PATIENTS AND MATERIAL

Formalin fixed, paraffin embedded, snap frozen material from patients with EITCL, ulcerative jejunitis, and those with refractory sprue evolving into overt EITCL were from centres participating in the German multicentre study on intestinal non-Hodgkin's lymphoma. Patients with microscopic overt infiltration of malignant tumour cells into their duodenal biopsies were excluded. As ulcerative jejunitis represents a special form of EITCL, results of patients with ulcerative jejunitis, refractory sprue evolving into overt EITCL, and primary overt EITCL were combined as the EITCL group. For patients with refractory sprue and coeliac disease, paraffin blocks of tissue taken for diagnostic reasons were drawn from the archives of the Institute of Pathology. Included in the study were 11 consecutive patients who presented initially with villous atrophy refractory to a strict gluten free diet and who were negative for antiendomysium and antigliadin antibodies under a gluten free diet. Coeliac disease was diagnosed according to the ESPGAN criteria or by detection of antibodies to endomysium.20 ,21

DNA ISOLATION

DNA was extracted from formalin fixed, paraffin embedded, duodenal and colonic biopsies and resected tumour according to the manufacturer's instructions (QIAamp tissue kit; Qiagen GmbH, Hilden, Germany). DNA was stored at 4°C for a maximum of four weeks until further usage.

PRIMER SETS

Oligonucleotide consensus primers for detection of Vγ1 to Vγ8 segments were used in a semi nested polymerase chain reaction (PCR) in conjunction with established Jγ primers, as described by Dippel and colleagues.24 Primers were synthesised on an automated DNA synthesiser (Model 381A; Applied Biosystems, Weiterstadt, Germany) and purified by high pressure liquid chromatography.

PCR CONDITIONS

A semi nested PCR was performed for detection of clonal TCR-γ rearrangements. In the first round of amplification, two separate reactions were carried out employing the same Vγ1–8 consensus primer in conjunction with the Jγ primer JGT1/2 and JGT3, respectively. An aliquot (1%) of the first two PCRs was used for reamplification which was performed with a nested Vγ1–8 primer and the same Jγ primers (table 1). DNA (200 ng) extracted from paraffin embedded tissue sections was used in the first PCR. The reaction (100 μl) was carried out after a hot start technique (AmpliWax PCR GEM 100; Perkin Elmer, Weiterstadt, Germany) with 2.0 mmol/l MgCl₂, 800 μmol/l deoxyribonucleotides, 1.25 U of Taq polymerase (Ampli Taq DNA Polymerase; Perkin Elmer), 10 μl of 10× PCR buffer (Perkin Elmer), and 200 ng of each primer. For reamplification, 1 μl of the first amplification product was used as a DNA template. Buffer conditions were identical with the exception of 1.5 mmol/l MgCl₂. Cycling parameters were 96°C (15 seconds) for denaturation, 60°C (40 seconds) for primer annealing, and 72°C for strand elongation for 30 seconds each (25 cycles). For denaturation, 96°C was extended in the first step to four minutes. PCR products were visualised on ethidium bromide stained 6% polyacrylamide gels. For negative controls water samples and for positive controls DNA from a T cell lymhoma cell line (PEER) were run in parallel with each PCR. Contamination was strictly avoided by separating DNA isolation, PCR set up, and processing into different rooms, using gloves at all times.

FLUORESCENT FRAGMENT ANALYSIS (GeneScan ANALYSIS)

For GeneScan analysis, reamplification was performed under the same conditions as described above employing fluorescence labelled Vγ1–8 primers. Deionised formamide (2.0 μl), 0.5 μl of loading buffer, and 0.5 μl of 6-carboxyrhodamine dye labelled DNA size standard buffer mixture (Genescan-500-ROX; Applied Biosystems, Weiterstadt, Germany) were mixed with 1 μl of PCR product and heated for three minutes to 90°C. The samples were immediately chilled and 5.0 μl loaded onto a 6% acrylamide gel containing 6 mmol/l urea and analysed in an automated DNA sequencer 373A. Size determination was performed on a Macintosh computer with the GeneScan software 672 (Applied Biosystems).24 Clonality was defined as the occurrence of one or two dominant bands.

CLONING AND SEQUENCING

Biclonal or monoclonal PCR products were cut out from the polyacrylamide gels and incubated overnight in 20–30 μl of distilled water. The supernatant (5 μl) was used for direct sequencing (BigDye; Perkin Elmer) employing the reamplification primers in two separate sequencing reactions. The reaction mixtures were analysed on a 377A automated DNA sequencer (Applied Biosystems) according to the manufacturer's instructions.

