Topic Article:
Patients with recessive dystrophic epidermolysis bullosa develop squamous-cell carcinoma regardless of type VII collagen expression
Celine Pourreyron, Georgie Cox, Xin Mao, Andrea Volz, Nuzhat Baksh, Tracy Wong, Hiva Fassihi, Ken Arita, Edel A. O’Toole, Jorge Ocampo-Candiani, Mei Chen, Ian R Hart, Leena Bruckner-Tuderman, Julio C. Salas-Alanis, John A. McGrath, Irene M. Leigh, Andrew P. South
Journal of Investigative Dermatology (2007) 127, 2438-2444; doi:10.1038/sj.jid.5700878

Recessive Dystrophic Epidermolysis Bullosa and Squamous-Cell Carcinoma: The Role of Type VII Collagen
Robert S. Kirsner¹ and Shasa Hu¹
Journal of Investigative Dermatology (2007) 127, 2292; doi:10.1038/sj.jid.5701086

Recessive dystrophic epidermolysis bullosa (RDEB) is a severe mechanobullous disease characterized on the molecular level by mutations in the COL7A1 gene, which encodes the keratinocyte-secreted protein type VII collagen, and clinically by an increased risk of developing squamous-cell carcinoma (SCC). The consequences of SCC development can be significant—55% of patients with the Hallopeau–Siemens subtype of RDEB die from metastatic SCC by age 40 (Fine et al., 1999). Thus, patients with RDEB require close monitoring for development and treatment of potential life-threatening cancers. A recent paper by Ortiz-Urda et al. (2005) reported, as part of a broader work describing the role of Ras and type VII collagen in epidermal tumorigenesis, that in an analysis of 10 patients with RDEB who developed SCC, all 10 displayed type VII collagen within their tumors. The results suggested that only a subset of patients who express type VII collagen develop SCC. These findings led Pourreyron and colleagues (2007) to attempt to confirm the results. They studied a discrete subset of 11 patients with RDEB (8 with the Hallopeau-Siemens subtype) who developed 17 SCCs. Of 10 patients (16 tumors), 8 demonstrated type VII collagen expression. In 2 patients (3 tumors) no type VII collagen was detected. The 2 patients and their tumors without type VII collagen detection exhibited compound heterozygous nonsense mutations within the region encoding the NC1 domain of the COL7A1 gene. Through the following questions we delve into this paper in greater detail.

QUESTIONS

1. Why is the NCI domain crucial to SCC development according to Ortiz-Urda et al.?

2. Is the hypothesis being tested that patients with dystrophic EB require the presence of type VII collagen for development of SCC?

3. Why were immunofluorescence and immunoblotting used in this study?

4. What controls were used?

5. Do the findings of this study contradict the findings of Ortiz-Urda et al.?

6. What clinical implications does the article suggest?

ANWERS

1. The article by Ortiz-Urda et al. (2005) hypothesized that type VII collagen was critical for Ras-driven epidermal tumorigenesis. According to their article, support for this hypothesis included the findings that: (1) type VII collagen expression in SCC from patients (N = 10) with recessive dystrophic epidermolysis bullosa (RDEB) and (2) epidermal tumors indistinguishable from SCC grew after engraftment onto immunodeficient mice only when: (a) the primary keratinocyte cell lines were from RDEB patients who expressed type VII collagen and (b) they were transfected to coexpress Ras and a specific NFκB inhibitor. Cell lines expressing type VII collagen (8 of 12 total), expressed the 145-kDa amino-terminal noncollagenous domain (NC1). Thus, retention of the NCI protein fragment of type VII collagen distinguished the tumorigenic subset. When cell lines devoid of type VII collagen expression had the capability to produce type VII collagen restored, their ability to produce tumors followed.

A region of the NC1 protein that encompasses fibronectin III–like repeats (FNC1), which binds to laminin 5, was found to be critical to Ras-driven tumorigenesis. For example, antibodies to FNC1 injected intraperitoneally (at the time of subcutaneous injection of tumorigenic normal cells) prevented tumor formation. Additionally, antibodies to FNC1 prevented invasion through the basement membrane zone but did not inhibit tumor-cell proliferation or trigger apoptosis, suggesting the FNC1’s role related to tumor invasiveness.

On further exploration of the interaction between FNC1 and laminin 5, antibodies to FNC1 were found to interfere with binding of FNC1 and laminin 5. In turn, antibodies to laminin 5 inhibited tumor invasion. Supporting these findings was the observation that cells from patients with junctional EB that lack laminin 5 do not invade. It was thus concluded that type VII collagen, specifically the NC1 domain, was required for Ras-driven tumor development.

