J Cutan Pathol 2001: 28: 351–355 Copyright C Munksgaard 2001

Printed in Denmark ¡ All rights reserved Journal ofCutaneous Pathology

ISSN 0303-6987

Heteroduplex analysis of T-cell receptor ggene rearrangement as an adjuvantdiagnostic tool in skin biopsies forerythrodermaBackground: Erythroderma, defined as red skin covering most of the Sarah Cherny1, Serena Mraz2,body surface often accompanied or followed by exfoliation, is the clin- Lyndon Su3, Jeff Harvell1,2 andical manifestation of at least six different underlying etiologies with Sabine Kohler1,2

allergic or irritant contact dermatitis, atopic/asteotic dermatitis, pity- Departments of 1Pathology and 2Dermatology,riasis rubra pilaris (PRP), psoriasis, and seborrheic dermatitis account- Stanford University Medical Center, Stanford

California, USA, Departments of 3Pathology anding for the majority of cases. Approximately 10% of cases are dueDermatology, University of Michigan, Annto adverse drug reactions with roughly another 10% due to cutaneousArbor, Michigan, USAT-cell lymphoma (CTCL), predominantly mycosis fungoides, or leuke-

mia. It is clear from multiple studies that the clinical diagnosis of theunderlying entity is often difficult, as these diseases can present in avery similar fashion. A skin biopsy is usually employed in this settingas a diagnostic tool. However, the histopathologic diagnosis of theunderlying cause is complicated by the subtlety of the distinguishinghistologic features. In this situation, an ancillary technique demonstrat-ing the presence of a monoclonal T-cell proliferation could help torule in or out CTCL in cases that clinically and histopathologically donot allow a definitive diagnosis.Methods: We retrospectively studied 25 biopsies from sixteen pa-tients who presented to the Stanford Dermatology Clinic with er-ythroderma. We examined the specimens morphologically and ana-lyzed the gamma chain of the T-cell receptor (TCR- g) by polymerasechain reaction (PCR) followed by heteroduplex analysis for clonality.We then correlated the results of our PCR and heteroduplex analyseswith the patients’ clinical outcomes.Results: Four biopsies, from three patients, contained clonal TCR-grearrangements; the four biopsies, all of which were equivocal histolog-ically, correlated to diagnoses of mycosis fungoides (MF) or Sezarysyndrome (SS). Twenty-one biopsies contained polyclonal T-cellpopulations. Eighteen of these biopsies represent patients with in-flammatory dermatoses. Three of these biopsies, all of which weretaken from a single patient, correlate to a diagnosis of MF.Conclusion: TCR-g PCR heteroduplex analysis seems to representan important adjuvant diagnostic tool that, used in conjunction withhistopathology and clinical history, could help to clarify the underlyingetiology of erythroderma.

Sabine Kohler, Dept. Pathology, StanfordUniversity Medical Center, Stanford, CA 94305,Cherny S, Mraz S, Su L, Harvell J, Kohler S. Heteroduplex analysisUSA

of T-cell receptor g gene rearrangement as an adjuvant diagnostic e-mail: skohler/stanford.edutool in skin biopsies for erythroderma.J Cutan Pathol 2001; 28: 351–355. C Munksgaard 2001. Accepted January 2, 2001

351

Cherny et al.

Erythroderma, defined as generalized erythema in-volving at least 90% of the total body surface areaand often accompanied by scaling, occurs in associ-ation with a variety of disorders.1 History of a pre-existing dermatosis such as psoriasis, pityriasis rubrapilaris (PRP), or atopic dermatitis may aid in identify-ing an underlying cause in some patients. However,in most cases, identification of the underlying etiologyis difficult since the histopathologic features that nor-mally distinguish among the different dermatoses aremore subtle in the context of erythroderma.3–5 Er-ythroderma may be the first manifestation of mycosisfungoides or Sezary syndrome (SS), which is a triadof erythrodermic MF, circulating Sezary cells in theperipheral blood (.5–10%), and clinically enlargedlymph nodes. Well-defined histologic criteria exist forpatch and plaque stage MF, but erythrodermic MFcan be more difficult to recognize histologically.6,7

