- Email: [email protected]
Increased dermal mast cell prevalence and susceptibility to development of basal cell carcinoma in humans
Methods 28 (2002) 90–96 www.academicpress.com
Increased dermal mast cell prevalence and susceptibility to development of basal cell carcinoma in humans Michele A. Grimbaldeston,a,* Lone Skov,b John J. Finlay-Jones,a and Prue H. Harta a
Department of Microbiology and Infectious Diseases, School of Medicine, Flinders University, Adelaide 5001, Australia b Department of Dermatology, Gentofte Hospital, University of Copenhagen, Copenhagen, Denmark Accepted 19 June 2002
Abstract Exposure to ultraviolet B (UVB) radiation (280–320 nm) is the primary etiologic factor associated with the development of basal cell carcinoma (BCC). The outgrowth of these keratinocyte-derived skin lesions is enhanced by the ability of UVB to impair an immune response that would otherwise eliminate them. Studies in a range of inbred mouse strains as well as mast cell-depleted mice reconstituted with mast cell precursors support a functional link between histamine-staining dermal mast cells and the extent of susceptibility to UVB-induced systemic immunomodulation. Humans, like mouse strains, display variations in dermal mast cell prevalence. In a study of Danish and South Australian BCC patients and control subjects, one 4-mm punch biopsy of non-sunexposed buttock skin was sampled from each participant. This skin site was investigated to avoid any changes in mast cell prevalence caused by sun exposure. Two sections (4 lm) per biopsy were immunohistochemically stained for detection of histamine-containing dermal mast cells. Computer-generated image analysis evaluated dermal mast cell prevalence in both sections by quantifying the total number of mast cells according to the total dermal area (expressed as mast cells per square millimeter). This technique enabled us to detect heterogeneity of dermal mast cell prevalence in buttock skin between individuals and provided evidence of an association between high dermal mast cell prevalence and BCC development in two diverse populations. We hypothesize that mast cells function in humans, as in mouse strains, by initiating immunosuppression following UV irradiation and, thereby, allowing a permissive environment for the development of BCC. Thus, a high dermal mast cell prevalence as demonstrable in buttock skin is a significant predisposing factor for development of BCC in humans. Ó 2002 Elsevier Science (USA). All rights reserved. Keywords: Ultraviolet immunosuppression; Histamine; cis-Urocanic acid; Skin cancer; Computer-generated image analysis
1. Introduction Many genetic and environmental factors contribute to the pathogenesis of skin cancer. Intermittent intense sun exposure, particularly to ultraviolet B (UV radiation of 280–320 nm), and a history of sunburn in childhood have been described as significant risk factors for the development of basal cell carcinoma (BCC)1 [1,2]. In addition to the initiation of skin neoplasms, UV irradiation also alters the cell-mediated immune response,
*
Corresponding author. Fax: +61-8-8204-4733. E-mail address: michele.grimbaldeston@flinders.edu.au (M.A. Grimbaldeston). 1 Abbreviations used: BCC, basal cell carcinoma; CHS, contact hypersensitivity; TNF-a, tumor necrosis factor, urocanic acid; UVB, UV radiation of wavelengths 280–320 nm.
thereby enabling cutaneous tumors to escape elimination [3]. An experimental model using contact hypersensitivity responses (CHS) to contact-sensitizing antigens is frequently used to study the impairment of cutaneous immunity by UV. Studies in humans have suggested that some individuals are more susceptible than others to the immunosuppressive properties of UVB [4,5]. More than 90% of patients with nonmelanoma skin cancer have impaired CHS responses to a sensitizing antigen following UVB exposure, or require higher concentrations of hapten for the same response [6,7]. Some of the mechanisms of UVB-induced immunosuppression that may determine susceptibility to the development of BCC have recently been investigated. Studies of cutaneous concentrations of UV-induced cisurocanic acid (UCA) [8], tumor necrosis factor (TNF)
1046-2023/02/$ - see front matter Ó 2002 Elsevier Science (USA). All rights reserved. PII: S 1 0 4 6 - 2 0 2 3 ( 0 2 ) 0 0 2 1 3 - X
M.A. Grimbaldeston et al. / Methods 28 (2002) 90–96
polymorphisms [9], and repair rates of UV-induced cyclobutane pyrimidine dimers and 6-4 photoproducts [10,11] have all failed to associate high levels of the immunosuppressive mediators cis-UCA and TNF with BCC development. Dermal mast cells are often localized at highest density in the papillary dermis, adjacent to blood and lymphatic vessels, and in close proximity to peripheral nerves [12]. They secrete a number of stored mediators including TNF-a, interleukin (IL)-4, and the biogenic amine histamine [13–16]. In murine studies, we have shown that mast cell-derived histamine is an important mediator of systemic immunomodulation caused by UVB [17]. Furthermore, if the sensitizing antigen is applied to a nonirradiated skin site, the extent of susceptibility to UVB-induced systemic suppression of CHS responses is determined by the prevalence of histamine-containing dermal mast cells at the irradiated site [13]. This correlation was observed in a number of different murine strains. However, the functional link was demonstrated by reconstituting dorsal skin of mast cell-depleted (W f =W f ) mice with bone marrow-derived mast cell precursors from wild-type littermates and thereby rendering them susceptible to UVB-induced immunomodulation [13]. Studies of dermal mast cell prevalence in BCC patients and control subjects from two latitudinally distinct countries, Denmark and South Australia, are discussed in this article. We have reported that patients with a history of BCC have a higher dermal mast cell prevalence in non-sun-exposed buttock skin [18]. The methodology of quantifying dermal mast cell prevalence in humans is also addressed, with discussion of previous and current techniques of quantification. The association between high dermal mast cell prevalence and BCC development provides support for the concept that mast cells may be important to UVB-induced systemic immunosuppression in humans and that suppression of immunity by UVB is essential for skin cancer development.
