Establishment and genetic characterization of an immortal tumor cell line derived from intestinal-type sinonasal adenocarcinoma

Abstract

Background

Intestinal-type sinonasal adenocarcinoma (ITAC) is a rare tumor etiologically related to professional exposure to wood dust. The overall prognosis is poor, mainly due to the difficulty to resect the tumor completely in this anatomically complex region. Therefore, there is great need for alternative treatments. However, the lack of a good tumor model system for ITAC has hampered the development and testing of new therapeutic agents. Here, we report the establishment and characterization of the first human ITAC cell line named ITAC-3.

Methods

The cell line was initiated from small explants of a T4bN0M0 colonic type ITAC from the ethmoid sinus. Growth and invasion parameters as well as genetic characteristics were analyzed.

Results

The population doubling time was 18 h and the cell line was capable of invasion in matrigel. Chromosomal analysis showed a tetraploid karyotype with both numerical and structural aberrations. High resolution microarray CGH analysis identified many copy number alterations, including homozygous deletions. TP53 carried a mutation c.818G>T in exon eight concurring with a strong nuclear protein overexpression. Immunohistochemical analysis showed protein overexpression of EGFR and normal expression of β-catenin and p16.

Conclusion

This is the first report of the establishment of a cell line derived from a primary ITAC. The genomic profile of the cell line was the same as the primary tumor from which it was derived. This new cell line will be a useful tool for the development and testing of new therapeutic agents for this tumor type.

Introduction

Intestinal-type sinonasal adenocarcinomas (ITACs) arise in the respiratory epithelium of the ethmoid sinus and the upper part of the nasal cavity [1, 16, 23]. ITAC is named intestinal-type because of their histopathological resemblance to colorectal adenocarcinoma. However, unlike colorectal adenocarcinoma, ITAC is etiologically associated with professional exposure to wood dust particles, making it a disease almost exclusive to carpenters and furniture makers [4, 16, 33].

Five histopathological subtypes of sinonasal ITAC are distinguished: papillary, colonic, solid, mucinous (alveolar goblet and signet ring), and mixed (transitional), the most frequent type being colonic (40%) [8]. ITAC is not known to arise from a clearly defined precursor lesion, however, squamous metaplasia and dysplasia in the vincinity of the tumor have been described [31]. Distant and lymph node metastasis are exceptional (5%–10%) [6, 30].

Standard therapeutic modalities include surgery followed by radiotherapy in advanced stages, sometimes with chemotherapy treatment. The overall prognosis is poor, mainly due to the difficulty to resect the tumor completely in this anatomically complex region. Survival is better in papillary and colonic type than in solid or mucinous type adenocarcinomas. When there is intracranial invasion, the 5-year survival is virtually zero [28]. The principal cause of mortality is local recurrence (40%–50% of all cases), occurring even when the tumor was extracted with wide tumor-free resection margins. Therefore, ITAC represents an important occupational health problem with serious consequences, needing better ways of prevention, early diagnosis and treatment.

Recent studies characterizing the alterations in molecular genetic and cellular pathways of ITAC contribute to develop new ways for prognosis and treatment. Because of its histopathological resemblance [22], most studies so far have focussed on a limited number of genes and proteins known to be involved in colorectal adenocarcinoma (CRC). K-ras mutations are less frequent, with reports of 0%–15% [25, 29, 34, 35] although one study reported 50% [10], about the same as in CRC. Mutations in TP53, with a frequency of 60%–70% in CRC, ranged between 18% and 57% in ITAC [10, 33, 34], and 50% of loss of heterozygosity was detected at 17p13, the chromosomal locus of TP53 [27]. Other genetic alterations known in CRC, such as APC, β-catenin, p16 and mismatch repair genes were absent in ITAC [10, 26, 35]. Aiming for a genome-wide view of recurrent genetic abnormalities, CGH (Comparative Genomic Hybridization) and microarray CGH analysis have been performed and revealed frequent gains on chromosome arms 5p, 7q, 8q, 12p, and 20q, and losses on 4, 5q, 8p, 17p and 18q [2, 11, 17]. This genomic profile is partly similar to that of CRC, but not completely.

