STUDY Lymphomatoid Keratosis An Epidermotropic Type of Cutaneous Lymphoid Hyperplasia: Clinicopathological, Immunohistochemical, and Molecular Biological Study of 6 Cases Eiichi Arai, MD; Michio Shimizu, MD; Tetsuya Tsuchida, MD; Seiichi Izaki, MD; Fumihiro Ogawa, MD; Takanori Hirose, MD Objective: To provide evidence that lymphomatoid kera- immunoglobulin heavy chain gene and T-cell receptor tosis should be categorized as an epidermotropic sub- (TCR)-␤ and TCRγ genes was also investigated by the type of cutaneous lymphoid hyperplasia. polymerase chain reaction method. ϩ Design: Clinicopathological, immunohistochemical, and Results: Immunohistochemically, epidermotropic CD20 molecular biological studies of epidermotropic and der- and/or CD79aϩ cells were present. In the upper dermal mal bandlike infiltrates of lymphocytes without ne- lymphocytic infiltrates, the CD3ϩ/CD79aϩ cell ratio ranged crotic keratinocytes, Civatte bodies, or Max-Joseph spaces from 5:5 to 8:2. The CD4ϩ/CD8ϩ cell ratio was within and solar lentigo or seborrheic keratosis adjacent to the normal limits. Rearrangements of the TCRγ gene were lesion, but with epidermal hyperplastic change (clini- demonstrated in 2 cases and of the TCRβ gene in 1 case. cally scaly plaque) in cases of lymphomatoid keratosis. Conventional histopathologic study as well as immuno- Conclusions: Our results indicate that lymphomatoid histochemical examinations for CD1a, CD3, CD4, CD8, keratosis is a clinically benign keratotic lesion but his- CD20, and CD79a and S100 protein and genotypic ex- tologically malignant, simulating mycosis fungoides. aminations were performed. Immunohistochemical findings showed a reaction pat- tern in all cases, but genotypical examination showed Setting: University departments comprising 2 sections some clonal dermatoses. Therefore, “lymphomatoid” of dermatology and 1 section of pathology. keratosis should be classed as a pseudolymphoma, namely, a subtype of cutaneous lymphoid hyperplasia Main Outcome Measures: Ratio of T to B cells and with epidermotropism. of CD4ϩ to CD8ϩ cells, and the phenotype of epidermo- tropic cells were evaluated. Gene rearrangement of the Arch Dermatol. 2007;143:53-59 HE TERM LYMPHOMATOID is noteworthy that the term pseudolym- keratosis was originally pro- phoma was originally adopted to indicate posed by Kossard1 and was a lesion clinically or histologically simu- thought to be a possible lating malignant lymphoma. Therefore, the variant of benign lichen- term lymphomatoid is considered to be Toid keratosis showing lymphomatoid fea- equivalent to so-called pseudolymphoma. tures. He considered that the case re- Epidermotropism is defined as an affin- ported by Evans et al2 as a unilesional ity of tumor cells for the epidermis. Vari- mycosis fungoides (MF) should be termed ous conditions display lymphocytic infil- lymphomatoid keratosis. Clinically, the case trates with tropism for the epidermis. reported by Evans et al2 was character- Epidermotropism was not included in the glossary of the fifth edition of Histopathol- ized by an asymptomatic scaly plaque, but Author Affiliations: ogy of the Skin by Lever and Schaumburg- Departments of Pathology the histological features were those seen 4 (Drs Arai, Shimizu, Ogawa, and Lever in 1975. However, it appeared in in MF. Therefore, the most important the sixth edition in 1983.5 In this text- Hirose) and Dermatology pathological feature of lymphomatoid (Dr Tsuchida), and Department book, epidermotropism was reportedly re- keratosis is epidermotropism resembling stricted to MF as follows5(p764): of Dermatology, Saitama 3 Medical Center (Dr Izaki), that seen in MF. Namely, the term lym- Saitama Medical University, phomatoid in lymphomatoid keratosis in- Presence of mononuclear cells in the epider- Saitama, Japan. dicates histological simulation of MF. It mis without spongiosis occurring in MF. The (REPRINTED) ARCH DERMATOL/ VOL 143, JAN 2007 WWW.ARCHDERMATOL.COM 53 ©2007 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 09/25/2021 cells lie either singly, surrounded by a clear halo, or aggre- mal collection of more than 3 lymphocytes, (2) disproportion- gated as in a Pautrier microabscess. Epidermotropism must be ate distribution of haloed lymphocytes in the epidermis, and differentiated from exocytosis. (3) linear aggregates of haloed lymphocytes at the dermoepi- dermal junction. However, there have been descriptions of epidermot- ropism in other tumors. In 1983, Burg and Braun- Falco6(p386) stated that in histiocytosis X, “An important his- IMMUNOHISTOCHEMICAL ANALYSIS tological feature is epidermotropism of the characteristic cells.” Burkert et al7 used the term pseudo-Pautrier ab- A set of 4-µm-thick tissue sections were prepared from