2. Description of methods 2.1. Subjects Patients with a history of histologically confirmed BCC and control subjects of similar sex, age, and skin phototype from two latitudinally distinct populations, one in Denmark and the other in South Australia, participated in the study. Volunteers consisted of 21 Danish and 26 Adelaide BCC patients together with 20 Danish and 25 Adelaide control subjects. Participants were between 42 and 89 years of age with a mean age of 69 years for both studies. All were of skin types I, II, and III according to the phototype classification system proposed by Fitzpatrick [19]. Recruitment of study subjects occurred at Gentofte Hospital, Hellerup, Flin-
91
ders Medical Centre, Adelaide, and the Repatriation General Hospital, Adelaide. None of the control subjects had been treated for a skin malignancy. BCC patients had no prior history of any other type of skin tumor such as squamous cell carcinoma or melanoma. Exposure to arsenicals, ionizing radiation, artificial sunlamps, or immunosuppressive drugs, or sunlight to the buttock was the criterion for exclusion. In all of the BCC patients, carcinomas occurred at sun-exposed skin sites (face, scalp, front of chest, back, arms, and legs). Thirty-nine of the forty-seven BCC patients had more than one BCC and of these 39 patients, 33 had lesions localized in two or more different anatomical sites. In the remaining 6 patients, BCCs were localized at the same anatomical site, such as the face or scalp, but separated in location or time of appearance. Six men and five women had a BCC on the leg but all 11 patients also had BCCs at skin sites that received more sun exposure such as the face and the chest. Three healthy female volunteers were recruited for the determination of the interbiopsy variability from a single donor. Two of the volunteers were of skin phototype II, and the third of skin phototype III. Their ages ranged from 25 to 45 years. The South Australian and Danish studies were approved and performed according to the guidelines of the Ethics Committee at the Gentofte Hospital, Hellerup, Denmark (KA95222) and the Clinical Investigations Committee at Flinders Medical Centre, Adelaide, South Australia. 2.2. Preparation of tissue samples Each participant in the study was administered local anesthetic (2% lidocaine with adrenaline) prior to having a 4-mm punch biopsy of skin from the upper buttock. The biopsies were fixed in 10% buffered formalin and paraffin embedded, ensuring a cross-sectional orientation of the cutaneous layers. All biopsies were coded for quantification of dermal mast cell prevalence. For the determination of intra-individual variance in mast cell prevalence in non-sun-exposed buttock skin, multiple biopsies (n ¼ 3–6) were taken from three volunteers. The buttock skin sites sampled were the upper left and right buttocks and lower left quadrant. Two and six months after initial sampling, another buttock biopsy was taken from each volunteer to determine variation in mast cell prevalence over time. 2.3. Immunohistochemical staining for histamine-containing dermal mast cells Sections (4 lm) of biopsy were cut, dewaxed, rehydrated with graded concentrations of alcohol, and immersed in 0.005 M Tris–HCl-buffered saline containing 0.1% calcium chloride and 0.1% trypsin powder (Difco,
92
M.A. Grimbaldeston et al. / Methods 28 (2002) 90–96
Becton Dickinson, Sparks, MD). Sections were digested with trypsin at 37 °C for 20 min followed by a wash in running water. After incubation at room temperature (RT) with 10% normal sheep serum (Silenus, Hawthorn, Australia) for 15 min, sections were immunohistochemically stained with a 1/1200 dilution of rabbit antihistamine antibody (Chemicon, Temecula, CA) for 16 h at 4 °C. Endogenous peroxide acitivity was quenched by immersion into 0.3% hydrogen peroxide in phosphatebuffered saline for 10 min at RT. After incubation with a 1/250 dilution of biotin-conjugated sheep anti-rabbit immunoglobulin (Silenus) for 45 min at RT and 1/500 dilution of peroxidase-conjugated streptavidin (Dako, Carpinteria, CA) for 30 min at RT, the reaction product was developed with the chromogen, 3; 30 -diaminobenzidine tetrahydrochloride (DAB) (Sigma Chemical, St. Louis, MO). Sections were counterstained with a weak Lillie-Mayers hematoxylin solution (1/1 with distilled water). Control sections were incubated with diluted normal rabbit serum (1/1200) or no primary antibody. Figs. 1A and B show histamine-containing dermal mast cells. 2.4. Quantification of histamine-containing dermal mast cells A video image analysis system (Video Pro 32, Leading Edge P/L, Adelaide, South Australia) was used to
quantify histamine-staining mast cells. The system consists of an Olympus BX40 microscope with a 2:5 camera eyepiece and a continuous interference filter monochromator for the enhancement of hematoxylin (blue) and DAB (brown) contrast. Images were captured with a Panasonic color digital CCD video camera (Model GP-KR222, automatic gain disabled) interfacing an Intel 700-MHz MMX(TM) processor-based personal computer with a VCG multimedia 32-bit color video digitizer card and Video Pro 32, Version 4, image analysis software. Depending on thickness of dermis (papillary + reticular layer), images of 68–490 consecutive fixed fields (image window, 512 512 pixels) were captured in each section and analyzed. Two consecutive sections from each biopsy were assessed for dermal mast cell prevalence and the mean was obtained. Images of fields were captured horizontally parallel to the epidermis followed by the papillary and reticular layers of the dermis using a 40 microscope objective. The hypodermal layer was excluded from measurement. Mast cells per fixed field of 140-lm width were counted (but expressed per mm2 ); the coefficient of variation for measurement of mean dermal area for 68–490 fields was <19%. A positive control section of buttock skin was included in each staining run and demonstrated minimal difference in mast cell staining intensity between runs. All measurement readings were standardized using a brightfield
Fig. 1. Histologic sections of histamine-containing dermal mast cells in buttock skin. (A) Magnification: 40 objective with 2:5 camera eyepiece. Arrows: mast cells. (B) Magnification: 100 objective with 2:5 camera eyepiece.
M.A. Grimbaldeston et al. / Methods 28 (2002) 90–96
93
image of the background illumination. This allowed correction of nonuniform illumination over a field of view.