Stable tumor cell lines as a model system for ITAC retaining the biological properties of primary cancer cells are an important tool to further advance the development and testing of new therapeutic agents. Here we report the establishment and characterization of the first human ITAC cell line named ITAC-3. The cell line was characterized by its morphology, doubling time, invasiveness in matrigel, chromosomal analysis, microarray CGH, TP53 and KRAS mutation and the protein expression of a number of cancer genes previously studied in series of primary ITAC in the literature. In addition, we demonstrate that the cell line is representative of the original primary tumor from which it was derived.

Material and methods

Patient and tumor characteristics

The primary tumor sample was obtained from a previously untreated male patient, 74 years of age, who had worked for 12 years as a carpenter 30 years before the time of diagnosis. The patient presented with a large tumor originating in the left ethmoid sinus, occupying the whole of the left nasal cavity with bilateral ethmoid invasion and affecting the cribriform plate and dura mater. There was no invasion into the brain at the time of diagnosis. The tumor was surgically resected with free margins and postoperatory treated with radiotherapy, staged as T4bN0M0 and classified as an intestinal-type sinonasal adenocarcinoma of the colonic type (or Papillary Tubular Cylinder Cell II). Informed consent was obtained from the patient and the study was approved by the ethical committee of our institute.

Cell line establishment

A fresh tumor sample from the operating theatre was cut into several small fragments, transferred to dry 25 cm2 culture flasks and covered with a drop of serum-free culture medium (HuMEC culture medium (Invitrogen, Barcelona, Spain) and incubated in 5% C0₂ at 37°C. Initial outgrowth of both tumor and fibroblast cells was observed after 7 days. Possible overgrowth by fibroblasts was prevented by repeated selective trypsinizations. After a period of 2 months a first passage was attempted and after 4 months cells started to grow at a faster rate and were passaged weekly. At the moment of writing this manuscript, the cell line has been in culture for 12 months and has been passaged more than 60 times. Possible contamination by Mycoplasma was checked using the LONZA MycoAlert Mycoplasma Detection Kit (LONZA, Rockland CE, USA).

Growth rate and invasiveness

The growth rate was assessed by seeding 3 × 10⁵ cells in 25 cm² culture flasks on day 0, and at 24 h intervals the cells were trypsinized, resuspended in medium and counted using a hemacytometer. This experiment was done in triplicate.

Cellular invasion was evaluated by studying the ability of cells to invade matrigel membranes (BioCoat Matrigel Invasion Chambers; BD Biosciences, San Jose CA, USA). A suspension containing 20 × 10³ cells taken up in medium without supplements were placed in the upper compartment and the lower compartment was filled with complete medium, and incubated 24 h. Subsequently, the membranes were stained with 0.5% crystal violet in methanol and invasion of cells was evaluated using a light microscope with a 10 X objective.

Short-term cultured keratinocyte cells derived form normal nasal mucosa and grown in a serum-free human keratinocyte growth medium (Keratinocyte SFM Invitrogen, Barcelona, Spain) were used as negative control.

Immunohistochemistry and immunofluorescence

The following antibodies were used: Anti-CK20 clone Ks20.8 (DAKO, Glostrup, Denmark), anti-p53 clone DO-7 (DAKO, Glostrup, Denmark), anti-p16 clone E6H4 (CINTEC, MTM Laboratories, Madrid, Spain), anti-EGFR clone 2-18C9 (DAKO, Glostrup, Denmark) and anti- β-catenin clone β-catenin -1 (BD Biosciences, San Jose CA, USA). Immunohistochemistry was performed using an automatic staining workstation (Dako Autostainer, Dako Cytomation, Glostrup, Denmark) with the Envision system and diaminobenzidine chromogen as substrate. Immunofluorescence took place on ITAC-3 cells grown on microscope slides. Results were visualized by fluorescent secondary antibodies (Goat-anti-Rabbit, Alexa 488 and Goat-anti-Mouse, Alexa 488 (InVitrogen Molecular Probes, Barcelona, Spain) and DAPI counterstain (Vector Laboratories, Burlingame CA, USA), and evaluated using the Olympus BX-61 fluorescence microscope.