forma- lin-fixed and paraffin-embedded blocks, deparaffinized in xy- scess for Langerhans cell microgranulomas, and Klein et 8 lene, rehydrated, and microwaved for 10 minutes at 30% al used the term epidermotropic nature to describe poro- power in citrate buffer, pH 9.0 (Dako Japan, Kyoto) for anti- carcinoma (malignant poroma). Moreover, Murphy and gen retrieval. Endogenous peroxidase activity was blocked us- Schwarting 9 used the term atypical epidermotropic infil- ing 0.3% hydrogen peroxide in methanol for 5 minutes. Im- trate for benign conditions such as lymphomatoid drug munostaining was carried out after antigen retrieval by eruption. Bayer-Garnar et al10 reported that CD68ϩ his- microwaving in 1 mM EDTA (pH 8.0) 2 times for 5 minutes tiocytes showed epidermotropism in benign lichenoid kera- each time at 900 W. Then the streptavidin-biotin-peroxidase tosis. In other words, all cells tropic for the epidermis have complex method was used for immunostaining with a Ven- so far been included in the concept of epidermotropism. tana OH system immunostainer (Ventana-Biotech, Tucson, Therefore, epidermotropism has been used mostly as Ariz). The primary antibodies against the lymphoid antigens were CD1a (O10; Immunotech, Marseille, France [undi- a descriptor for MF, which is epidermotropism in the nar- luted]), CD3 (PS-1; Novocastra, Newcastle-upon-Tyne, En- row sense. All of the cases we have examined have fallen gland [1:100 dilution]), CD4 (1F6; Novocastra [1:40 dilu- into this category, which shows complete histological simi- tion]), CD8 (C8/144B; Dakopatts, Glostrup, Denmark [1:100 larity to MF. The epidermotropism of MF is a phenom- dilution]), CD20 (L26; Dakopatts [1:100 dilution]), CD68 enon whereby lymphocytes infiltrate into the epider- (PG-M1; Dakopatts [1:200 dilution]), CD79a (JCB117; Da- mis, slippery with perinuclear halo, and do not attack the kopatts [1:100 dilution]), ␬chain (A8B5; Dakopatts [1:100 di- basement membrane or keratinocytes. This phenom- lution]), ␭ chain (C4; Immunotech [1:50 dilution]), and S100 enon occurs in specific conditions caused by long- protein (Dakopatts [1:300 dilution]). Human tonsillar tissue standing chronic inflammation in the skin-associated lym- was used as a normal control. The main features evaluated 11 during immunohistochemical testing were (1) the B to T-cell phoid tissue of the upper dermis. Kazakov et al and ϩ ϩ 12 ratio; (2) the ratio of the CD4 to CD8 cells; (3) the pheno- Massone et al have described epidermotropism in sev- type of epidermotropic cells; and (4) the proliferation of eral of the following conditions: (1) disproportionate epi- CD1a- and/or S100 protein-positive cells in the epidermis and dermotropism (presence of many lymphocytes distrib- dermis. uted irregularly), (2) basilar lymphocytes (presence of linear aggregates of small hyperconvoluted lympho- GENOTYPIC ANALYSIS cytes located at the dermoepidermal junction), (3) large epidermal lymphocytes (intraepidermal lymphocytes For genotypic studies, DNA was extracted from 10-µm-thick larger than those in the dermis), (4) Pautrier microab- paraffin sections by proteinase K (200 µg/mL) digestion. The scesses (presence of intradermal collection of more than supernatant containing DNA was used directly for polymerase 3 lymphocytes), and (5) haloed lymphocytes (presence chain reaction amplification. The polymerase chain reaction was of a perinuclear clear space), all of which may be recog- used to detect rearrangement of the immunoglobulin heavy chain nized as epidermotropism in the narrow sense. (IgH) gene using 3 VH primers (FR1c, FR2a, and FR3a) and 2 13 ␤ Based on these facts, we performed clinicopathologi- JH primers (LJH and VLJH) ; the T-cell receptor (TCR) chain ␤ ϩ ␤ ϩ cal, immunohistochemical, and molecular biological stud- gene using the following 5 primer sets: 1 (V J1), 2 (V J2), ␤ ϩ ␤ ϩ ␤ ϩ 14 ␥ ies of 6 cases of clinically benign scaly plaque and epi- 3 (D1 J1), 4 (D1 J2), and 5 (D2 J2) ; and the TCR chain gene using the following 20 mixed primer sets: mix 1 (V V , dermotropism histologically similar to that in MF. We 2 8 V3,V4,V9,J1.12.1,J1.2,J1.32.3,J1.12.1 in,J1.2 in,J1.32.3 in)andmix2(V5, investigated whether lymphomatoid keratosis should be 15 V10,V11,V12,J1.12.1,J1.2,J1.32.3,J1.12.1 in,J1.2 in,J1.32.3 in). classified as an epidermotropic subtype of cutaneous lym- phoid hyperplasia. RESULTS METHODS CLINICAL FINDINGS CASE SELECTION The clinical data are demonstrated in Table 1. The pa- tients ranged in age from 36 to 78 years, with a mean age A retrospective review of the files of Saitama Medical Univer- of 59 years, and the male-female ratio was 1:1. The size sity Hospital and Saitama Medical Center Hospital, Saitama, Ja- (largest diameter) of the lesions ranged from