3. Results and discussion 3.1. Interbiopsy variability Until recently quantification of dermal mast cell prevalence was performed using eyepiece graticules or graded on a semiquantitative histological scale [20–22]. Most studies have evaluated mast cell prevalence by randomly choosing a limited number of fields in the dermis [23,24] rather than quantifying mast cell density according to the total dermal area of the biopsy. As mast cells in the skin are not randomly distributed but localized close to nerves, blood vessels, and at high densities in the papillary dermis, random selection of a few high-power microscope fields may not provide an accurate assessment. For this reason we elected to evaluate dermal mast cell prevalence by quantifying the total number of mast cells according to the total dermal area (expressed as mast cells/mm2 ). Fields were captured consecutively in two sections of biopsy and the mean mast cell prevalence was obtained. In a study on the interbiopsy variability in dermal mast cell numbers from a single donor, we observed that there were minor differences in prevalence in anatomically comparable nonsun-exposed skin (Table 1) [18]. For two of the donors the dermal mast cell prevalence in multiple biopsies of buttock skin did not vary significantly. Three of the initial biopsies from donor 2 had a similar mast cell prevalence; however, a fourth biopsy had a higher mast cell prevalence caused by a perifollicular mast cell infiltrate associated with low-grade inflammation of a hair follicle. This was not observed in any other biopsy investigated. The mast cell prevalences quantified in bi-
Fig. 2. Mast cells per square millimeter of dermal area in biopsies of non-sun-exposed skin from South Australian and Danish BCC patients and control subjects. Each point represents the mean mast cell number expressed per fixed field area for 346–964 measurements. The lines represent the median for each group. BCC, basal cell carcinoma (Adelaide patients n ¼ 26; Danish patients n ¼ 21); control (Adelaide subjects n ¼ 25; Danish subjects n ¼ 20). For the Danish study, median values are 38 mast cells/mm2 (range ¼ 18–65) and 26 mast cells/ mm2 (range ¼ 11–43); for the Adelaide study median values are 35 mast cells/mm2 (range ¼ 18–77) and 24 mast cells/mm2 (range ¼ 9–48) for BCC patients and control subjects, respectively. Dermal mast cell prevalence was significantly higher in BCC than in the control group for both the Adelaide (P ¼ 0:0003, Mann–Whitney U) and Danish (P ¼ 0:01, Mann–Whitney U) studies. The data for the Danish BCC patients and control subjects have been published previously [18].
opsies taken 2 and 6 months after the initial biopsies were similar to those originally observed and demonstrated few changes in mast cell prevalence over this period. In contrast to our findings, another group reported variation in mast cell prevalence in different sections of the same biopsy sample and variance between adjacent biopsies of non-sun-exposed upper arm
Table 1 Dermal mast cell prevalence interbiopsy variability Subject
Sex
Age (years)
Skin phototype
Mast cell prevalence (per mm2 )a Initial biopsy
2 months later
6 months later
1
F
32
II
26
29
2
F
45
III
43
43
3
F
25
II
23 31 27 24 43 37 43 53a 26 26 23
ND
ND
a
Source. Previously published data [18]. Perifollicular mast cell infiltrate associated with low-grade inflammation of a hair follicle. ND, not done.
94
M.A. Grimbaldeston et al. / Methods 28 (2002) 90–96
skin [22]. In this aforementioned study [22] the entire dermal area was evaluated for mast cell prevalence using an eyepiece graticule, which may introduce error by failing to accurately quantify dermal area in fields containing features other than dermis, such as regions of the epidermal–dermal interface. With the advent of computer-generated image analysis such difficulties of quantifying dermal area are overcome as software programs enable accurate editing. Furthermore, Eady et al. [22] evaluated 1- to 1.5-lm-thick sections compared with 4-lm-thick sections in our study. The thicker sections of biopsy used by us may reduce multiple counting of mast cell fragments. 3.2. Dermal mast cell prevalence in BCC patients Patients with a history of BCC have significantly higher dermal mast cell prevalence than age-matched control subjects [18]. Fig. 2 shows the dermal mast cell
prevalence of BCC patients and control subjects from two populations, South Australia and Denmark. A similar association between high dermal mast cell prevalence and BCC development exists in both the South Australian and Danish studies of BCC patients. No significant linear correlation was observed between dermal mast cell prevalence and age of BCC patients and control subjects in the two populations (Figs. 3A and B). The distribution of data in Fig. 2 suggests that dermal mast cell prevalence is only one susceptibility factor for BCC development. BCC patients with lower dermal mast cell prevalence may have other unknown predisposing genetic factors or etiologic reasons associated with the development of BCC, such as loss of heterozygosity in the PTCH tumor suppressor gene [25], exposure to a larger UV dose, or more exposures to UV. Although increased numbers of mast cells have been reported in perilesional skin [26], there is evidence to suggest that the presence of BCCs does not increase
Fig. 3. Mast cells per square millimeter of dermal area of non-sun-exposed buttock skin versus age (years). (A) Linear regression analysis r2 ¼ 0:003 (p ¼ 0:81) for Danish BCC patients (j); r2 ¼ 0:003, (p ¼ 0:93) for Danish control subjects (). The data for the Danish BCC patients and control subjects have been published previously (12). (B) Linear regression analysis: r2 ¼ 0:03 (P ¼ 0:4) for South Australian BCC patients (d); r2 ¼ 0:265 (P ¼ 0:1) for South Australian control subjects ().
M.A. Grimbaldeston et al. / Methods 28 (2002) 90–96
mast cell numbers at distant skin sites [27]. It may be proposed that UV irradiation of the skin plays an important role not only in the induction of BCCs but also on dermal mast cell prevalence. Although there are contradictory reports of increases in the number of dermal mast cells in photodamaged skin [20,28], in our study of BCC patients non-sun-exposed buttock skin was investigated to avoid the potential for superimposed effects on mast cell prevalence caused by sun exposure. We propose that the prevalence of mast cells in buttock skin reflects the prevalence of mast cells in the skin at a time in life critical for the initiation of BCCs, that is, childhood [1,2]. We hypothesize that mast cells function in humans, as in mice, by initiating immunosuppression, and thus encourage BCC development. Studies in mice provide evidence that UVB indirectly activates dermal mast cells to release mediators, such as histamine, by inducing isomerization of trans-UCA to cis-UCA which in turn acts on the sensory nerves of the skin, stimulating neuropeptide release and mast cell degranulation [29]. In contrast to this more immediate effect of cis-UCA, UVB irradiation also stimulates keratinocytes to produce nerve growth factor (expressed maximally 8–12 h after UV exposure) which sustains the activation of sensory cfibers for neuropeptide-induced mast cell activation [30]. Future studies investigating dermal mast cell prevalence and extent of susceptibility to UVB-induced impairment of CHS responses in BCC patients and control subjects may provide the important evidence that, in humans as in murine strains, mast cells are functionally important in the mechanisms of immunomodulation caused by UVB. Ultimately, identification of the mechanism(s) that signals immune suppression may provide possibilities for intervention and have implications for development of desirable sunscreen properties.
