Expression Pattern of Aquaporin 1 and Aquaporin 3 in Melanocytic and Nonmelanocytic Skin Tumors

Giovana Osorio, MD, PhD; Teresa Zulueta-Dorado, MD, PhD; Patricia González-Rodríguez, PhD; José Bernabéu-Wittel, MD, PhD; Julian Conejo-Mir, MD, PhD; Reposo Ramírez-Lorca, PhD; Miriam Echevarría, PhD


Am J Clin Pathol. 2019;152(4):446-457. 

In This Article


The ongoing interest in the pathophysiologic and molecular mechanisms underlying disease in general and cancer in particular has led to a rapidly growing field aimed at providing clinicians with the tools required to improve the diagnosis, treatment, and/or prognosis of cancer. As a result, numerous specific therapeutic targets are now known. Aquaporins have provided new insight into the basic mechanisms underlying skin physiology and pathology and may therefore prove to be promising therapeutic targets in the diagnosis and treatment of skin cancer.

The analysis of healthy skin samples showed positive immunoreactivity for AQP1 in vascular endothelium and erythrocytes. Both eccrine and apocrine glands, which are secretory epithelia and therefore experience significant water flow, as well as dermal fibroblasts, were also positive. All these findings are consistent with previous studies in which AQP1 expression was shown in these tissues using the polymerase chain reaction technique and tissue microarrays[12] or immunohistochemistry.[13] However, in contrast to the findings published by other authors,[1] we never detected AQP1 expression in melanocytes from the healthy skin samples analyzed despite all our efforts to detect such positive immunoreactivity in this type of cell. The reason for this discrepancy remains unknown.

The expression pattern for AQP3 in healthy skin was very characteristic and in agreement with that reported previously.[9,14] Thus, as expected, intense staining for AQP3 was observed in keratinocytes throughout the epidermis and around pilosebaceous follicles. Eccrine and apocrine glands, fibroblasts, blood vessels, and erythrocytes also stained positive for AQP3, whereas melanocytes did not exhibit expression of this protein in healthy skin.

The expression of both AQPs in tumor tissue was also very different and specific for each protein. Thus, none of the tumor types analyzed expressed AQP1 in the carcinoma cells themselves, whereas all types of skin carcinoma studied exhibited marked AQP3 expression. This all-or-none expression pattern for these AQPs in the carcinomas studied herein limits their potential use as specific markers for this type of disease. However, this does not diminish interest in the potential roles played by each of these proteins in tumor development.

The tumor cells in all BCCs, SCCs, and melanomas studied were negative for AQP1, although marked AQP1 expression was observed in the stroma surrounding the tumor masses, as well as in the neovessels and lymphocytes infiltrating this stroma. Another interesting observation was the intense staining of AQP1 observed in all common melanocytic nevi but in none of the atypical melanocytic nevi. This unexpected AQP1 staining pattern is of particular interest given that melanocytes in healthy skin did not express AQP1, and neither did atypical nor malignant melanocytic tumors. In addition, and to our knowledge, no studies in the literature have analyzed AQP1 expression in the human skin tumors studied herein, and there are no known precedents of an upregulation/downregulation pattern for AQP1 expression in any other type of tumors. In light of our findings, it would be tempting to propose that AQP1 plays a key role in the transformation of normal melanocytes into a benign neoplastic state (nevus) and that the subsequent malignant transformation is independent of this protein. As such, it would be of interest to study the molecular basis underlying the regulation of this particular on/off process of AQP1 expression observed in melanocytic skin tumors.

The marked AQP1 expression observed in the peritumoral stroma, fibroblasts, and, in particular, the vascular endothelium of neovessels surrounding and infiltrating the tumor mass is in agreement with previous findings in various other tumor types. Thus, AQP1 expression has been associated with carcinoma of the colon, breast, brain, and lung, as well as with hemangioblastoma and multiple myeloma.[15–21] Animal studies have suggested a key role played by AQP1 in angiogenesis and cell migration within the vascular endothelium,[2] which is fundamental for the formation of new vessels to supply the tumor. It has been reported that AQP1 localizes in the main border of membrane protrusions during cell migration and is associated with various transporters, Na/H+ and Cl/HCO3 exchangers,[3] in keeping with the rapid movement of ions and water required during migration.

We also found the fibroblasts and lymphocytes surrounding tumor cells to be immunopositive for AQP1, which is consistent with a possible role for this protein in the inflammatory phenomena triggered by the immune system as part of its immunovigilance against tumors. The inflammatory infiltration and fibrosis surrounding the tumor also demonstrate the presence of this protein in those processes related to tumor development.

On the other hand, high expression of AQP3 was detected in tumor cells from both BCC and SCC in all samples evaluated in this study. Similarly, although AQP3 expression was not detected in melanocytes from healthy skin, intense staining of this protein was observed in both benign and malignant melanocytic lesions. Some authors have suggested that AQP3 overexpression in tumors is related to the basal expression of this protein and the function of the tissue or organ.[5] However, there is little evidence for the validity of this hypothesis in melanomas as, under normal conditions, melanocytes do not express AQP3 and melanomas are not glandular-type tumors.

Previous studies have linked AQP3 to BCC and SCC or epidermal tumors,[22] in which intense immunohistochemical staining was found in tumor areas of human primary SCCs, such as esophageal and lingual cancers.[22] In addition, a higher incidence of AQP3 expression and H-score was found in SCCs compared to BCCs in human skin biopsy specimens that were immunohistochemically analyzed.[23] Also, in cell lines with epidermoid differentiation (keratocarcinoma), endogenous AQP3 expression has been demonstrated at both the messenger RNA and protein levels,[8] and evidence for the importance of AQP3 in tumor formation also has been obtained from animal models. The hypothesis that glycerol transport via AQP3 is a determining factor for proliferation and cutaneous carcinogenesis based on its role in cellular energy adenosine triphosphate (ATP) production was postulated some time ago.[10] The incorporation of glycerol into lipid synthesis, ATP-facilitated mitogen-activated protein kinase signaling, and a positive-feedback loop through which cell proliferation increases AQP3 expression also have been proposed as part of the pathophysiologic mechanism of AQP3.[10] Nowadays, it is believed that AQP3 is involved in the migration, proliferation, and invasion of skin tumors, as well as in the worsening of the epithelial-to-mesenchymal transition.[24]

In conclusion, the results of this study are consistent with the involvement of both AQP1 and AQP3 in the development of skin tumors. AQP1 appears to be relevant for the neoangiogenesis required to sustain tumor cells and for the inflammatory process inherent to the tumor, which is in agreement with previous studies showing an association between AQP1 expression and the formation of new blood vessels as well as with a more aggressive and potentially metastatic profile in various types of human cancers.[2,17,25–28] Similarly, AQP3 overexpression in the three types of tumor analyzed—namely, BCC, SCC, and melanoma—probably confers a greater proliferative capacity on the tumor cell,[28,29] together with an increased passage of glycerol and/or solutes such as hydrogen peroxide,[29] which increase the energy reservoir in the cell and/or the potentially carcinogenic oxidative stress inside the cell.