While Southern blotting is the gold standard to de-termine T-cell receptor (TCR) clonality in skin biop-sies, this procedure has important disadvantages thathave made polymerase chain reaction (PCR)-basedmethods a preferred alternative. Southern blotting re-quires a large amount of high quality DNA that canonly be extracted from unfixed tissue, involves ex-posure to radioactive isotopes, and is only able to dis-cern clonal T-cell populations that comprise at least5% of the total T-cell population. In contrast, thePCR-based methods are comparatively fast andsimple, can be performed on small quantities of DNAderived from either fresh or paraffin-embedded tissuesamples, and can detect clonal populations that makeup less than 1% of the T-cells in the sample.3

While multiple other studies have shown the utilityof PCR analysis of the TCR as a tool in detecting aclonal population of T-cells in skin biopsies,3,8–11

none to our knowledge has approached the topic ina retrospective fashion, for the specific disease entityof erythroderma. This is important not only for deter-mining the procedure’s ability to detect clonality inthe early stages of erythroderma but also for quantify-ing the false-positive rate among the inflammatorydermatoses.

Material and methodsPatientsWe obtained 25 formalin-fixed, paraffin-embeddedtissue samples from sixteen patients, all of whom pre-sented with erythroderma to the Stanford Derma-tology Clinic. Nine men and seven women were in-cluded in the study, ranging in age from 54 to 86years, with an average age of 72.2 years. All tissue-samples were taken at the time of presentation witherythroderma and were obtained between 1996 and1999, and only the patients for whom we had suf-ficient follow-up clinical information, as determined

352

by review of the patients’ charts, were included in thestudy.

PCR and heteroduplex analysisAmplification and heteroduplex analysis of the TCR-g chain was performed as previously described.12

Consensus primers for the TCR-g V-J junction wereused to amplify genomic DNA isolated from paraffin-embedded tissue. The amplification of the TCR-g V-J junction was performed in a Perkin-Elmer 2400thermal cycler with a single set of consensus primers:V1 (5ø-TACATCCACTGGTACCTACACCA-3ø)for the V g 1 family (V g 1–8) and J1/2 (5ø-CCCGT-CGACTACCTTGGAAATGTTGTATTCTTC-3ø)for the J g 1 and J g 2 segments. Positive and negativecontrols were included in all PCR reactions.

Conditions were as follows: 100–500 ng of genomicDNA was added to a reaction mixture containing2.5U AmpliTaq Gold polymerase (Perkin-Elmer, Fos-ter City, CA, USA), 10 pM of each primer, 125 mMof each dNTP, in Perkin-Elmer PCR buffer II (10mM Tris-HCl, pH 8.3, 50 mM KCl) with 1.5 mMMgCl2. After 95 æC denaturation the TCR-g wasamplified as follows: 20 cycles of 95 æC denaturation(40 s), 65 æC annealing (30 s), decreasing by 1 æC/cycle, 72 æC extension (30 s), followed by 15 cycleswith the same denaturing and extension temperaturesand an annealing temperature of 45C, followed by afinal 72 æC extension for 5 min.

PCR products were first visualized on a 2% agar-ose gel to document successful amplification. TwentymL of the PCR product with 2 mL of 0.05 M EDTAadded was heated to 95 æC for 10 min and then slowlycooled to 37 æC over 30 min using the thermal cycler,then placed on ice. The samples were then electro-phoresed on a Mutation Detection Enhancement(MDE) gel (FMC Bioproducts, Rocklands, ME,USA), for 12 h at 80 V in tris-borate-EDTA buffer.Heteroduplex patterns were then determined bystaining with ethyidium bromide and photography onultraviolet (UV) light.

We excluded any samples pre-dating 1996 from thestudy due to deterioration of the DNA. Some suchsamples produced a band on gel electrophoresis fol-lowing b-globin PCR, performed prior to TCR-gPCR to check the integrity of the DNA. However,the samples did not produce bands following TCR-gPCR, indicating degradation of the relevant areas ofDNA.

