quantification of CaSki DsRed i.m. tumors, imaged extern-ally with a fluorescence stereomicroscope using the DsRed filter set (8 × magnification). A single group of animals was imaged live and externallyover time ( n  =  5). Images were segmented based on a DsRed fluorescentintensity threshold and two measured diameters were used to calculate tu-mor volume based on an ellipsoid geometry. Mean volume is plotted as afunction of time after injection of 2 . 5 × 10 5 cells; error bars represent 1 SD.(C) growth rates of orthotopic cervical tumors. Primary tumor weight wasmeasured by excising and weighing primary tumors at various times afterimplantation. Means of groups of five to eight animals are plotted at eachtime point (error bars represent 1 SD). The volume of an independent setof 5 ME-180 DsRed cervical tumors was measured over time by externalfluorescent imaging of the abdomen using a fluorescent stereomicroscope(1.6 × magnification). DsRed fluorescent images were segmented based onan intensity threshold and two measured diameters were used to calculatethe volume based on an ellipsoid geometry (error bars represent 1 SD).  278  R.A. Cairns & R.P. Hill Table 1.  Stability of fluorescent protein expression  in vitro  and  in vivo . Stability of the fluorescentphenotype was assessed by plating cells at colony forming density and determining the number of fluorescent and non-fluorescent colonies that developed after 10to 14 days. Thenumber ofcoloniescounted per plate ranged from 20 to 250.  In vitro  stability 1  In vivo  stability 2 (fluorescent colonies ± SD) (fluorescent colonies)CaSki DsRed 99.3% ( ± 0.8) 63.1% (3.7, 57.7, 72.9, 89.0, 92.3)CaSki EGFP 74.1% ( ± 5.4) 64.7% (40.0, 58.1, 68.1, 70.1, 86.2)ME-180 DsRed 89.0% ( ± 5.7) NDME-180 EGFP 86.2% ( ± 2.6) ND 1 Cells were plated in triplicate after growth of the cells in culture for 20 days in the absence of selection (mean ± SD). 2 Cells were injected intravenously into SCID mice, and after 14 day, the lungs were removed,digested to single-cell suspensions and plated in triplicate. The mean and the values for the fivemice assessed are reported. diculardiametersweremeasuredandusedtocalculatetumorvolume assuming an ellipsoid geometry (Figure 1B). Tu-mors could be detected immediately after implantation andtheir growth could be observed over a period of 13 days.After the tumors reached a diameter of   > 7.5 mm, at ap-proximately 13 days, areas of necrosis and hemorrhagepreventedfurtheraccurate quantificationof tumorsize basedon the fluorescent signal. When tumor volumes were calcu-lated from the fluorescent images for days 7–13, they werenot significantly different from the tumor weights obtainedby measurements of leg diameter, assuming a density of 1 g/cm 3 (Figures 1A and B). Orthotopic growth and spontaneous metastasis During the developmentof the orthotopicimplantation tech-nique, both surgical implantation of tumor fragments, andinjection of tumor cell suspensions were employed in orderto initiate cervical tumors. However, it was found that theinjectiontechniquedidnot producereproducibleresults, andtherefore, the surgical implantation technique was adopted.After surgical orthotopic implantation of 2–3 mm 3 tumorfragments, primary masses developedat the cervix and grewto involve the entire reproductive tract (Figure 2). These tu-mors developed into large intraperitoneal masses with littleinvasionof other abdominalstructures such as the bladderorintestine.ME-180 and CaSki tumors were excised and weighed asa function of time in order to determine the growth rate of primary tumors (Figure 1C). As was observed at the i.m.site, CaSki tumors were rapidly growing, whereas ME-180tumors developed at a slower rate. While it is difficult toderive accurate  in vivo  doubling times from the availabledata, the orthotopic tumors and the i.m. tumors appeared togrow at similar rates. Also, the DsRed and EGFP variants of each cell line grew at similar rates (data not shown), as wasobserved in the i.m. tumors. The cervical tumors expressedthe fluorescent proteins, and could be imaged using a fluor-escent stereomicroscope from outside the animal. A set of 5ME-180DsRed tumorswas imagedasa functionoftime andthe fluorescent images were used to quantify tumor growth(Figure 1C). The tumor volumes that were calculated fromthese images were in the same range as the tumor massesobtained by excising and weighing primary tumors. Also,when cervical tumor weight was measured after dissectionat the end of the experiment, the mean value was 0.9  ± 0.2 g, which is slightly lower, but not significantly differentfrom the volume calculated from the optical signal at day42 (1 . 1 ± 0 . 