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


Population-based Characterization of Samples From Primary Skin Tumors

The mean age of the patients from whom the 70 samples analyzed were obtained was 60 years (range, 21–89), excluding the controls. Thirty-eight (54.3%) of the patients were men and 32 (45.7%) were women. A total of 68 (97.1%) samples were primary tumors, with the remaining 2.9% (n = 2) corresponding to relapses. The most common location was the head and neck (45.7%, n = 32), followed by the trunk (30%, n = 21), upper extremities (15.7%, n = 11), and lower extremities (8.6%, n = 6). A more detailed description of the patients' demographic characteristics and the various biopsy subtypes is provided in Table 1, Table 2, and in the supplementary material (all supplemental materials can be found at American Journal of Clinical Pathology online).

Expression of AQP1 and AQP3 in Healthy Skin

In healthy skin, as can be observed in Image 1, AQP1 was expressed in the vascular endothelium and erythrocytes. The eccrine glands, which are secretory epithelia and therefore experience significant water flow, and dermal fibroblasts were also positive for AQP1 (Table 2). In contrast, expression of AQP3 was restricted to the epidermal keratinocytes, with the intensity thereof being higher in the stratum basale and decreasing toward the more superficial layers of the skin. Keratinocytes in pilosebaceous follicles were also immunoreactive for AQP3, as were the eccrine and apocrine glands. Finally, fibroblasts, blood vessels, and erythrocytes also stained positive for AQP3 (Image 1). The staining patterns for AQP1 and AQP3 in the different cells found in normal skin are summarized in Table 2.

Image 1.

Expression of aquaporin 1 (AQP1) and aquaporin (AQP3) in healthy skin. A, Negative staining for AQP1 in the epidermis but positive staining in vascular endothelium (arrows) and fibroblasts (arrowheads) (x20). B, Higher magnification of a small area shown in A (x40). C (x10), D (x40), Positive staining for AQP3 in epidermal keratinocytes (arrows). Keratinocytes found in the pilosebaceous follicle were also immunoreactive for AQP3 (arrowhead). E, Higher magnification of epidermal keratinocytes showing AQP3 staining. Staining for AQP3 was observed in keratinocytes of apocrine (F) and eccrine (G) glands (x40).

Expression of AQP1 and AQP3 in Tumors

Of the 70 tumors evaluated, 92.9% (n = 65) were negative for AQP1, with a further 1.4% (n = 1) exhibiting focal immunoreactivity and 5.7% (n = 4) diffuse staining. All AQP1-positive samples corresponded to common melanocytic nevi. Lesion size was inversely related to signal intensity in all cases (P = .01). With respect to AQP3 immunostaining, 98.6% (n = 69) of tumors exhibited diffuse positive staining, with focal positive staining being observed in the remaining 1.4%. None of the tumors analyzed were negative in the immunohistochemical study for this protein. Lesion size was inversely related to signal intensity in all cases (P = .01).

AQP1 and AQP3 Expression in Common and Atypical Nevi. Immunostaining for AQP1 was intense in all common melanocytic nevi (100%; Image 2), which is of interest given that the melanocytes in healthy skin were not positive and atypical or malignant melanocytic tumors were also negative for this protein (Table 2). Tumoral melanocytes were negative for AQP1 in all samples of atypical melanocytic nevi (100%, Image 2). Positive staining for AQP1 was found in dermal capillaries and peritumoral connective tissue.

Image 2.

Aquaporin 1 (AQP1) and aquaporin 3 (AQP3) expression in common and atypical melanocytic nevi. A, Epidermis is consistently negative, but nevi cells are strongly positive for AQP1 (arrows) (x20). B, Higher magnification of a small area shown in A (x40). C, Positive staining for AQP3 is observed in the entire epidermis and in all melanocytes of nevi theca cells (x20). D, Cells forming nevi thecas are shown at a larger magnification (arrows) (x40). E, All melanocytic thecas are negative for AQP1 (x20). F, Intense positive signal for AQP3 is observed in melanocytic nevus cells (x40). Samples correspond to common melanocytic nevi biopsy specimens (A-D) and atypical melanocytic nevi (E, F).

Although AQP3 expression has not been reported in either normal or proliferating melanocytes, staining of these cells was found in all common melanocytic nevi analyzed in this study, as well as in all atypical nevi (Image 2 and Table 2).

AQP1 and AQP3 Expression in Primary Skin Melanomas. As was the case for atypical melanocytic nevi, the tumoral melanocytes found in melanomas were all negative for AQP1. However, the small neovessels, peritumoral stroma, and melanophages surrounding the tumor were immunopositive for this protein Image 3.

Image 3.

Aquaporin 1 (AQP1) (A-D) and aquaporin 3 (AQP3) (E-H) expression in different types of melanomas: (A, E) nodular melanoma, (B, F) lentigo maligna melanoma, (C, G) superficial spreading melanoma, and (D, H) acral lentiginous melanoma. A-D, All melanocytic cells are negative for AQP1 staining; however, positive staining was detected in neovessels inside the tumor area, serving as an internal positive control (x40). Staining is seen in blood vessels (A, C, and D) (arrows) and melanophages (B) (arrowheads). E-H, All melanocytic cells by contrast show intense staining for AQP3 (x40).

It should be noted that positive staining for AQP1 was found in the neval cells from two melanoma samples; one was the superficial spreading type and the other was lentigo maligna, both of which grew on a preexisting neval lesion, although the melanoma cells themselves remained negative. This finding is consistent with the AQP1 staining pattern observed for each type of melanocytic tumor.

Similar to the situation for common and atypical nevi, AQP3 immunostaining was positive in all tumor cells from melanomas (Image 3), independent of their histologic subtype (Table 2). These findings suggest that AQP3 may be a marker for proliferating melanocytes as this protein is found in all melanocytic tumoral lesions but is not expressed in basal melanocytes.

AQP1 and AQP3 Expression in BCCs and SCCs. The tumor cells in all basal cell carcinomas studied were negative for AQP1, although marked AQP1 expression was observed in the stroma surrounding the tumor mass (Image 4 and Table 2). In addition, AQP1 immunostaining was negative in all SCCs. However, positive staining for AQP1 was observed in the peritumoral stroma and in the neocapillaries and lymphocytes infiltrating this stroma.

Image 4.

Aquaporin 1 (AQP1) (A, B, E, and F) and aquaporin 3 (AQP3) (C, D, G, and H) expression in basal cell carcinoma (A-D) and squamous cell carcinoma (E-H). AQP1 expression was not detected in either the epidermis (A, x10) or tumor cells (B, x40), but a strong positive signal was detected in the surrounding stromal tissue. A high rate of AQP3 expression was observed in basal cell carcinoma (C, x20). Higher magnification of the tumor zone of C is shown (D, x40). In squamous cell carcinoma, tumor cells were negative for AQP1 immunostaining, but capillaries in the surrounding stroma tissue (arrows) did express AQP1 (E, x20; F, x40). Intense staining for AQP3 was seen in tumor cells of squamous cell carcinomas (G, x20; H, x40).

High AQP3 expression was found in both BCC and SCC cells, independently of the tumor subtype (Image 4). As both of these neoplasms are derived from epidermal keratinocytes, AQP3 expression in them was to be expected (Table 2).