ResultsClinical dataClinical follow-up was available for all patients in thestudy, and ranged from 8 to 37 months, with an aver-age of 19 months. Of the nine men and seven women

Heteroduplex analysis for erythroderma

Table 1. PCR Results and clinical follow-up

patient . age and sex follow-up clinical follow-up PCR

1 70, F 37 months drug reaction ª2 73, F 33 months psoriasis ª3 71, M 31 months asteotic/atopic dermatitis ª4a 86, M 29 months contact dermatitis ª4b 86, M 29 months contact dermatitis ª5 72, M 28 months asteotic/atopic dermatitis ª6a 72, M 18 months contact dermatitis ª6b 72, M 18 months contact dermatitis ª7a 56, M 17 months MF ª7b 56, M 17 months MF π7c 56, M 17 months MF ª7d 56, M 17 months MF ª8a 83, F 17 months contact dermatitis ª8b 83, F 17 months contact dermatitis ª9 86, F 16 months asteotic/atopic dermatitis ª

10 73, F 15 months drug reaction ª11 79, F 14 months psoriasis ª12a 71, M 12 months PRP, diagnosed 12/98 ª12b 71, M 12 months PRP, diagnosed 12/98 ª13 77, F 12 months drug reaction ª14 58, M 10 months MF/SS π15a 74, M 9 months drug reaction ª15b 74, M 9 months drug reaction ª16a 54, M 8 months MF/SS π16b 54, M 8 months MF/SS π

mean: 72.2 mean: 19 morange: 54–86 range:8 –37 mo

9 M, 7 W

included in the study, three were diagnosed with MF,as determined by file review; the remaining thirteenpatients’ erythroderma was determined to be due toa variety of conditions: asteotic/atopic dermatitis(3 patients), generalized contact dermatitis (3 pa-tients), drug reactions (4 patients), psoriasis (2 pa-tients), and PRP (1 patient).

Diagnoses for the three patients with MF werebased on the following: Patient .7 had 30% circu-lating Sezary cells and developed lymphadenopathywithin one year after presenting with erythroderma.At the time of original presentation, however, the pa-tient had no lymphadenopathy and no blood workwas done. Patient .14 had a history of stage 1a my-cosis fungoides prior to presenting with erythrodermaand had 30% circulating Sezary cells and lymph-adenopathy at the time of presentation with erythrod-erma. Patient .16 had biopsy-confirmed involvementof an inguinal lymph node with mycosis fungoidesand had .30% circulating Sezary cells at the time ofpresentation with erythroderma.

Diagnoses for patients with drug reactions (patients.1, 10, 13, 15) were based on the observations thatthe patients’ symptoms cleared after removal of theoffending agent. Patients who were determined tohave atopic dermatitis (patients .3, 5, 9) had signifi-cant atopic histories and improved with aggressiveskin emoliation and mid-strength topical steroid ther-apy. Patients with a diagnosis of contact dermatitis(patients .4, 6, 8) cleared with topical and/or oralcorticosteroid therapy and did not recur after avoid-

353

ance of the offending agent. Patients .6 and .8 hadpositive reactions to allergic patch testing, and patient.4 had a suspected source of allergy found throughclinical history. The patients with psoriasis (patients.2 and 11) both had a history of plaque-type psoriasisprior to presenting with erythroderma. Patient .2 hadmild patch-type disease for 5 years prior to develop-ing erythroderma and continued to have typicalpatches of psoriasis that waxed and waned in severityafter the erythroderma was treated with aggressivetopical therapy. Patient .11 was known to have a longhistory of severe psoriasis prior to developing erythro-derma and, with improved control of her disease, shecontinues to have typical plaques of psoriasis. Patient.12 had classical ‘‘islands of sparing’’ and had pal-mar-plantar hyperkeratosis and coloration typical ofpityriasis rubra pilaris. Furthermore, his response totreatment with an oral retinoid was in keeping forPRP.