Chromosome analysis, SKY and FISH

Metaphase preparations were made according to standard procedures. G-band conventional karyotyping was performed and described according to the ISCN. SKY was done as described previously [12] using commercial SKY Paint probes (Applied Spectral Imaging, Migdal Ha’Emek, Israel). FISH was performed using a directly labeled dual probe EGFR/centromere 7 (Kreatech Diagnostics, Amsterdam, The Netherlands) following the manufacturer’s recommendations. Images were captured using the Olympus BX-61 fluorescence microscope mounted with a SpectraCube system and analyzed with the SkyView and FISH view software version 1.6.1 (Applied Spectral Imaging, Migdal Ha’Emek, Israel).

DNA extraction

DNA was extracted with Qiagen extraction kits (Qiagen GmbH, Hilden, Germany) from the cell line, the primary tumor and from normal blood lymphocytes of the same patient.

Microsatellite instability (MSI)

Ten ng of DNA from the cell line, the primary tumor and from blood lymphocytes was amplified in a multiplex PCR using a MSI analysis kit consisting of primers for five nearly monomorphic mononucleotide markers: BAT-25, BAT-26, NR-21, NR-24, and MONO-27 according to the manufacturer’s recommendations (Promega Biotech Iberica, Barcelona, Spain), as described previously [21].

TP53 and KRAS mutation analysis

TP53 mutations in exons 5–9 and KRAS in exon 2 (codons 12 and 13) were analyzed by direct sequencing using the ABI PRISM 3100 and 3730 Genetic Analyzer, (Applied Biosystems, Foster City CA, USA) using the following primers.

• TP53 Exon 5–6: Forward TGACTTTCAACTCTGTCTCC Reverse GCCACTGACAACCACCCTTA

• TP53 Exon 7: Forward CCAAGGCGCACTGGCCTCATC Reverse TCAGCGGCAAGCAGAGGCT

• TP53 Exon 8–9: Forward GACCTGATTTCCTTACTGCCTC Reverse GACTGGAAACTTTCCACTTGA

• KRAS Exon 2: Forward TACTGGTGGAGTATTTGATAGTG Reverse CTGTATCAAAGAATGGTCCTG

Sense and antisense sequencing was performed for confirmation.

Microarray CGH

Microarray CGH analysis was performed as described by Buffart et al. [3], using a 180 k oligonucleotide array (SurePrint G3 Human CGH Microarray Kit 4 × 180 K, Agilent Technologies, Palo Alto CA, USA). Images were acquired using a Microarray scanner G2505B (Agilent Technologies, Palo Alto CA, USA). Analysis and data extraction were quantified using feature extraction software (version 9.1, Agilent Technologies, Palo Alto CA, USA). Gains and losses were defined as deviations of 0.2 or more from log2 ratio = 0.0. Homozygous deletions (HDs) were defined as two or more consecutive oligonucleotides with a Log2 ratio value of < −2. The possibility of copy number variations (rather than copy number alterations) was excluded by using normal DNA of the same patient as reference.

Results

Cell morphology

ITAC-3 grew as a tightly packed monolayer of small cobblestone shaped cells with marked nucleoli visible in the nucei, not unlike the morphology of cells that can be seen in a H&E section of the patient’s primary tumor (Fig. 1). Immunological staining of intestinal-type differentiation marker cytokeratin 20 showed a only a few positively staining cells in the primary tumor.

Fig. 1.