4. Concluding remarks Few studies have evaluated dermal mast cell prevalence in humans. In these investigations, differences in methodology have contributed to the contradictory findings of mast cell density in skin. Variations in methodology that may limit accurate quantification include (a) size and depth of biopsy (sometimes not extending to the subcutaneous fatty tissue); (b) thickness of sections; (c) number of consecutive sections investigated; and (d) method of quantification (semiquantitative analysis or according to total dermal area employing computer-generated image analysis). Though labor intensive, meticulous counting of mast cells according to dermal area in sequential fields throughout the dermis limits inaccuracy that may arise from the selection of a few random fields. This technique has enabled us to detect heterogeneity of dermal mast cell
95
prevalence in buttock skin between individuals and has provided evidence of an association between high dermal mast cell prevalence and BCC development. In conclusion, in our investigations of the mechanisms of UVB-induced immunosuppression, we have shown that high dermal mast cell prevalence, a factor that determines susceptibility to UVB-induced immunomodulation in mouse strains, is associated with the development of BCC in humans. Thus, we hypothesize that mast cells function in humans, as in mice, by initiating immune suppression and, thereby, allowing a permissive environment for BCC development.
Acknowledgments The authors thank Dr. Gillian Marshman, Dr. Karen Koh, and Dr. Dorota Long for taking the skin biopsies. The work was supported by the National Health and Medical Research Council of Australia, the Anti-Cancer Foundation of South Australia, and the Flinders Medical Centre Foundation (to P.H.H. and J.J.F.J.). M.A.G. is supported by the South Australian Joyner Scholarship for Medicine.
References [1] A. Kricker, B.K. Armstrong, D.R. English, P.J. Heenan, Int. J. Cancer 60 (1995) 489–494. [2] B.K. Armstrong, A. Kricker, J. Photochem. Photobiol. B. 63 (2001) 8–18. [3] M.L. Kripke, Adv. Cancer Res. 34 (1981) 69–106. [4] K.D. Cooper, Photochem. Photobiol. 63 (1996) 400–406. [5] L. Skov, H. Hansen, J.N. Barker, J.C. Simon, O. Baadsgaard, Clin. Exp. Immunol. 107 (1997) 585–588. [6] T. Yoshikawa, V. Rae, W. Bruins-Slot, J.W. Van den Berg, J.R. Taylor, J.W. Streilein, J. Invest. Dermatol. 95 (1990) 530–536. [7] J.W. Streilein, J.R. Taylor, V. Vincek, I. Kurimoto, J. Richardson, C. Tie, J.P. Medema, C. Golomb, J. Invest. Dermatol. 103 (Suppl) (1994) S107–S111. [8] F. De Fine Olivarius, J. Lock-Andesen, F. Gronhoj Larsen, H.C. Wulf, J. Crosby, M. Norval, Br. J. Dermatol. 138 (1998) 986–992. [9] A.H. Hajeer, J.T. Lear, W.E.R. Ollier, M. Naves, J. Worthington, D.A. Bell, A.G. Smith, W.P. Bowers, P.W. Jones, R.C. Strange, A.A. Fryer, Br. J. Dermatol. 142 (2000) 441–445. [10] G. Xu, E. Snellman, C.T. Jansen, K. Hemminki, J. Invest. Dermatol. 115 (2000) 95–99. [11] J. Kibitel, V. Hejmadi, L. Alas, A. O’Connor, B.M. Sutherland, D. Yarosh, Photochem. Photobiol. 67 (1998) 541–546. [12] D.D. Metcalfe, D. Baram, Y.A. Mekori, Physiol. Rev. 77 (1997) 1033–1079. [13] P.H. Hart, M.A. Grimbaldeston, G.J. Swift, A. Jaksic, F.P. Noonan, J.J. Finlay-Jones, J. Exp. Med. 187 (1998) 2045–2053. [14] S.J. Galli, M. Maurer, C.S. Lantz, Curr. Opin. Immunol. 11 (1999) 53–59. [15] S.J. Galli, Curr. Opin. Hematol. 7 (2000) 32–39. [16] Y.A. Mekori, D.D. Metcalfe, Immunol. Rev. 173 (2000) 131–140. [17] P.H. Hart, A. Jaksic, G. Swift, M. Norval, A.A. El-Ghorr, J.J. Finlay-Jones, Immunology 91 (1997) 601–608.
96
M.A. Grimbaldeston et al. / Methods 28 (2002) 90–96
[18] M.A. Grimbaldeston, L. Skov, O. Baadsgaard, B.G. Skov, G. Marshman, J.J. Finlay-Jones, P.H. Hart, J. Invest. Dermatol. 115 (2000) 317–320. [19] T.B. Fitzpatrick, Arch. Dermatol. 124 (1988) 869–871. [20] J. Bhawan, W. Andersen, J. Lee, R. Labadie, G. Solares, J. Cutan. Pathol. 22 (1995) 154–159. [21] J. Bhawan, C.H. Oh, R. Lew, K.S. Nehal, R.R. Labadie, A. Tsay, B.A. Gilchrest, Am. J. Dermatopathol. 14 (1992) 224–230. [22] R.A.J. Eady, T. Cowen, T.F. Marshall, V. Plummer, M.W. Greaves, Br. J. Dermatol. 100 (1979) 623–633. [23] B. Hellstr€ om, H.J. Holmgren, Acta Anatom. 10 (1950) 82–90. [24] T. Hase, K. Shinta, T. Murase, I. Tokimitsu, M. Hattori, R. Takimoto, R. Tsuboi, H. Ogawa, Br. J. Dermatol. 142 (2000) 267–273.
[25] M.R. Gailani, M. St ahle-B€ackdahl, D.J. Leffell, Nat. Genet. 14 (1996) 78–81. [26] T.R. Humphreys, M.R. Monteiro, M.D. Murphy, Dermatol. Surg. 26 (2000) 200–204. [27] N.F. Eaglstein, A.D. Hernandey, J.E. Allen, J. Dermatol. Surg. Oncol. 8 (1992) 943–947. [28] R.M. Lavker, A.M. Kligman, J. Invest. Dermatol. 90 (1988) 325– 330. [29] Z. Khalil, S.L. Townley, M.A. Grimbaldeston, J.J. Finlay-Jones, P.H. Hart, J. Invest. Dermatol. 117 (2001) 886–891. [30] S.L. Townley, M.A. Grimbaldeston, I. Ferguson, R.A. Rush, S.H. Zhang, X.F. Zhou, J.M. Conner, J.J. Finlay-Jones, P.H. Hart, J. Invest. Dermatol. 118 (2002) 396–401.
Increased dermal mast cell prevalence and susceptibility to development of basal cell carcinoma in humans Michele A. Grimbaldeston,a,* Lone Skov,b John J. Finlay-Jones,a and Prue H. Harta a
Department of Microbiology and Infectious Diseases, School of Medicine, Flinders University, Adelaide 5001, Australia b Department of Dermatology, Gentofte Hospital, University of Copenhagen, Copenhagen, Denmark Accepted 19 June 2002
Abstract Exposure to ultraviolet B (UVB) radiation (280–320 nm) is the primary etiologic factor associated with the development of basal cell carcinoma (BCC). The outgrowth of these keratinocyte-derived skin lesions is enhanced by the ability of UVB to impair an immune response that would otherwise eliminate them. Studies in a range of inbred mouse strains as well as mast cell-depleted mice reconstituted with mast cell precursors support a functional link between histamine-staining dermal mast cells and the extent of susceptibility to UVB-induced systemic immunomodulation. Humans, like mouse strains, display variations in dermal mast cell prevalence. In a study of Danish and South Australian BCC patients and control subjects, one 4-mm punch biopsy of non-sunexposed buttock skin was sampled from each participant. This skin site was investigated to avoid any changes in mast cell prevalence caused by sun exposure. Two sections (4 lm) per biopsy were immunohistochemically stained for detection of histamine-containing dermal mast cells. Computer-generated image analysis evaluated dermal mast cell prevalence in both sections by quantifying the total number of mast cells according to the total dermal area (expressed as mast cells per square millimeter). This technique enabled us to detect heterogeneity of dermal mast cell prevalence in buttock skin between individuals and provided evidence of an association between high dermal mast cell prevalence and BCC development in two diverse populations. We hypothesize that mast cells function in humans, as in mouse strains, by initiating immunosuppression following UV irradiation and, thereby, allowing a permissive environment for the development of BCC. Thus, a high dermal mast cell prevalence as demonstrable in buttock skin is a significant predisposing factor for development of BCC in humans. Ó 2002 Elsevier Science (USA). All rights reserved. Keywords: Ultraviolet immunosuppression; Histamine; cis-Urocanic acid; Skin cancer; Computer-generated image analysis
1. Introduction Many genetic and environmental factors contribute to the pathogenesis of skin cancer. Intermittent intense sun exposure, particularly to ultraviolet B (UV radiation of 280–320 nm), and a history of sunburn in childhood have been described as significant risk factors for the development of basal cell carcinoma (BCC)1 [1,2]. In addition to the initiation of skin neoplasms, UV irradiation also alters the cell-mediated immune response,
*
Corresponding author. Fax: +61-8-8204-4733. E-mail address: michele.grimbaldeston@flinders.edu.au (M.A. Grimbaldeston). 1 Abbreviations used: BCC, basal cell carcinoma; CHS, contact hypersensitivity; TNF-a, tumor necrosis factor, urocanic acid; UVB, UV radiation of wavelengths 280–320 nm.
thereby enabling cutaneous tumors to escape elimination [3]. An experimental model using contact hypersensitivity responses (CHS) to contact-sensitizing antigens is frequently used to study the impairment of cutaneous immunity by UV. Studies in humans have suggested that some individuals are more susceptible than others to the immunosuppressive properties of UVB [4,5]. More than 90% of patients with nonmelanoma skin cancer have impaired CHS responses to a sensitizing antigen following UVB exposure, or require higher concentrations of hapten for the same response [6,7]. Some of the mechanisms of UVB-induced immunosuppression that may determine susceptibility to the development of BCC have recently been investigated. Studies of cutaneous concentrations of UV-induced cisurocanic acid (UCA) [8], tumor necrosis factor (TNF)
1046-2023/02/$ - see front matter Ó 2002 Elsevier Science (USA). All rights reserved. PII: S 1 0 4 6 - 2 0 2 3 ( 0 2 ) 0 0 2 1 3 - X
M.A. Grimbaldeston et al. / Methods 28 (2002) 90–96
polymorphisms [9], and repair rates of UV-induced cyclobutane pyrimidine dimers and 6-4 photoproducts [10,11] have all failed to associate high levels of the immunosuppressive mediators cis-UCA and TNF with BCC development. Dermal mast cells are often localized at highest density in the papillary dermis, adjacent to blood and lymphatic vessels, and in close proximity to peripheral nerves [12]. They secrete a number of stored mediators including TNF-a, interleukin (IL)-4, and the biogenic amine histamine [13–16]. In murine studies, we have shown that mast cell-derived histamine is an important mediator of systemic immunomodulation caused by UVB [17]. Furthermore, if the sensitizing antigen is applied to a nonirradiated skin site, the extent of susceptibility to UVB-induced systemic suppression of CHS responses is determined by the prevalence of histamine-containing dermal mast cells at the irradiated site [13]. This correlation was observed in a number of different murine strains. However, the functional link was demonstrated by reconstituting dorsal skin of mast cell-depleted (W f =W f ) mice with bone marrow-derived mast cell precursors from wild-type littermates and thereby rendering them susceptible to UVB-induced immunomodulation [13]. Studies of dermal mast cell prevalence in BCC patients and control subjects from two latitudinally distinct countries, Denmark and South Australia, are discussed in this article. We have reported that patients with a history of BCC have a higher dermal mast cell prevalence in non-sun-exposed buttock skin [18]. The methodology of quantifying dermal mast cell prevalence in humans is also addressed, with discussion of previous and current techniques of quantification. The association between high dermal mast cell prevalence and BCC development provides support for the concept that mast cells may be important to UVB-induced systemic immunosuppression in humans and that suppression of immunity by UVB is essential for skin cancer development.
2. Description of methods 2.1. Subjects Patients with a history of histologically confirmed BCC and control subjects of similar sex, age, and skin phototype from two latitudinally distinct populations, one in Denmark and the other in South Australia, participated in the study. Volunteers consisted of 21 Danish and 26 Adelaide BCC patients together with 20 Danish and 25 Adelaide control subjects. Participants were between 42 and 89 years of age with a mean age of 69 years for both studies. All were of skin types I, II, and III according to the phototype classification system proposed by Fitzpatrick [19]. Recruitment of study subjects occurred at Gentofte Hospital, Hellerup, Flin-
91
ders Medical Centre, Adelaide, and the Repatriation General Hospital, Adelaide. None of the control subjects had been treated for a skin malignancy. BCC patients had no prior history of any other type of skin tumor such as squamous cell carcinoma or melanoma. Exposure to arsenicals, ionizing radiation, artificial sunlamps, or immunosuppressive drugs, or sunlight to the buttock was the criterion for exclusion. In all of the BCC patients, carcinomas occurred at sun-exposed skin sites (face, scalp, front of chest, back, arms, and legs). Thirty-nine of the forty-seven BCC patients had more than one BCC and of these 39 patients, 33 had lesions localized in two or more different anatomical sites. In the remaining 6 patients, BCCs were localized at the same anatomical site, such as the face or scalp, but separated in location or time of appearance. Six men and five women had a BCC on the leg but all 11 patients also had BCCs at skin sites that received more sun exposure such as the face and the chest. Three healthy female volunteers were recruited for the determination of the interbiopsy variability from a single donor. Two of the volunteers were of skin phototype II, and the third of skin phototype III. Their ages ranged from 25 to 45 years. The South Australian and Danish studies were approved and performed according to the guidelines of the Ethics Committee at the Gentofte Hospital, Hellerup, Denmark (KA95222) and the Clinical Investigations Committee at Flinders Medical Centre, Adelaide, South Australia. 2.2. Preparation of tissue samples Each participant in the study was administered local anesthetic (2% lidocaine with adrenaline) prior to having a 4-mm punch biopsy of skin from the upper buttock. The biopsies were fixed in 10% buffered formalin and paraffin embedded, ensuring a cross-sectional orientation of the cutaneous layers. All biopsies were coded for quantification of dermal mast cell prevalence. For the determination of intra-individual variance in mast cell prevalence in non-sun-exposed buttock skin, multiple biopsies (n ¼ 3–6) were taken from three volunteers. The buttock skin sites sampled were the upper left and right buttocks and lower left quadrant. Two and six months after initial sampling, another buttock biopsy was taken from each volunteer to determine variation in mast cell prevalence over time. 2.3. Immunohistochemical staining for histamine-containing dermal mast cells Sections (4 lm) of biopsy were cut, dewaxed, rehydrated with graded concentrations of alcohol, and immersed in 0.005 M Tris–HCl-buffered saline containing 0.1% calcium chloride and 0.1% trypsin powder (Difco,
92
M.A. Grimbaldeston et al. / Methods 28 (2002) 90–96
Becton Dickinson, Sparks, MD). Sections were digested with trypsin at 37 °C for 20 min followed by a wash in running water. After incubation at room temperature (RT) with 10% normal sheep serum (Silenus, Hawthorn, Australia) for 15 min, sections were immunohistochemically stained with a 1/1200 dilution of rabbit antihistamine antibody (Chemicon, Temecula, CA) for 16 h at 4 °C. Endogenous peroxide acitivity was quenched by immersion into 0.3% hydrogen peroxide in phosphatebuffered saline for 10 min at RT. After incubation with a 1/250 dilution of biotin-conjugated sheep anti-rabbit immunoglobulin (Silenus) for 45 min at RT and 1/500 dilution of peroxidase-conjugated streptavidin (Dako, Carpinteria, CA) for 30 min at RT, the reaction product was developed with the chromogen, 3; 30 -diaminobenzidine tetrahydrochloride (DAB) (Sigma Chemical, St. Louis, MO). Sections were counterstained with a weak Lillie-Mayers hematoxylin solution (1/1 with distilled water). Control sections were incubated with diluted normal rabbit serum (1/1200) or no primary antibody. Figs. 1A and B show histamine-containing dermal mast cells. 2.4. Quantification of histamine-containing dermal mast cells A video image analysis system (Video Pro 32, Leading Edge P/L, Adelaide, South Australia) was used to
quantify histamine-staining mast cells. The system consists of an Olympus BX40 microscope with a 2:5 camera eyepiece and a continuous interference filter monochromator for the enhancement of hematoxylin (blue) and DAB (brown) contrast. Images were captured with a Panasonic color digital CCD video camera (Model GP-KR222, automatic gain disabled) interfacing an Intel 700-MHz MMX(TM) processor-based personal computer with a VCG multimedia 32-bit color video digitizer card and Video Pro 32, Version 4, image analysis software. Depending on thickness of dermis (papillary + reticular layer), images of 68–490 consecutive fixed fields (image window, 512 512 pixels) were captured in each section and analyzed. Two consecutive sections from each biopsy were assessed for dermal mast cell prevalence and the mean was obtained. Images of fields were captured horizontally parallel to the epidermis followed by the papillary and reticular layers of the dermis using a 40 microscope objective. The hypodermal layer was excluded from measurement. Mast cells per fixed field of 140-lm width were counted (but expressed per mm2 ); the coefficient of variation for measurement of mean dermal area for 68–490 fields was <19%. A positive control section of buttock skin was included in each staining run and demonstrated minimal difference in mast cell staining intensity between runs. All measurement readings were standardized using a brightfield
Fig. 1. Histologic sections of histamine-containing dermal mast cells in buttock skin. (A) Magnification: 40 objective with 2:5 camera eyepiece. Arrows: mast cells. (B) Magnification: 100 objective with 2:5 camera eyepiece.
M.A. Grimbaldeston et al. / Methods 28 (2002) 90–96
93
image of the background illumination. This allowed correction of nonuniform illumination over a field of view.