Table 1 summarizes the clinical data. The PCRanalysis was performed in a retrospective fashion anddid not play a role in establishing a final clinical diag-nosis.

Molecular studiesFour of the 25 biopsies contained clonal T-cell pro-liferations as shown by TCR-g rearrangements; allfour biopsies were from patients with MF. Thetwenty-one biopsies with polyclonal T-cell popula-tions represent patients who, through file review, weredetermined to have atopic dermatitis, allergic or irri-tant contact dermatitis, drug reactions, PRP, psoriasis,pemphigus foliaceous, or, in one case, MF. Throughserial dilutions of our positive control (DNA extractedfrom a lymph node with peripheral T-cell lymphoma)in our negative control (DNA extracted from tonsil),we established that a clonal T-cell population com-prising less than 1% of the total DNA produced apositive result on heteroduplex analysis. Results of theTCR- g PCR are included in Table 1.

DiscussionThe difficulty of identifying the underlying etiologyof erythroderma through histopathology alone is wellestablished, and several articles address this issue.1, 2,

4, 5 Walsh et al. examined 56 skin biopsies from 40erythrodermic patients and found that the mean ac-curacy of the histopathological diagnoses was 53%,with a range of 48–66%; accuracy varied by pathol-ogist as well as by diagnosis, e.g., the greatest degreeof accuracy was observed in cases of spongiotic der-matitis and the least in those of PRP.4 Kohler et al.describe the difficulties in identifying erythrodermicMF and Sezary syndrome patients: although clearlydefined histologic parameters exist for diagnoses of

Cherny et al.

MF in well-developed patch and plaque stages, thesame histologic parameters do not hold true for thosepatients who present with erythroderma.6 Botella-Es-trada et al. also acknowledge the difficulty of assigningetiology to erythroderma through histopathologyalone, citing that several biopsies were necessary be-fore establishing a final diagnosis in eight out of atotal of 36 biopsies from erythrodermic patientsevaluated. Although the 36 patients collectively repre-sented a variety of final diagnoses, of the eight pa-tients that required multiple biopsies, five of them hadCTCL, suggesting that erythrodermic CTCL has anespecially challenging histopathology.1 Clearly, itwould be beneficial to both patient and physician tobe able to establish a diagnosis more quickly andmore accurately with diminished need for multiplebiopsies from each patient.

Multiple studies have evaluated the utility of PCRas a diagnostic aide in evaluating potential CTCLcases. Guitart and Kaul analyzed skin biopsies from48 patients with ‘‘possible CTCL.’’ Twenty-six of the48 patients proved to have CTCL; of those 26, 19(73%) showed a clonal proliferation with PCR.10 Sig-noretti et al., in a similar study, found a clonal pro-liferation in 20 out of 21 (95%) of T-cell lymphomacases.9

To the best of our knowledge, no studies have fo-cused on the utility of TCR- g PCR in erythrodermicpatients. With this study, we aimed to evaluate theeffectiveness of establishing clonality in CTCL pre-senting as erythroderma. Additionally, we evaluatedthe potential for false-positives among the polyclonalinflammatory infiltrates associated with the majorityof erythrodermic patients. In our study, 25 biopsiesfrom sixteen patients were analyzed by TCR-g PCRand the results were correlated with the patients’ clin-ical outcomes as determined through file review. Fourof the 25 biopsies contained clonal TCR-g rearrange-ments; all four represented biopsies from patients whoproved to have MF. These results correspond to100% specificity and a 0% false-positive rate in thislimited sample size. The remaining 21 negative TCR-g PCR results, indicating polyclonal T-cell popula-tions, represent thirteen different patients; 12 of thesepatients had inflammatory dermatoses while one pa-tient had MF. Over the course of one month, thisparticular patient had four biopsies, .7a-d, all ofwhich were included in our study. The first two, 7aand 7b, were taken while the patient was on topicalsteroids, prescribed for the patient’s ongoing papularrash of unknown etiology. The second set, 7c and 7d,were taken after the patient had discontinued his top-ical therapy for a period of 3 weeks. The histopath-ology was similarly equivocal in both sets of biopsies.During our study, one of the four biopsies produceda clonal band, whereas the remaining three did not.This not only underscores the importance of multiple