2. Drawing on the implications of the article by Ortiz-Urda et al. (2005), Pourreyron et al. (2007) set out to determine the role of type VII collagen in SCC development in patients with RDEB. They hypothesized that patients with dystrophic EB require the presence of type VII collagen for the development of SCC. Because it is never possible to prove a hypothesis, the researchers set about to try to disprove the null or opposite hypothesis. As more studies disprove the null, the more the hypothesis of interest is supported. In this case, the authors tested the null hypothesis that patients with dystrophic EB who lack type VII collagen expression still develop SCC.

The idea that patients with dystrophic EB require type VII collagen for development of SCC is curious from both an epidemiologic and a clinical standpoint. Previous reports demonstrated that patients with dystrophic EB have greater risk for SCC compared with other EB subtypes. The distinguishing feature of dystrophic EB is mutations in type VII collagen. The most severe mutations seen in patients with RDEB (who are thought to lack type VII collagen) are associated with the highest risk of SCC development. If type VII collagen was important, one might expect a greater risk of SCC development in other subtypes of EB, in which collagen VII is not affected. Although SCC can occur in other types of EB, it does not occur as often as in dystrophic EB (Fine et al., 1999).

In this study, the authors found patients with dystrophic EB lacking type VII collagen who did develop SCC. By doing so they disproved the hypothesis that patients with dystrophic EB require the presence of type VII collagen for the development of SCC.

3. In testing the null hypothesis that patients with dystrophic EB who lack type VII collagen can develop SCC it is critical to ensure that patients said to lack type VII collagen expression truly lack type VII collagen (including both mRNA and protein expression). One might imagine that using immunofluorescence to detect antibodies to collagen VII would be sufficient to prove that the SCCs have no type VII collagen. There are several potential flaws in the use of this approach as the sole method of determining whether type VII collagen is present. One is that immunofluorescence detects the binding of an antibody to an antigen. In this study, of 10 patients (16 tumors) who were studied for presence of type VII collagen, 8 demonstrated collagen VII; in 2 patients (3 tumors) there was no detectable type VII collagen. However, it is possible that the protein (in this case, type VII collagen) is produced but then degraded; hence, immunofluorescence might be negative. Another situation in which immunofluorescence might be falsely negative is when low levels of the protein of interest interfere with staining or when the protein does not contain the binding region for the antibody employed. Thus, lack of staining of collagen VII in SCC from REDB patients does not necessarily confirm that those patients do not produce type VII collagen.

Therefore, further testing is required. Specifically, one should examine both keratinocytes from nonaffected patients and, importantly, keratinocytes isolated from normal skin of RDEB patients. One method uses immunoprecipitation, which can provide a means to separate a specific protein from whole-cell lysates or culture supernatants. Additionally, one can use immunoprecipitation to confirm the identity of a protein of interest or to study its biochemical characteristics, posttranslational modifications, and expression levels. However, with immunoprecipitation, one needs to know the protein of interest. Unfortunately, when a protein is mutated, degraded, or denatured, immunoprecipitation may be negative. In such cases immunoblotting may be superior in detecting a variety of proteins of different molecular weight that may have been produced by the RDEB keratinocytes.

4. Although studying a population of patients is of interest, findings take on greater importance when appropriate controls are included. In this study, in addition to studying SCCs from patients with RDEB, it was important to study other patients (or cells) for comparison and to ensure that the techniques were performed correctly.

In this study several controls were used. The investigators first examined type VII collagen expression in non-EB SCC frozen tumors and reported that 20 of 20 demonstrated type VII collagen (all expression was restricted to the basement membrane). Second, the investigators, after identifying 3 (of 16) SCCs from 2 (of 10) patients with RDEB lacking type VII collagen staining, studied type VII collagen expression in normal skin and in keratinocytes isolated from 9 other RDEB patients. All had type VII collagen expression similar to that seen in SCCs.

The investigators then studied isolated keratinocytes from non-EB SCC tumors and from RDEB SCC tumors (6 patients). All the keratinocytes from non-EB SCC tumors demonstrated type VII collagen by immunoblotting. Of the 6 RDEB SCC tumors, 2 of 6 RDEB SCC keratinocytes demonstrated type VII collagen.