Photomicrograph of a representative H&E stained paraffin section of the original primary tumor showing a colonic type (PTCC-II) ITAC. A complex, mixed tubulopapillary pattern can be seen; the papillae are covered by an intricate meshwork of tubules (a). A higher magnification of the primary tumors cells (b) shows a morphological appearance similar to the in vitro ITAC-3 cells (c) at 70%–80% confluence, passage 28. Immunohistochemical staining with antibody anti-CK20 was virtually negative (d). Original magnification 200x (A, C, D) and 400x (B)

Growth characteristics

The growth curve of ITAC-3 cells showed a slow start after seeding, leading even to a decrease of cells after 24 h. In the logarithmic phase, between 24 and 96 h, a population doubling time of approximately 18 h was observed. After passage ten, this growth has remained stable up until the present time (passage 60). Using the matrigel invasion assay, we found that ITAC-3 had a high potential of invasion with a large number of the originally seeded 20 × 10³ cells present in the matrix, while the normal keratinocyte cells did not invade at all. ITAC-3 was negative for Mycoplasma.

Genetic profile

Microarray CGH analysis of ITAC-3 showed copy number alterations involving most chromosomes (Fig. 2a) including gains at 1q, 3q, 4p, 5p, 6p+q, 13q, 18p, 19p+q, and 20p+q, and losses at 1p, 3p, 4p, 4q, 5q, 9q, 10p+q, 12p, 12q, 13q, 15q, 16p+q, 17p+q, 18q and 21q. High level amplifications were not observed. Thirteen small homozygous deletions were present at 1p, 3p, 3q, 4p, 4q, 5q, 13q, 17q, 22q and Xq, some of which indicated complete absence of DNA material (Table 1). One HD occurred at 13q14.1 very close to where RB1 is located and one other at 17p12 within the gene MAP2K4. Microarray CGH analysis of DNA extracted from frozen tumor tissue of the primary tumor from which ITAC-3 was derived showed an almost identical pattern of copy number alterations (Fig. 2a–b), however, the amplitude of the gains and losses was lower in the primary tumor sample, probably due to the contamination of normal cells, such as stroma and infiltrating lymphocytes.

Fig. 2.

Microarray CGH analysis of the ITAC-3 cell line at passage eight (a) and its original primary tumor (b). All data points are expressed as log2-ratios, ordered continuously from left to right as chromosomes 1 up to chromosomes X (here numbered as 23). Many chromosomes show copy number alteration, including small homozygous deletion but no high level amplification. Representative DAPI banding (c) and SKY karyotype (d) of cell line ITAC-3, showing both numerical and structural chromosomal rearrangements (the latter marked with asterisks). Only two unbalanced translocations involving different chromosomes were observed, both involving chromosome 7. Based on chromosome banding and SKY analysis the following composite karyotype was determined: 88–102, <4n>, -Y, -Y, i(1)(q10), −2, −3, del(3)(p11), del(4)(p?)(q?), del(4)(p?)(q?), del(4)(p?)(q?), +del(4)(p?)(q?), +dup(4)(p?), del(5)(q?), +i(5)(p10), +6, der(7)t(7;12)(p11;q22), dic(7;21)(p11;p11), +i(7)(p10), +9, −10, −10, −11, −12, +del(13)(q?), +del(13)(q?), −15, −16, −16, −17, −17, dup(18)(p?)dup(18)(q?), del(18)(q?), del(18)(q?), +19, +20, +20, +20, +20, +20, −21 [cp10]

Table 1.