3. Results and discussion 3.1. Interbiopsy variability Until recently quantification of dermal mast cell prevalence was performed using eyepiece graticules or graded on a semiquantitative histological scale [20–22]. Most studies have evaluated mast cell prevalence by randomly choosing a limited number of fields in the dermis [23,24] rather than quantifying mast cell density according to the total dermal area of the biopsy. As mast cells in the skin are not randomly distributed but localized close to nerves, blood vessels, and at high densities in the papillary dermis, random selection of a few high-power microscope fields may not provide an accurate assessment. For this reason we elected to evaluate dermal mast cell prevalence by quantifying the total number of mast cells according to the total dermal area (expressed as mast cells/mm2 ). Fields were captured consecutively in two sections of biopsy and the mean mast cell prevalence was obtained. In a study on the interbiopsy variability in dermal mast cell numbers from a single donor, we observed that there were minor differences in prevalence in anatomically comparable nonsun-exposed skin (Table 1) [18]. For two of the donors the dermal mast cell prevalence in multiple biopsies of buttock skin did not vary significantly. Three of the initial biopsies from donor 2 had a similar mast cell prevalence; however, a fourth biopsy had a higher mast cell prevalence caused by a perifollicular mast cell infiltrate associated with low-grade inflammation of a hair follicle. This was not observed in any other biopsy investigated. The mast cell prevalences quantified in bi-
Fig. 2. Mast cells per square millimeter of dermal area in biopsies of non-sun-exposed skin from South Australian and Danish BCC patients and control subjects. Each point represents the mean mast cell number expressed per fixed field area for 346–964 measurements. The lines represent the median for each group. BCC, basal cell carcinoma (Adelaide patients n ¼ 26; Danish patients n ¼ 21); control (Adelaide subjects n ¼ 25; Danish subjects n ¼ 20). For the Danish study, median values are 38 mast cells/mm2 (range ¼ 18–65) and 26 mast cells/ mm2 (range ¼ 11–43); for the Adelaide study median values are 35 mast cells/mm2 (range ¼ 18–77) and 24 mast cells/mm2 (range ¼ 9–48) for BCC patients and control subjects, respectively. Dermal mast cell prevalence was significantly higher in BCC than in the control group for both the Adelaide (P ¼ 0:0003, Mann–Whitney U) and Danish (P ¼ 0:01, Mann–Whitney U) studies. The data for the Danish BCC patients and control subjects have been published previously [18].
opsies taken 2 and 6 months after the initial biopsies were similar to those originally observed and demonstrated few changes in mast cell prevalence over this period. In contrast to our findings, another group reported variation in mast cell prevalence in different sections of the same biopsy sample and variance between adjacent biopsies of non-sun-exposed upper arm
Table 1 Dermal mast cell prevalence interbiopsy variability Subject
Sex
Age (years)
Skin phototype
Mast cell prevalence (per mm2 )a Initial biopsy
2 months later
6 months later
1
F
32
II
26
29
2
F
45
III
43
43
3
F
25
II
23 31 27 24 43 37 43 53a 26 26 23
ND
ND
a
Source. Previously published data [18]. Perifollicular mast cell infiltrate associated with low-grade inflammation of a hair follicle. ND, not done.
94
M.A. Grimbaldeston et al. / Methods 28 (2002) 90–96
skin [22]. In this aforementioned study [22] the entire dermal area was evaluated for mast cell prevalence using an eyepiece graticule, which may introduce error by failing to accurately quantify dermal area in fields containing features other than dermis, such as regions of the epidermal–dermal interface. With the advent of computer-generated image analysis such difficulties of quantifying dermal area are overcome as software programs enable accurate editing. Furthermore, Eady et al. [22] evaluated 1- to 1.5-lm-thick sections compared with 4-lm-thick sections in our study. The thicker sections of biopsy used by us may reduce multiple counting of mast cell fragments. 3.2. Dermal mast cell prevalence in BCC patients Patients with a history of BCC have significantly higher dermal mast cell prevalence than age-matched control subjects [18]. Fig. 2 shows the dermal mast cell
prevalence of BCC patients and control subjects from two populations, South Australia and Denmark. A similar association between high dermal mast cell prevalence and BCC development exists in both the South Australian and Danish studies of BCC patients. No significant linear correlation was observed between dermal mast cell prevalence and age of BCC patients and control subjects in the two populations (Figs. 3A and B). The distribution of data in Fig. 2 suggests that dermal mast cell prevalence is only one susceptibility factor for BCC development. BCC patients with lower dermal mast cell prevalence may have other unknown predisposing genetic factors or etiologic reasons associated with the development of BCC, such as loss of heterozygosity in the PTCH tumor suppressor gene [25], exposure to a larger UV dose, or more exposures to UV. Although increased numbers of mast cells have been reported in perilesional skin [26], there is evidence to suggest that the presence of BCCs does not increase
Fig. 3. Mast cells per square millimeter of dermal area of non-sun-exposed buttock skin versus age (years). (A) Linear regression analysis r2 ¼ 0:003 (p ¼ 0:81) for Danish BCC patients (j); r2 ¼ 0:003, (p ¼ 0:93) for Danish control subjects (). The data for the Danish BCC patients and control subjects have been published previously (12). (B) Linear regression analysis: r2 ¼ 0:03 (P ¼ 0:4) for South Australian BCC patients (d); r2 ¼ 0:265 (P ¼ 0:1) for South Australian control subjects ().
M.A. Grimbaldeston et al. / Methods 28 (2002) 90–96
mast cell numbers at distant skin sites [27]. It may be proposed that UV irradiation of the skin plays an important role not only in the induction of BCCs but also on dermal mast cell prevalence. Although there are contradictory reports of increases in the number of dermal mast cells in photodamaged skin [20,28], in our study of BCC patients non-sun-exposed buttock skin was investigated to avoid the potential for superimposed effects on mast cell prevalence caused by sun exposure. We propose that the prevalence of mast cells in buttock skin reflects the prevalence of mast cells in the skin at a time in life critical for the initiation of BCCs, that is, childhood [1,2]. We hypothesize that mast cells function in humans, as in mice, by initiating immunosuppression, and thus encourage BCC development. Studies in mice provide evidence that UVB indirectly activates dermal mast cells to release mediators, such as histamine, by inducing isomerization of trans-UCA to cis-UCA which in turn acts on the sensory nerves of the skin, stimulating neuropeptide release and mast cell degranulation [29]. In contrast to this more immediate effect of cis-UCA, UVB irradiation also stimulates keratinocytes to produce nerve growth factor (expressed maximally 8–12 h after UV exposure) which sustains the activation of sensory cfibers for neuropeptide-induced mast cell activation [30]. Future studies investigating dermal mast cell prevalence and extent of susceptibility to UVB-induced impairment of CHS responses in BCC patients and control subjects may provide the important evidence that, in humans as in murine strains, mast cells are functionally important in the mechanisms of immunomodulation caused by UVB. Ultimately, identification of the mechanism(s) that signals immune suppression may provide possibilities for intervention and have implications for development of desirable sunscreen properties.