5. Pourreyron et al. (2007) did not test whether type VII collagen is required for Ras-driven tumorigenesis; rather, they studied the role of type VII collagen in patients with dystrophic EB. More specifically, they tested whether collagen VII is required for SCC development. As noted above and as part of a more general theme, Ortiz-Urda et al. (2005) demonstrated that collagen VII (specifically FNC1) was required for Ras-driven SCC-like tumor development, which included studying a subset of patients with RDEB. Once conclusion based on that work is that patients with RDEB who do not express any type collagen VII would not be at the same high risk of developing SCC and would require less vigilant monitoring. If true, this would run counter to conventional wisdom, because it has long been thought that patients with RDEB with the least (or no) expression of collagen VII have both the most severe clinical manifestations and the greatest risk of developing SCC.

Pourreyron et al. found that SCC developed in their REDB patients regardless of type VII collagen expression. They studied a discrete subset of 11 patients with RDEB (8 with the Hallopeau-Siemens subtype) who developed 17 SCCs. In 10 patients (16 tumors) who were studied for the presence of type VII collagen, 8 demonstrated type VII collagen expression. In 2 patients (3 tumors) no type VII collagen was detectable. In these 2 patients, nonsense mutations within the NC1 domain were detected. However, no Ras mutations were identified in any of the SCCs studied. Because neither the SCCs with collagen VII nor those without collagen VII expression contained Ras mutations, the authors neither confirmed nor disproved that collagen VII (or specifically the NC1 domain) is essential for Ras-driven tumorigenesis, which was the finding of Ortiz-Urda et al.

Overall, the study by Ortiz-Urda et al. was oriented toward tumorogenesis in Ras-driven SCC development, whereas the study by Pourreyron et al. focused on SCC development in RDEB patients and as such did not find a relationship between collagen VII expression and Ras mutations. The data in toto may be taken to indicate that at least two factors may be at work in the development of SCC in RDEB patients: that type VII collagen is critical through a Ras-driven process and that type VII collagen expression is not critical.

6. To the reader without specific expertise in EB—in particular, RDEB—this article highlights a number of important clinically relevant features.

First, EB represents a widely heterogeneous group of diseases. Even within subsets of RDEB, numerous unique mutations can occur and thus individualize a patient’s prognosis and course. Second, the subgroup of RDEB patients are at high risk not only for developing SCC but also for succumbing to SCC at a young age (Fine et al., 1999). Third, many might have casually learned that in addition to genetics, the difference between recessive and dominant DEB was the complete (vs. partial) absence of type VII collagen expression in the former. As this paper highlights, this is an oversimplification. RDEB patients may have either expression or lack of expression of type VII collagen. This point, gleaned from previous work (Ortiz-Urda et al.), suggested that expression of type VII collagen might stratify patients’ risk of SCC development. Fourth, in the development of SCC, Ras appears to be an important oncogene, and, when present, type VII collagen—specifically the FNC1 region and its binding to laminin 5—appears critical for Ras-driven tumorigenesis. The FNC1 region, in combination with laminin 5, appears to act by enhancing tumor invasiveness. Fifth, this interaction (and each individual component) involves potentially therapeutic targets in the prevention of tumor invasion, spread, and metastasis. Sixth, other mechanisms of carcinogenesis besides Ras and type VII collagen expression are likely important in the increased risk of patients with RDEB for development of SCC. For example, SCC can develop in RDEB patients in the absence of Ras mutations or type VII collagen expression. Finally, this study confirms that, regardless of type VII collagen expression, RDEB patients may still be at risk for developing SCC, highlighting the fact that close monitoring and follow-up continue to be essential in these patients.

REFERENCES

Fine JD, Johnson LB, Suchindran C, Bauer EA, Carter M, McGuire J et al. (1999) Cancer and inherited epidermolysis bullosa: lifetable analyses of the National Epidermolysis Bullosa Registry population. In: Epidermolysis Bullosa: Clinical, Epidemiologic and Laboratory Advances and the Findings of the National Epidermolysis Bullosa Registry (Fine JD, Bauer EA, McGuire J, Moshell A, eds), Johns Hopkins University Press: Baltimore, 175–92

Ortiz-Urda S, Garcia J, Green CL, Chen L, Lin Q Veitch DP et al. (2005) Type VII collagen is required for Ras-driven human epidermal tumorigenesis. Science 307:1773–6

Pourreyron C, Cox G, Mao X, Volz A, Baksh N, Wong T et al. (2007) Patients with recessive dystrophic epidermolysis bullosa develop squamous-cell carcinoma regardless of type VII collagen expression. J Invest Dermatol 127:2438–44

¹Department of Dermatology and Cutaneous Surgery, University of Miami Miller School of Medicine, Miami, Florida, USA