Homozygous deletions as detected by microarrayCGH

Chromosome	Log2 ratio	Begin (bp)	End (bp)	Size (bp)	Gene
4p16.1	−4,5	9775301	9873149	97848	SLC2A9
4p15.1	−3,5	34418775	34525462	106687	–
4q22.1	−5	94794589	95042493	247904	–
4q35.2	−7	189707059	190425991	718932	–
5q31.1	−5	133026200	133166028	139828	–
13q14.2	−5,5	47759152	47950933	191781	(near RB1)
13q21.1	−6	57500241	57692811	192570	MAP2K4
17p12	−5	11919137	11934628	15491	–
Xq21.31	−5	87581689	88091482	509793	–
Xq21.31	−4	90222177	90278224	56047	–
Xq22.3	−5	103900994	104234719	333725	IL1RAPL2
Xq27.3	−5	143727135	143748649	21514	–
Xq27.3	−5	144827517	145115847	288330	SLITRK2

Chromosome banding and SKY analysis demonstrated a tetraploid karyotype with both numerical and structural aberrations. A representative karyogram and a full description of the karyotype is given in Fig. 2c–d. Two structural abnormal chromosomes involved more than one chromosome, both affecting chromosome 7. A number of small, fragmented chromosomes that were unidentifiable by DAPI banding were shown by SKY to consist of parts of chromosomes 4, 13 or 18. Microarray CGH confirmed the marked instability of these three chromosomes, that resulted in many different copy number levels of parts of the same chromosome. MSI analysis showed a microsatellite stable phenotype.

TP53, KRAS and other cancer genes

Sequencing of exons 5–9 of TP53 revealed a missense mutation c.818G>T [p.Arg273Leu] in exon 8 (Fig. 3a–c). This concurred with a strong nuclear p53 protein overexpression in the original tumor as well as in the cell line (Fig. 3e–f). The cell line was homozygous for the mutation indicating that the other allele is lost (Fig. 3b), which concurs with the microarray CGH analysis showing a deletion of whole chromosome 17 (Fig. 3d). The primary tumor did show a double peak at c.818 (Fig. 3c), but this is probably due to the normal cell contamination in the tumor DNA sample. KRAS did not carry a mutation.

Fig. 3.

TP53 gene copy number, mutation and p53 expression. Sequence analysis of part of exon 8 of a normal control DNA (a), the cell line (b) and the primary tumor (c). A missense mutation was identified both in the cell line as in the primary tumor at c.818G>T in exon 8. Microarray CGH analysis showed a deletion of whole chromosome 17 in the cell line (d). Strong nuclear p53 overexpression both in the original primary tumor (e) as in the ITAC-3 cell line (f). Original magnification 200x (E) and 1000x (F)

Microarray CGH indicated a small gain at the locus of EGFR on chromosome 7p12 (Fig. 4a). Using FISH we confirmed the copy number increase of EGFR. A total of five signals were seen per nucleus. On metaphase spreads, we identified three signal pairs (a red EGFR and a green centromere seven signal together), two chromosomes with only the signal of centromere seven and one chromosome carrying a centromere signal flanked on two sides by EGFR signals, indicating isochromosome 7p (Fig. 4b). Immunohistochemistry showed a moderate overexpression of EGFR, mostly in the cell cytoplasm, both in the primary tumor and in the cell line (Fig. 4c–d).

Fig. 4.

DNA copy number and protein expression analysis of EGFR. a Microarray CGH data of the cell line showing all clones representing chromosome 7, displayed from 7p to 7q (left to right). A stretch of clones from 54,846,884 to 55,258,775 have log2-ratios of around 0.5, signifying a gain of material at 7p11.2 (arrow). EGFR is the only gene located in this amplification. b FISH analysis on metaphase spreads of the cell line using a dual EGFR/centromere seven probe, labeled with PlatinumBright550 (red) and PlatinumBright495 (green), respectively. The metaphase shows three red/green pairs indicating a normal chromosome 7, two green centromeric signals without the red EGFR signal, and one centromeric green signal surrounded by two red EGFR signals, indicating isochromosome 7p. These latter two red EGFR signals seem stronger and may explain the gain of EGFR seen in the microarray CGH result. c EGFR protein overexpression detected in the cytoplasm and the cell membrane in the original primary tumor by immunohistochemistry (C) and in the cell line by immunofluorescence (d). Original magnification 200x (C) and 600x (D)

Both the cell line and the primary tumor demonstrated a normal membrane expression of β-catenin and a normal nuclear expression of p16.

