Cancerization of cutaneous flap reconstruction for oral squamous cell carcinoma: report of three cases studied with the mtDNA D‐loop sequence analysis

Foschini M P, Morandi L, Marchetti C, Cocchi R, Eusebi L H, Farnedi A, Badiali G, Gissi D B, Pennesi M G & Montebugnoli L (2011) Histopathology58, 361–367 Cancerization of cutaneous flap reconstruction for oral squamous cell carcinoma: report of three cases studied with the mtDNA D‐loop sequence analysis


Introduction
Oral squamous cell carcinoma (OSCC) is the most common type of cancer affecting the oral cavity. Despite its frequency, diagnosis is often delayed, and in order to achieve complete tumour removal major surgical resection is often required. More extensive resection of OSCC is possible as tissue defects being reconstructed routinely with skin flaps obtained from the radial forearm and a microvascular anastomosis at the site of reconstruction. 1 Recurrences and second primary OSCC are not uncommon in the oral cavity, with a development rate of 2-3% of new cases per year. 2 Several cases of squamous cell carcinoma or dysplasia arising from the grafted skin have been described (Table 1); however, knowledge on the behaviour of such lesions in the grafted skin is still scanty. 3,4 In 1953, Slaughter et al. introduced the concept of 'field cancerization' as a pathogenetic pathway for multiple primary OSCC arising in different areas, associated with preneoplastic lesions. 5 The authors 5 based the concept of 'field cancerization' on histological observations from OSCCs and associated preneoplastic lesions, such as dysplasia or in situ carcinoma, which developed in different areas in a single patient. 5 Subsequently, the concept of field cancerization has been supported by further clinical and histological evidence, and in the last decade several studies, based on molecular techniques, have evaluated the genetic basis of this concept. The oral mucosa undergoes malignant transformation through the development of genetically altered keratinocytes that progressively gain further mutations, most probably as a result of continued exposure to carcinogens, such as tobacco and alcohol. [6][7][8] Mutated keratinocytes with a growth advantage expand gradually and replace normal epithelial cells of the oral mucosa, favouring the development of a second primary OSCC. 9,10 Nevertheless, data supporting the concept of field cancerization are difficult to reconcile with the development of dysplasia and squamous carcinoma in cutaneous skin grafted into the oral cavity.
To date, no molecular studies have been performed on squamous carcinomas arising in the grafted skin; furthermore, the clonal relationship between primary OSCCs and secondary neoplastic changes appearing in the skin grafts still needs to be assessed, in order to differentiate recurrences from second primary OSCCs.
The high frequency rate of mtDNA mutations in tumours, [11][12][13] especially those found in the D-loop region, a non-coding region, along with numerous mitochondrial genomes present in a single cell, makes mtDNA a reliable marker for clonality assays from microdissected paraffin-embedded tissue samples. [14][15][16] We evaluated the clonal relationship between the primary cancer affecting the oral mucosa and the secondary neoplastic changes appearing in the skin graft by screening the mtDNA D-loop region in three OSCC patients.

Case histories case 1
In 2004, a 61-year-old female, a heavy smoker, presented with an ulcerated mass located in the retromolar trigone of the right mandible. An incisional biopsy diagnosed an invasive OSCC. Thereafter, the right mandible was removed surgically and substituted with a peroneal bone and an osteocutaneous flap comprising forearm skin. In February 2006 the patient presented with an exophytic mass of the right maxillary gingiva, which was diagnosed as OSCC by an incisional biopsy. The patient underwent right maxillectomy with reconstruction. In April 2006 the patient presented a polypoid lesion located in the centre of the cutaneous mandibular graft used to reconstruct the defect caused by the surgical resection of the first OSCC. The lesion was removed surgically.
The patient never discontinued smoking. She died of OSCC, 4 years after the initial surgery.  as invasive OSCC ( Figure 1B). The patient was a heavy smoker and previous personal and familial history was unremarkable. Radical surgical excision of the neoplastic mass was performed at the time and a cutaneous graft of forearm skin was used to repair the defect ( Figure 1C). No radiotherapy was performed. The patient discontinued smoking and was put on follow-up. Four years later, in April 2008, an ulcerated lesion appeared at the periphery ( Figure 1D) of the cutaneous graft, close to the margin with the oral mucosa. An incisional biopsy was performed to define the nature of the ulcer, diagnosing in situ OSCC of the cutaneous graft ( Figure 1E). The lesion was then removed completely.
In October 2008 a third lesion was observed, consisting of an erythroplasic area located at the right side of the skin graft, with no spatial relationship with the in situ OSCC. Five years after the first surgical resection the patient is alive, still on follow-up and free of neoplastic disease. In February 2005 a 52-year-old male patient, a heavy smoker, presented with an area of leukoplakia, ulcerated in the centre, in the anterior third of the right margin of the tongue. After the histological diagnosis of invasive OSCC, radical surgical excision of the neoplastic lesion was performed together with bilateral submandibular lymph node dissection. A forearm free flap was used to reconstruct the defect of the right margin of the tongue.
In August 2009 the patient presented with an exophytic lesion with focal ulceration of the cutaneous graft. The lesion was diagnosed as squamous cell carcinoma on incisional biopsy, and thereafter it was removed surgically.
Five months after the last surgical resection the patient is alive, on follow-up, and free of neoplastic disease.

Materials and methods
All tissues had been formalin-fixed and paraffin-embedded. Staging was performed according to the TNM staging system 17 and grading was carried out according to the criteria defined by Kademani et al. 18 Moreover, in all three cases, the skin flap specimens which included the second tumours were all sampled completely for histological examination.

microdissection and mtdna sequencing analysis
Pertinent lesions were microdissected using the laserassisted SL lcut microtest GmbH distributed by Nikon (Firenze, Italy; http://www.mmi-micro.com), as described previously. [19][20][21] Different portions of normal epithelial tissue located remotely from the neoplastic lesions were dissected and used as reference DNA. DNA was extracted using the QIAamp Ò DNA Micro kit (Qiagen, Hilden, Germany), following the manufacturer's instructions. An extraction control, to which no tissue was added, was processed in parallel with each sample extraction. The mtDNA D-loop sequence analysis was performed by amplifying four overlapping segments of about 300 bp, covering the whole region from position 16.056 to position 729, according to Anderson et al. 22 (see http://www.mitomap.org for revised Cambridge mtDNA reference sequence). Primers were designed using primer3 (http://www-genome.wi.mit. edu/cgi-bin/primer/primer3_www.cgi/; see Table 2 for sequence details). These primers were selected to avoid amplification of human mitochondrial pseudogenes in the nuclear genome. 23 Polymerase chain reaction (PCR) products were sequenced directly using a CEQ2000 XL instrument (Beckman Coulter, Inc., Fullerton, CA, USA), following the manufacturer's instructions.
Phylogenetic and cluster analyses were conducted using Molecular Evolutionary Genetics Analysis (MEGA) software version 3.1 [19][20][21] (http://www. megasoftware.net), using the neighbour-joining method (NJ) 24 and Kimura-2 parameter with Gamma model that corrects for multiple hits, taking into account transitional and transversional substitution rates and differences in the site substitution rates. 25 Every NJ tree was tested for standard error by the bootstrap method. 26 Results case 1

Histology
Histological examination of the tumour removed from the right mandible (indicated as T1 in the mtDNA analysis) confirmed the diagnosis of OSCC, staged T2N0M0, 17 grade 3. 18 Resection margins were tumour free, but a small neoplastic nest was present close (<5 mm) to the medial margin.
The tumour of the right maxilla (indicated as T2 in the mtDNA analysis) was diagnosed as a second primary OSCC, as it arose at a distance >30 mm from the first OSCC and was staged T4N0M0, 17 grade 3. 18 All resection margins were tumour free (distance >5 mm).
The lesion in the skin graft (indicated as T3 in the mtDNA analysis) showed in situ squamous cell carcinoma. 27 In addition, nests of neoplastic cells invaded the superficial dermis. Therefore, the diagnosis of OSCC with early dermal invasion was reported. Resection margins were free of carcinoma and of dysplastic features.

mtDNA analysis
The phylogenetic NJ tree revealed that the two normal epithelial samples (indicated as N1 and N2) clustered together, while the three lesions were considered to be independent entities with a clonal relationship between T1 (OSCC arising in the retromolar trigone of the right mandible) and T3 (the in situ OSCC arising on the skin graft) (Figure 2). T2 (the OSCC arising on the maxilla) appeared to be completely separate from T1 and T3, indicating the lack of a clonal relationship.

Histology
The tumour affecting the floor of the mouth (indicated as T1 in the mtDNA analysis) was a conventional type of OSCC, staged T1N0M0, 17 grade 2. 18 All resection margins were tumour free, with a minimum distance between the tumour and the closest margin of >5 mm. The lesion appearing in the skin graft (T2) met the criteria for in situ OSCC. 27 The resection margins were free of tumour and dysplastic lesions.
The oral mucosa adjacent to the skin graft margin (T3) was composed of atypical keratinocytes, arranged in an irregular architecture, replacing the lower two-thirds of the squamous epithelium. Therefore, the diagnosis of moderate dysplasia was made. 27

mtDNA analysis
The phylogenetic NJ tree showed a close genetic relationship between T1 and T2. Conversely, the dysplastic lesion which appeared in 2008 (T3) had several mutations that were not in common with the previous lesions. As expected, in the tree, normal epithelia (N1, N2 and N3) were located remotely from the four lesions ( Figure 3).