354

biopsies, especially in patients who may be clinicallysuspicious for MF, but also adds weight to the pres-ence of a positive clonal result in even one of thebiopsies. It is likely that the three negative biopsiescontained infiltrates that were below the threshold ofdetection. This patient’s diagnosis was confirmed oneyear after the original presentation by the presence ofcirculating Sezary cells and lymph node involvementof T-cell lymphoma.

In conclusion, a more sensitive and specific diag-nostic procedure for patients presenting with erythro-derma would mean more appropriate and earlier in-stituted therapy, especially in the case of patients withCTCL, and avoidance of serial biopsies, particularlywhen the histology is ambiguous. Our study is basedon a small sample of patients with a relatively shortfollow-up period, and similar studies with larger num-bers of subjects would be necessary to substantiate ourconclusions. In our hands, TCR-g PCR heteroduplexanalysis seems to represent an important adjuvant di-agnostic tool that, used in concert with histopathologyand clinical history, can help to clarify underlyingetiology of erythroderma.

References1. Botella-Estrada R, Sanmartin O, Oliver V, Febrer I, Aliaga A.

Erythroderma. A clinicopathological study of 56 cases. ArchDermatol 1994; 130: 1503.

2. King LE Jr. Erythroderma. Who, where, when, why, and how.Arch Dermatol 1994; 130: 1545.

3. Kohler S, Zehnder JL. Use of the polymerase chain reactionin the evaluation of cutaneous T-cell infiltrates. Dermatol Clin1999; 17: 657.

4. Walsh NM, Prokopetz R, Tron VA, et al. Histopathology inerythroderma: review of a series of cases by multiple observers.J Cutan Pathol 1994; 21: 419.

5. Zip C, Murray S, Walsh NM. The specificity of histopath-ology in erythroderma. J Cutan Pathol 1993; 20: 393.

6. Kohler S, Kim YH, Smoller BR. Histologic criteria for thediagnosis of erythrodermic mycosis fungoides and Sezary syn-drome: a critical reappraisal. J Cutan Pathol 1997; 24: 292.

7. Smoller BR, Detwiler SP, Kohler S, Hoppe RT, Kim YH.Role of histology in providing prognostic information in my-cosis fungoides. J Cutan Pathol 1998; 25: 311.

8. Andersen WK, Li N, Bhawan J. Polymerase chain reaction-denaturing gradient gel electrophoresis (PCR/DGGE)-baseddetection of clonal T-cell receptor gamma gene rearrange-ments in paraffin-embedded cutaneous biopsies in cutaneousT- cell lymphoproliferative diseases. J Cutan Pathol 1999; 26:176.

9. Signoretti S, Murphy M, Cangi MG, Puddu P, Kadin ME,Loda M. Detection of clonal T-cell receptor gamma gene re-arrangements in paraffin-embedded tissue by polymerasechain reaction and nonradioactive single-strand confor-mational polymorphism analysis. Am J Pathol 1999; 154: 67.

10. Guitart J, Kaul K. A new polymerase chain reaction-basedmethod for the detection of T-cell clonality in patients withpossible cutaneous T-cell lymphoma. Arch Dermatol 1999;135: 158.

Heteroduplex analysis for erythroderma

11. Chhanabhai M, Adomat SA, Gascoyne RD, Horsman DE.Clinical utility of heteroduplex analysis of TCR gamma generearrangements in the diagnosis of T-cell lymphoproliferativedisorders. Am J Clin Pathol 1997; 108: 295.

355

12. Kohler S, Jones CD, Warnke RA, Zehnder JL. PCR-heterod-uplex analysis of T-cell receptor gamma gene rearrangementin paraffin-embedded skin biopsies. Am J Dermatopathol2000; 22: 321.