4. Concluding remarks Few studies have evaluated dermal mast cell prevalence in humans. In these investigations, differences in methodology have contributed to the contradictory findings of mast cell density in skin. Variations in methodology that may limit accurate quantification include (a) size and depth of biopsy (sometimes not extending to the subcutaneous fatty tissue); (b) thickness of sections; (c) number of consecutive sections investigated; and (d) method of quantification (semiquantitative analysis or according to total dermal area employing computer-generated image analysis). Though labor intensive, meticulous counting of mast cells according to dermal area in sequential fields throughout the dermis limits inaccuracy that may arise from the selection of a few random fields. This technique has enabled us to detect heterogeneity of dermal mast cell
95
prevalence in buttock skin between individuals and has provided evidence of an association between high dermal mast cell prevalence and BCC development. In conclusion, in our investigations of the mechanisms of UVB-induced immunosuppression, we have shown that high dermal mast cell prevalence, a factor that determines susceptibility to UVB-induced immunomodulation in mouse strains, is associated with the development of BCC in humans. Thus, we hypothesize that mast cells function in humans, as in mice, by initiating immune suppression and, thereby, allowing a permissive environment for BCC development.
Acknowledgments The authors thank Dr. Gillian Marshman, Dr. Karen Koh, and Dr. Dorota Long for taking the skin biopsies. The work was supported by the National Health and Medical Research Council of Australia, the Anti-Cancer Foundation of South Australia, and the Flinders Medical Centre Foundation (to P.H.H. and J.J.F.J.). M.A.G. is supported by the South Australian Joyner Scholarship for Medicine.
References [1] A. Kricker, B.K. Armstrong, D.R. English, P.J. Heenan, Int. J. Cancer 60 (1995) 489–494. [2] B.K. Armstrong, A. Kricker, J. Photochem. Photobiol. B. 63 (2001) 8–18. [3] M.L. Kripke, Adv. Cancer Res. 34 (1981) 69–106. [4] K.D. Cooper, Photochem. Photobiol. 63 (1996) 400–406. [5] L. Skov, H. Hansen, J.N. Barker, J.C. Simon, O. Baadsgaard, Clin. Exp. Immunol. 107 (1997) 585–588. [6] T. Yoshikawa, V. Rae, W. Bruins-Slot, J.W. Van den Berg, J.R. Taylor, J.W. Streilein, J. Invest. Dermatol. 95 (1990) 530–536. [7] J.W. Streilein, J.R. Taylor, V. Vincek, I. Kurimoto, J. Richardson, C. Tie, J.P. Medema, C. Golomb, J. Invest. Dermatol. 103 (Suppl) (1994) S107–S111. [8] F. De Fine Olivarius, J. Lock-Andesen, F. Gronhoj Larsen, H.C. Wulf, J. Crosby, M. Norval, Br. J. Dermatol. 138 (1998) 986–992. [9] A.H. Hajeer, J.T. Lear, W.E.R. Ollier, M. Naves, J. Worthington, D.A. Bell, A.G. Smith, W.P. Bowers, P.W. Jones, R.C. Strange, A.A. Fryer, Br. J. Dermatol. 142 (2000) 441–445. [10] G. Xu, E. Snellman, C.T. Jansen, K. Hemminki, J. Invest. Dermatol. 115 (2000) 95–99. [11] J. Kibitel, V. Hejmadi, L. Alas, A. O’Connor, B.M. Sutherland, D. Yarosh, Photochem. Photobiol. 67 (1998) 541–546. [12] D.D. Metcalfe, D. Baram, Y.A. Mekori, Physiol. Rev. 77 (1997) 1033–1079. [13] P.H. Hart, M.A. Grimbaldeston, G.J. Swift, A. Jaksic, F.P. Noonan, J.J. Finlay-Jones, J. Exp. Med. 187 (1998) 2045–2053. [14] S.J. Galli, M. Maurer, C.S. Lantz, Curr. Opin. Immunol. 11 (1999) 53–59. [15] S.J. Galli, Curr. Opin. Hematol. 7 (2000) 32–39. [16] Y.A. Mekori, D.D. Metcalfe, Immunol. Rev. 173 (2000) 131–140. [17] P.H. Hart, A. Jaksic, G. Swift, M. Norval, A.A. El-Ghorr, J.J. Finlay-Jones, Immunology 91 (1997) 601–608.
96
M.A. Grimbaldeston et al. / Methods 28 (2002) 90–96
[18] M.A. Grimbaldeston, L. Skov, O. Baadsgaard, B.G. Skov, G. Marshman, J.J. Finlay-Jones, P.H. Hart, J. Invest. Dermatol. 115 (2000) 317–320. [19] T.B. Fitzpatrick, Arch. Dermatol. 124 (1988) 869–871. [20] J. Bhawan, W. Andersen, J. Lee, R. Labadie, G. Solares, J. Cutan. Pathol. 22 (1995) 154–159. [21] J. Bhawan, C.H. Oh, R. Lew, K.S. Nehal, R.R. Labadie, A. Tsay, B.A. Gilchrest, Am. J. Dermatopathol. 14 (1992) 224–230. [22] R.A.J. Eady, T. Cowen, T.F. Marshall, V. Plummer, M.W. Greaves, Br. J. Dermatol. 100 (1979) 623–633. [23] B. Hellstr€ om, H.J. Holmgren, Acta Anatom. 10 (1950) 82–90. [24] T. Hase, K. Shinta, T. Murase, I. Tokimitsu, M. Hattori, R. Takimoto, R. Tsuboi, H. Ogawa, Br. J. Dermatol. 142 (2000) 267–273.
[25] M.R. Gailani, M. St ahle-B€ackdahl, D.J. Leffell, Nat. Genet. 14 (1996) 78–81. [26] T.R. Humphreys, M.R. Monteiro, M.D. Murphy, Dermatol. Surg. 26 (2000) 200–204. [27] N.F. Eaglstein, A.D. Hernandey, J.E. Allen, J. Dermatol. Surg. Oncol. 8 (1992) 943–947. [28] R.M. Lavker, A.M. Kligman, J. Invest. Dermatol. 90 (1988) 325– 330. [29] Z. Khalil, S.L. Townley, M.A. Grimbaldeston, J.J. Finlay-Jones, P.H. Hart, J. Invest. Dermatol. 117 (2001) 886–891. [30] S.L. Townley, M.A. Grimbaldeston, I. Ferguson, R.A. Rush, S.H. Zhang, X.F. Zhou, J.M. Conner, J.J. Finlay-Jones, P.H. Hart, J. Invest. Dermatol. 118 (2002) 396–401.