2 g assuming a density of 1 g/cm 3 ). Quantitat-ive external imaging of the CaSki cervical tumors was notpossible due to artefacts created by large areas of hemor-rhagic necrosis that blocked the fluorescent signal, as wasthe case for the larger CaSki i.m. tumors described above.The histology of the primary tumors did not vary substan-tially between the i.m. site and the orthotopicsite (Figure 3).The CaSki tumors were very poorlydifferentiated, with sub-stantial areas of necrosis and hemorrhage. The SiHa tumorswere also poorly differentiated, however, little necrosis orhemorrhage was evident. Finally, the ME-180 tumors werewell-differentiated and showed little if any necrosis.Animals bearing tumors greater that 0.4 g were dissectedat a fluorescentstereomicroscopein orderto assess spreadof disease. Metastasis to the local lymph nodes and lungs wasevident for all tumor cell lines (Figure 2, Table 2). Duringdissection, small numbers of fluorescent tumor cells couldbe easily detected so that in some cases, what appeared tobe individual cells could be visualised colonizing both thelungs and the local nodes. CaSki tumors were extremelyaggressive, with all but one of the animals showing spreadto the lymph nodes and lungs. Several animals also hadmetastases at other sites, which included liver, kidney, andintestine. ME-180andSiHatumorsalsospreadreadilytothelocal lymph nodes but were less aggressive in their spreadto lung, with the proportion of animals showing pulmonarymetastasis being 30% and 20%, respectively. Interestingly,no animals in any group were found to have lung metastasesin the absence of lymph node involvement.Althoughmetastasis has onlybeenassessed at a few timepoints, in all tumor types, the larger tumors tended to ex-hibit more extensive metastatic spread, both in terms of thenumber of sites involved and the size of the individual le-sions (data not shown). To examine the relationship betweentumor size and metastatic spread in more detail, orthotopicCaSki tumors were initiated by injecting a cell suspension,   An orthotopic model of cervical carcinoma  279 Figure 2.  Orthotopic tumor growth and metastasis. (A, C, E, G, I, K) show brightfield images; (B, D, F, H, J, L) show corresponding fluorescence images.(A and B) show a ME-180 DsRed cervical tumor six weeks after implantation. (C and D) show a SiHa EGFP cervical tumor six weeks after implantation.(E and F) show local lymph-node metastases in a mouse bearing a ME-180 DsRed tumor six weeks after implantation (primary tumor removed). (G andH) show local lymph-node metastases in a mouse bearing a SiHa EGFP tumor six weeks after implantation (primary tumor removed). (I and J) showan individual lung lobe from a mouse bearing a ME-180 DsRed tumor six weeks after implantation in which metastatic lesions can be observed in thefluorescent image. (K and L) show a lung lobe from a mouse bearing a CaSki EGFP tumor two weeks after implantation. Table 2.  Incidence of spontaneous metastasis from orthotopiccervical tumors. Metastatic spread was assessed in animals bear-ing cervical tumors (0.4–2.0 g) using a fluorescent stereomicro-scope.Lymph node Lung Other sitesCaSki 94% (15/16) 94% (15/16) 25% (4/16)ME-180 90% (9/10) 30% (3/10) 0% (0/10)SiHa 80% (4/5) 20% (1/5) 40% (2/5) rather than by suturing a tumor fragment in place. We hadpreviouslyfoundthat thistechniqueproducesalargeamountof heterogeneity in tumor size, likely due to the difficultyin injecting a consistent volume into the small target tissue.While this would normally be undesirable, in this case itallowed the relationship between tumor size and metastaticburden to be investigated at a single time point. Tumor sizeand lung metastatic burden were measured after 10 days,and a correlation was found between the two parameters(Figure 4). This is consistent with clinical data where tumorsize correlates with the presence of involved lymph nodes atpresentation [12]. Discussion The cell lines generated for use in this study displayed thecharacteristicsrequiredofafluorescenttumormodelsystem.They could be readily distinguished from their parental pop-ulationsandfrommousetissue  invivo  basedonfluorescenceintensity. Although the quantification of the stability of thefluorescent phenotype indicated that not all cells maintainedtheir fluorescence, especially during growth in the lungs of mice, this did not prevent small lesions from being detected in vivo . However, it does suggest that not all tumor cells willbe identified using fluorescence-based detection modalities,and this must be considered when performing experimentsusing such a model. Further investigation of the mechanismresponsiblefortheinactivationofthereportergenemayhelpto improvethe model in this regard.We also foundthat therewere no differencesbetweenthe cells stably transfectedwitheither the EGFP or DsRed fluorescent proteins in any of theparameters that were measured, including plating efficiency, in vitro  and  in vivo  growth rate, and tumor histology.The optical detectionof fluorescent tumorcells  in vivo  asdescribed here is a powerful tool for the study of both solidtumors and metastases in laboratory animals, and has beendiscussedpreviously,althoughthestabilityofthefluorescent
Search
Similar documents
View more...
Tags
Related Search
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks