In vivo characterization of the inflammatory infiltrate and apoptotic status in imiquimod-treated basal cell carcinoma
Version of Record online: 19 FEB 2009
© 2009 The International Society of Dermatology
International Journal of Dermatology
Volume 48, Issue 3, pages 312–321, March 2009
How to Cite
De Giorgi, V., Salvini, C., Chiarugi, A., Paglierani, M., Maio, V., Nicoletti, P., Santucci, M., Carli, P. and Massi, D. (2009), In vivo characterization of the inflammatory infiltrate and apoptotic status in imiquimod-treated basal cell carcinoma. International Journal of Dermatology, 48: 312–321. doi: 10.1111/j.1365-4632.2009.03916.x
- Issue online: 19 FEB 2009
- Version of Record online: 19 FEB 2009
Background Imiquimod use in the treatment of basal cell carcinoma (BCC) has proven to be successful in a large percentage of cases, inducing tumor regression; however, the exact cellular mechanism has not been fully clarified.
Aim To measure the morphological changes in the tumor microenvironment and the markers of apoptosis in skin biopsies from patients with BCC before and after imiquimod treatment.
Methods In this open label study, skin biopsies obtained from 11 patients with BCC were evaluated before and after imiquimod treatment for: (i) morphological changes in the tumor microenvironment, with specific emphasis on the immunophenotype of inflammatory cells around the tumor; and (ii) markers of apoptosis, including expression of death receptors.
Results Imiquimod treatment induced a significant increase in the mononuclear inflammatory response. In the majority of cases, the cellular infiltrate was predominantly composed of CD3+/CD4+ T cells, suggesting that the effector response is mediated by CD3+/CD4+ lymphocytes, with a minor cytotoxic and natural killer (NK) component. An increase in the cytotoxic CD3+/CD8+ T-cell population was also observed. Imiquimod treatment was associated with a marked increased in CD20+ B cells, and a less pronounced enhancement in cells of monocyte–macrophage origin (CD68+) surrounding, or within, the tumor. This finding indicates either that macrophages play a minor role in the imiquimod-induced response, or the recruitment of these cells is related to time and dose. Imiquimod treatment decreased CD1A+ Langerhans cells in the epidermis and increased the number of CD1A+ dendritic cells within the tumor aggregates. Imiquimod reduced Bcl-2 expression, but no difference was found in Bax, Fas/FasL, and p53 expression in BCC cells.
Conclusions Our results support the hypothesis that imiquimod activity in the treatment of BCC is partly a result of a pro-inflammatory action mediated by CD3+/CD4+ lymphoid cells and of a pro-apoptotic activity associated with decreased Bcl-2 expression.
Clinical studies have demonstrated that imiquimod 5% cream is a safe and effective treatment for cutaneous basal cell carcinoma (BCC).1–3 Current data suggest that imiquimod exerts a triple action in the treatment of BCC: (i) Toll-like receptor (TLR)-dependent induction of immune responses; (ii) adenosine receptor-associated augmentation of inflammation; and (iii) direct pro-apoptotic activity. It is well known that imiquimod exerts its tumoral effect through the binding of TLR-7 and TLR-8, which belong to a family of receptors that recognize constituents of microorganisms, such as lipopolysaccharide (LPS), bacterial DNA, or double-stranded RNA. The stimulation of TLRs triggers interaction with adaptor molecules, such as myeloid differentiation primary-response protein 88 (MyD88), mediating a complex intracellular signaling pathway which ultimately leads to the activation of nuclear factor-kappaB (NF-κB). NF-κB regulates the expression of downstream target genes that encode proinflammatory mediators, such as interleukin-1 (IL-1), tumor necrosis factor-α (TNF-α), interferon-α (IFN-α), and IFN-β,4 inducing a tumor-directed cellular immune response. Recent evidence has also indicated that imiquimod acts as an adenosine receptor antagonist at the A2A receptor subtype, thereby suppressing an important feedback mechanism of inflammation by the antagonism of an adenosine receptor-dependent increase of cyclic adenosine monophosphate (cAMP) and a concomitant receptor-independent inhibition of cAMP production.5 In agreement with this pathogenetic hypothesis, previous reports on imiquimod-treated BCC tissue specimens have shown unequivocally that tumor regression is associated with the development of a prominent inflammatory cell infiltrate. The relative contribution of CD8+ T lymphocytes, CD4+ T-helper lymphocytes, CD20 B lymphocytes, dendritic cells, and macrophages in vivo to the overall antitumoral effect remains to be clarified, however.
Recent experimental and clinical data have indicated that imiquimod exerts a direct proapoptotic activity in cancer cells both in vitro and in vivo.6,7 At least in part, apoptosis induction by imiquimod is mediated through the Bcl-2-dependent release of mitochondrial cytochrome c and subsequent activation of caspase-9.7,8 Whether this action is also related to membrane-bound death receptors, rather than involving caspase activation alone, is less clear. Indeed, controversial results have been published on Bcl-2, Fas and FasL immunohistochemical expression by BCC neoplastic cells on imiquimod treatment in vivo.7–11
The aims of this study were as follows: (i) to evaluate the effect of topical imiquimod 5% cream on the treatment of superficial and nodular BCCs; (ii) to investigate the morphological changes in the possible residual tumor and peritumoral microenvironment induced by imiquimod treatment; (iii) to characterize the immunophenotype of inflammatory cells around the tumor before and after treatment; and (iv) to investigate a series of parameters reflecting tumor apoptotic status and expression of death receptors (Fas and FasL) before and after treatment. Frequently, topical application of imiquimod is associated with a proinflammatory response. To avoid this confounding factor, the second biopsy after imiquimod treatment was obtained only after this inflammatory response was turned off, as demonstrated by macroscopic examination of the skin.
Materials and Methods
Patients enrolled in the study attended the Department of Dermatology, University of Florence, Italy. All patients older than 18 years with BCC on the head and neck region (excluding nose, mouth, ears, and eyes), trunk (excluding the anogenital area), or extremities, suitable for treatment by surgical excision, and consecutively observed at the Pigmented Lesion/Skin Cancer Clinic, Department of Dermatology, University of Florence, were eligible for the study. Patients with recurrent or previously treated tumors were excluded. The age, ethnic background, and medical history of the patients were carefully reviewed to identify those affected by nevoid BCC (Gorlin) syndrome. Prior to entering the study, all patients were informed concerning the procedures and gave their consent. At the first visit, the tumor size was measured and photographs of the lesion were taken. Before treatment, each lesion was submitted to a punch biopsy (4 mm in diameter) under local anesthesia (1 mL mepivacaine 2%). Biopsy specimens were formalin-fixed and paraffin-embedded. For conventional histopathologic examination, sections, 5 µm thick, were cut from each tissue block and stained with hematoxylin and eosin. Additional serial sections were cut to be processed for immunohistochemistry.
Topical treatment with imiquimod 5% cream
Treatment (imiquimod 5% cream; Aldara®, 3M Pharmaceuticals, Bracknell, Berkshire, UK) was administered according to recent in vivo studies: daily application of 5% imiquimod cream at night for at least 8 h five times per week until the occurrence of signs of erosion, when the lesion was surgically excised with 3 mm margins and sent for histopathologic examination.
Immunohistochemical analyses were performed on both the pretreatment biopsy and the post-treatment excision specimens. Paraffin sections were incubated with antibodies against the following: CD45 (Dako S.p.A., Milan, Italy), CD68 (clone PGM-1, Dako), CD20 (Dako), CD79a (Dako), CD3 (Cell Marque Corp., Hot Springs, AR, USA), CD4 (Dako), CD8 (Dako), CD56 (NCAM, monoclonal antibody, Clone1B6; Novocastra, Newcastle, UK), T-cell intracytoplas antigen-1 (TIA-1) (C-20, polyclonal antibody; Santa Cruz Biotechnologies, Santa Cruz, CA, USA), CD1A (Zymed, Freemont, CA, USA), Bcl-2 (Dako), Bax (Dako), Fas (Novocastra), FasL (Novocastra), and p53 (Dako). Immunohistochemistry was performed by the standard avidin–biotin–peroxidase complex (ABC) method (Lab Vision, Fremont, CA, USA) with diaminobenzidine as chromogen and hematoxylin as counterstain. Immunoreactivities were semiquantitatively and independently evaluated by two pathologists [−, negative; +, < 20% of the cells/high-power field (hpf); ++, 20–50% of the cells/hpf; +++, > 50% of the cells/hpf]. Discrepancies were resolved in a second revision at a multihead microscope.
The study sample included 11 patients with a total of 12 BCCs. Their clinicopathologic characteristics are summarized in Table 1. Treatment was administered for a total of 20 days for six BCCs, 15 days for three, and 10 days for the remaining three. Compliance with treatment was 100% and no adverse events occurred, except increased crusting and erythema that extended peripherally to the area of application. Seven cases showed areas of superficial central erosion. At the time of excisional surgery, all tumors showed a decrease in size compared with the initial greatest diameter measurement. In five cases, excisional surgery was performed immediately at the end of treatment, whereas, in the other seven cases, the excisional biopsy was performed 1–3 weeks after the end of treatment because of the need to wait for a decrease in inflammatory effect in order to identify the margin of the tumor for surgery. All of the patients showed progressive healing of the excisional surgical site. None of the patients showed any evidence of recurrence after a mean follow-up of 1 year.
|Case||Age (years)||Sex||Size (cm)||Site||IMQ therapy (days)||Excision time (weeks)||Histotype||Residual tumor|
All pretreatment biopsies showed aggregates of basaloid cells, surrounded by a mild inflammatory infiltrate, consistent with BCC. On imiquimod treatment, six excisional specimens showed significant residual tumor (Fig. 1), in two cases only minimal aggregates of tumor cells were detected, and in the remaining four cases no residual BCC was found. After imiquimod therapy, all excisional specimens showed a moderate (three cases) to marked (nine cases) inflammatory infiltrate, extending around the neoplastic aggregates or in a lichenoid arrangement in the superficial dermis (Fig. 2). A correlation with the time of excision demonstrated that the inflammatory infiltrate was greater in cases submitted to surgery immediately after treatment. The infiltrate was composed mainly of lymphohistiocytes, with scattered plasma cells, neutrophils, eosinophils, and melanophages. Five cases displayed areas of epidermal ulceration with scale-crusts; in one case, there were signs of superficial erosion. In four cases, the epidermis showed features of pseudoepitheliomatous hyperplasia.
Immunohistochemical analysis (Table 2)
|Case||Infiltrate density||CD45||CD3||CD20||CD79a||CD4 : CD8||CD68||CD56||Bcl-2 (% of tumor cells)||Bax (% of tumor cells)||Fas (% of tumor cells)||FasL (% of tumor cells)||p53 (% of tumor cells)|
|1||+++||+++||++||+||+||2 : 1||+||+ (rare cells)||40||80||–||–||50|
|2||+++||+++||+++||++||+||3 : 1||+++||+||30||90||–||+||90|
|3||+++||+++||+++||++||++||3 : 1||++||+||10||100||–||+||65|
|4||+++||+++||+++||+||+||2 : 1||++||30||80||–||+||60|
|5||+++||+++||+++||++||+||3 : 1||+||+ (rare cells)||30||80||–||+||60|
|6||+++||+++||+++||++||+||2 : 1||+++||–||ne||ne||ne||ne||ne|
|7||+++||+++||++||+||+||2 : 1||++||–||5||90||–||+||100|
|8||+++||+++||+++||++||+||3 : 1||++||+||20||80||–||+||40|
|9||+++||+++||+++||++||+||3 : 1||+++||+ (rare cells)||ne||ne||ne||ne||ne|
|10||++||++||++||+||+||3 : 1||+||+||30||100||–||+||40|
|11*||++||++||++||+||+||2 : 1||++||+||ne||ne||ne||ne||ne|
|12*||++||++||++||+||+||2.5 : 1||+||ne||ne||ne||ne||ne||ne|
Immunohistochemical analyses demonstrated that, in pretreatment biopsies, neoplastic cells were surrounded by a scant inflammatory infiltrate mainly composed of T lymphocytes (CD45+, CD3+) intermingled with scattered histiocytes (CD68+). B lymphocytes (CD20+) and CD79a+ plasma cells were rare to absent; most T lymphocytes expressed CD4 antigen, with a CD4 : CD8 ratio ranging from 4 : 1 to 6 : 1 (Fig. 3). Before treatment, lymphocytes expressing TIA-1 were virtually absent. CD56 staining was observed in neoplastic cells (with both membranous and cytoplasmic staining), but only rare to no inflammatory cells were CD56+. Immunostaining with CD1A showed a variable number of Langerhans cells within the epidermis, as well as scattered CD1A+ dendritic cells within the tumor aggregates and the peritumoral dermis.
On imiquimod treatment, the overall number of inflammatory cells (CD45+ lymphocytes and CD68+ macrophages) was higher in comparison with pretreatment biopsies. T lymphocytes (CD45+; CD3+) were still the predominant cell population, although there was a relative increase in the number of B lymphocytes (CD20+), some of which aggregated in clusters, in comparison with pretreatment biopsies (Fig. 4). In the majority of cases, we observed a relative increase in CD8+ cytotoxic/suppressor lymphocytes, but still CD4+ lymphocytes predominated (from CD4 : CD8 = 2 : 1 to CD4 : CD8 = 3 : 1). Only in one case, in which the excisional biopsy was performed 2 weeks after treatment, did we demonstrate a significant increase in CD8+ lymphocytes, with a CD4 : CD8 ratio around 1 : 1. TIA-1 staining was observed in scattered CD8+ cytotoxic lymphocytes. In the majority of cases, the number of natural killer (NK)-like CD56+ cells was low, but mildly increased in comparison with pretreatment specimens in one case. Overall, after treatment, the number of CD1A+ cells in the overlying epidermis, adnexal structures, and dermis was significantly reduced, whereas increased numbers of CD1A+ cells were present within the neoplastic aggregates (Figs 5 and 6).
Variable percentages (40–80%) of BCC cells from pretreatment biopsies were weakly positive for the antiapoptotic marker Bcl-2, whereas, in all post-treatment biopsies, decreased Bcl-2 staining was observed (Fig. 7). Inflammatory cells in a peritumoral location expressed high Bcl-2 levels. The expression of the proapoptotic marker Bax was high in the majority of neoplastic cells, as well as in inflammatory cells in pretreatment biopsies, with no significant changes on imiquimod treatment. p53 was variably expressed in neoplastic cells (40–90% of tumor cells) in pretreatment biopsies. In the majority of cases, the p53 levels remained unchanged on imiquimod treatment, but, in one case, p53 expression was increased. Neoplastic cells did not express Fas and weakly expressed FasL before treatment, with no differences between pre- and post-treatment biopsies (Fig. 8).
Imiquimod is a well-established treatment for BCC that has been proven to be successful in a large percentage of cases, inducing regression of the tumor; however, the precise mechanism operating at the cellular level remains to be clarified. We first determined whether the regression of BCCs could correlate with a specific profile of inflammatory cells in the dermis surrounding and infiltrating the residual tumor. Controversial data have been reported in the literature with regard to the role of CD8+ T lymphocytes, CD4+ T-helper lymphocytes, dendritic cells, and macrophages under these circumstances. In agreement with previous observations,8,9,11,12 in our study, imiquimod treatment was associated with a significant increase in the mononuclear inflammatory response. Although we demonstrated a relative increase in the cytotoxic CD3+/CD8+ T-cell population, as reported previously,12 in the majority of cases, the lymphocytic infiltrate was mainly composed of CD3+/CD4+ T cells, suggesting that the effector response is probably mediated by CD3+/CD4+ lymphocytes, with a minor cytotoxic and NK component. One important difference between our study and previous reports was that the excision was performed at different time intervals after the end of treatment, because we decided to wait for the macroscopic resolution of the local inflammatory reaction, which sometimes is rather extensive. It is possible that, in an initial phase, the cellular infiltrate is predominantly composed of inflammatory CD3+/CD4+ T cells, followed by activation of dendritic cells, associated with the inflammatory response to imiquimod, with a relatively lower tumor suppressor CD4+/CD8+ cell component. These data are in agreement with those obtained during the spontaneous regression of BCC, when elevated levels of IFN-γ, IL-2, TNF-β, and CD3δ are found, thus supporting a role for T-helper-1-type cytokines in the immune response against the tumor.13,14
It is worth noting that imiquimod represents a potent stimulus for CD20+ B cells, possibly in a more complex interaction with other immune cells, including plasmacytoid dendritic cells (PDCs).15–18 Following imiquimod treatment, PDCs are recruited and activated in BCC, producing IFN-αin vivo, and thus probably playing a major role in the responsiveness to imiquimod in humans.15 In addition, the number of PDCs correlates with the clinical outcome of the tumor after imiquimod treatment, suggesting that the antitumor effect of the drug is mediated, at least in part, by the recruitment of PDC-like cells to the skin.
In our study, the overall increase in the cells of monocyte–macrophage origin (CD68+) surrounding or within the tumor was lower than expected, suggesting that either macrophages play a minor role in the imiquimod-induced response, or the recruitment of these cells is strictly related to time and dose. In any case, given the complexity of macrophage-related biological activities, further studies aimed to clarify their functional state/activity are necessary to better understand their possible role in imiquimod-induced tumor clearance. An intriguing finding of our study was that, on treatment, CD1A Langerhans cells in the epidermis were significantly decreased, whereas increased numbers of CD1A dendritic cells were present within the tumor aggregates. The decrease in the epidermal Langerhans cells may reflect an imiquimod-induced migration to draining lymph nodes, as shown previously.16 Our observations indicate that more detailed investigation is required to identify and characterize the different subsets of epidermal, dermal, and tumor-associated dendritic cells on imiquimod treatment.
This study was supported by a grant from 3M Pharmaceuticals, Bracknell, Berkshire, UK.