Histology
Histological examination of the OSCC removed from the right margin of the tongue (indicated as T1 in the mtDNA analysis) showed features of a conventional well-differentiated OSCC and no lymph node metastases were present. Thus, the tumour was staged T1N0M0, 17 grade 2. 18 The cancer was surrounded by an area of in situ carcinoma (T1IS). Resection margins were tumour free; however, a small neoplastic nest was present close (distance <5 mm) to the deep margin.
The histological analysis of the lesion developed in the skin graft (T2) showed in situ squamous cell carcinoma. 27 In addition, well-differentiated invasive OSCC was present.

mtDNA analysis
The phylogenetic NJ tree revealed that the two normal epithelial samples (N1 and N2) were clustered together, while the two lesions could be considered as independent entities. Furthermore, a moderate relationship between the T1 in situ carcinoma (T1IS) and both the in situ carcinoma (T2IS) and the OSCC (T2) arising from the skin graft was observed ( Figure 4). However, T1 (the OSCC arising in the tongue) appeared to have no clonal relationship and was divergent from the other carcinomas. This might be the consequence of new random acquired mutations during cellular proliferation.

Discussion
To define a second primary OSCC, mainly clinical criteria, such as distance between the first and the second primary tumour of >20 mm or a time interval between the tumours of >3 years, have been proposed. 28 However, increasing evidence suggests that in some cases, regardless of clinical characteristics, primary and second tumours may have a common clonal origin.
Several methods are available nowadays to assess the clonal relationship between two lesions. Clonal assessment by mtDNA analysis 29 is based on the concept that mitochondrial DNA is present in high copy number in each cell (10 3 -10 4 ) and that the vast majority of these copies are identical at birth (homoplasmic). In other words, neoplastic cells preserve the  Cutaneous flap cancerization 365 high copy number but show a high frequency of mutations in the mtDNA, especially in the D-loop region. Coller et al. 12 demonstrated that the mtDNA D-loop region is not involved in transformation and disease progression, as it would appear that mutations in this mtDNA segment do not offer a proliferation advantage to the cell. Therefore, these acquired mutations may be considered to be a reliable marker to assess clonality, as indicated previously. [14][15][16] In all three cases presented the neoplastic lesions arising in the skin graft showed a clonal relationship with the previous OSCC. Therefore, on the basis of the results obtained by mtDNA analysis, the present cases of OSCC arising in the skin graft can be considered as recurrences of the primary OSCC rather than a second primary OSCC.
However, these results seem difficult to reconcile with the clinical and histological features of these cases. All lesions arising in the skin graft were characterized by the presence of in situ OSCC. Furthermore, in cases 2 and 3, the skin graft lesions developed more than 3 and 5 years after the removal of the first OSCC, respectively.
These observations are, in turn, difficult to reconcile with the molecular analysis, which shows a clonal relationship between the primary OSCC and the neoplastic lesion affecting the graft. However, it should be borne in mind that we are dealing with skin grafted in the oral cavity and, as is well known, skin tumour cells can spread through the epidermis, as seen commonly in Borst-Jadasson phenomenon, 30 horizontal growth of melanomas and Paget's disease of the nipple.
All these phenomena, which are a part of routine surgical pathology, have been explained by demonstrating the capacity of epidermal keratinocytes to produce cytokines that allow and induce cell movements.
Currently available data indicate that the skin basically maintains its morphology, even when implanted into the oral cavity. Indeed, diseases typically affecting the skin, such as psoriasis 31 and focal acantholytic dyskeratosis, 32 have been reported in skin grafts. Therefore, it is not implausible to consider that skin grafted into the oral cavity may retain its ability to produce cytokines, continuing to give instructive signals. 4 Taken together, our results suggest that single neoplastic cells from the original OSCC can spread laterally, reach the epidermis of the skin graft and implant there, giving rise to the OSCC.
A hypothesis of the intraepithelial spread of neoplastic keratinocytes in the oral cavity has been proposed and described by Braakhius et al. 7 as the 'patch-fieldcarcinoma model'. Starting from a patch of genetically altered cells, through clonal expansion in a lateral direction, a field lesion develops and grows gradually, taking over the normal epithelium. Within this genetically altered field, a transforming event leads eventually to the development of a subclone with invasive growth that turns into a carcinoma.
Within this background, it is possible that after radical resection of the tumour, the genetically altered field is still present in the patient. Therefore, the field expansion continues and may spread into the skin graft used to reconstruct the tissue defect and be the cause of a new cancer.
In conclusion, primary OSCC and tumours arising in grafted skin might develop from a common genetically altered field. Moreover, the spread of the clonal cell population to the cutaneous flap might be stimulated by cytokines produced by the grafted skin.More studies are needed to evaluate the molecular relationship between primary and second OSCC in order to identify those patients who are at higher risk of developing a second tumour and thus require more assiduous long-term follow-up.