Discussion

Sinonasal ITAC is an aggressive tumor associated with a high morbidity. Despite improvements in surgery in this anatomically complex region, still about half of all patients die within 5 years of the first diagnosis, mostly due to local recurrence and intracranial invasion. Knowledge on the genetic changes involved in ITAC is increasing but the preclinical development and testing of new therapeutic agents is hampered by the lack of an appropriate model system. Gelbard and co-workers described a mouse model for sinonasal cancer in which they implanted tumor cells orthotopically in the maxillary sinus or soft palate, enabling the study of local invasion, intracranial extension and lymph node or distant metastasis [7]. However, the applied cell lines were not originally derived from sinonasal tumors. In this paper, we describe the establishment and characterization of the first sinonasal intestinal-type adenocarcinoma cell line, growing as an in vitro cell culture.

The primary tumor from which ITAC-3 was derived is of the colonic or PTCC-II subtype, which generally show an intermediate clinical course between papillary and solid or mucinous tumors [20]. Nevertheless, this patient presented with a large, advanced stage tumor that invaded into the other fossa and affected the dura mater. After a disease-free time of 8 months, the patient developed brain metastases and died of the disease. The primary tumor showed a very low expression of intestinal differentiation marker CK20. In general ITACs stain positive for CK20 but absence of staining has been demonstrated before [5, 24].

Many of the copy number alterations found by microarray CGH in this cell line have been found in earlier series of primary ITAC studied by conventional and microarray CGH [2, 11, 17], including gains on 3q, 5p, 7p and 20 and losses at 4q, 5q, 10, 17p, 18q and 19p. Exceptions are gains of 8q and 12p and loss of 8p, which were not observed in the cell line. TP53 mutations and p53 protein overexpression have been found between 18% and 57% in primary ITAC [10, 13–15, 19, 27, 34] and was also found in this cell line. A recent paper showed overexpression of EGFR in 18 of 55 ITAC (32%) and gene copy number increase in 10/18 cases [9]. ITAC-3 showed a moderate copy number increase of a small region at chromosome 7p12 (including the locus of EGFR) and also a moderate EGFR protein overexpression (Fig. 4). Microarray CGH analysis revealed several small homozygous deletions (Table 1). We ruled out the possibility that these might be copy number variations by using normal DNA from the same patient as reference DNA. Some of the homozygous deletions actually indicated the complete absence of DNA material, two of which occurred within genes, one in 13q14.1 at RB1 and one in 17p12 at MAP2K4. Using immunofluorescence we were able to confirm the loss of Rb protein expression in the cell line. HDs at both RB1 and MAP2K4 have been reported in colon cancer [18]. The importance of HDs is that they indicate certain cellular processes to be deficient, and they could be useful targets of new pharmacological therapies [32]. In a similar fashion, specific gains such as EGFR in ITAC-3 may justify the study of anti-EGFR agents and perhaps additional pharmaceuticals that target other players in the EGFR signaling pathway.

The absence of nuclear β-catenin staining is in line with Perez Ordonez et al. who found only membranous staining in all ten analyzed cases of ITAC [26]. However in one study, eight of 20 cases (40%) showed membranous, cytoplasmic and nuclear positivity [10]. In CRC nuclear staining of β-catenin is seen in virtually 100% of cases. In a series of 21 ITAC, Perrone et al. found 67% promotor methylation and 45% loss of heterozygosity of CDKN2A, the gene coding for p16 [27]. The ITAC3 cell line and its primary tumor CDKN2A did not show an abnormal copy number, nor was there a loss of p16 protein expression.

In conclusion, this paper describes the first stable tumor cell line derived from a sinonasal ITAC exhibiting a high proliferative and invasive potential and carrying a genomic profile that matches the original primary tumor from which it was derived. Together with the high resolution genetic data obtained, this cell line will be a useful tool for preclinical testing of new therapeutic possibilities.