HISTOLOGY AND IMUNOHISTOCHEMISTRY

Endoscopically removed or resected material was fixed in 10% formalin. Villous atrophy was graded according to Marsh and colleagues25: grade I, infiltrative lesion; grade II, infiltrative-hyperplastic lesion; grade III, flat-destructive lesion with subtotal villous atrophy; and grade IV, flat-destructive lesion with total villous atrophy. Sections of formalin fixed paraffin embedded tissue were used for immunohistochemistry applying the immunoalkaline phosphatase method with the following antibodies: βF1 (TCR β chain) from T cell Sciences (Cambridge, USA), 123C3 (CD56) from Monosan (Uden, Netherlands), and 1F6 (CD4), C8-144 (CD8), and CD3 (polyclonal) from Dako (Glostrup, Denmark). Pretreatment was done using high pressure cooking for antigen retrieval in a 10 mmol/l citrate buffer, pH 6.0. Pretreatment for βF1 consisted of proteolytic digestion. IELs and positively stained IELs were counted per 750 epithelial cells under blinded conditions. Results are expressed as (positively stained) IELs per 100 epithelial cells and per cent of positively stained IELs.

STATISTICAL ANALYSIS

Results were compared using the Mann-Whitney U test and the Kruskal-Wallis test for unpaired samples. Results of patients with refractory sprue due to defined disorders were not evaluated in this test because of different disease entities. For evaluation of TCR-γ gene rearrangement, the χ² test was used.

Specificity was defined as the frequency of a negative test (clonal rearrangement of the TCR-γ gene and/or loss of T cell antigens) in patients without EITCL (coeliac disease and refractory sprue due to defined disorders) and sensitivity as the frequency of a positive test in patients with EITCL (primary overt EITCL, ulcerative jejunitis, and refractory sprue evolving into overt EITCL).

HISTOLOGICAL FINDINGS

IELs were increased in all patients with EITCL above 40 per 100 epithelial cells. In patient EITCL8, overt infiltration of the duodenum by the lymphoma was found on histology and immunohistology. Data from this patient were therefore excluded from the statistical analysis. All patients had grade III–IV enteropathy except for patient EITCL2 who had grade II enteropathy with a patchy increase in CD56 positive IELs (table 1). Multiple ulcerations without evidence of lymphomatous infiltration were observed in the duodenum of patients with ulcerative jejunitis.

In the first patient with collagenous sprue (CoS1), a 35 μm thick subepithelial band-like structure, positive for van Gieson's staining, was seen. In the other patient with collagenous sprue (CoS2), subepithelial collagen deposition reached a thickness of 50 μm. In the patient with common variable immunodeficiency, weakly periodic acid-Schiff positive macrophages and a reduced number of plasma cells were found. A non-specific lymphocytic infiltrate was seen in the patient with ulcerative colitis associated sprue syndrome. Finally, the patient with autoimmune enteropathy presented with signs of chronic inflammation in the lamina propria, vacuolisation of the surface epithelium, but no significant increase in IELs.

Patients with coeliac disease showed a dense mononuclear infiltrate in the lamina propria. Median IEL counts in coeliac disease were higher in comparison with the other groups but this difference was not statistically significant (table 1, fig 1A).

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Absolute numbers of intraepithelial T lymphocytes (IELs) (A), and positively marked IELs with anti-CD3 (B), CD8 (C) and T cell receptor (TCR)-β (E) per 100 epithelial cells are shown. In addition, the percentage of positively marked IELs with anti-CD8 (D) and anti-TCR-β (F) of the total IELs was calculated. Clonal rearrangements of the TCR-γ gene are depicted by filled symbols. The number of CD 8 and TCR-β positive IELs in duodenal biopsies of patients with enteropathy-type intestinal T cell lymphomas (EITCLs) were significantly reduced in comparison with patients with coeliac disease (C–F). In most cases, clonality was associated with reduced expression of T cell antigens. RSUC, refractory sprue of uncertain cause; RSDD, refractory sprue due to defined disorders. For other abbreviations see table 1.

IMMUNOHISTOLOGY

CD3 staining of the duodenal biopsies confirmed the histological data, except in two cases in which lower numbers of CD3 positive IELs were counted (EITCL1 and EITCL2; fig 1B). In patient EITCL2, IELs and tumour cells were CD56 positive. Duodenal biopsies from patients with coeliac disease showed significantly more CD3 positive IELs compared with patients in the EITCL group, and compared with patients with refractory sprue of uncertain cause (fig 1B, fig 2). The numbers of CD3 positive IELs in patients with sprue syndromes due to defined disorders were similar to those of the EITCL group (fig 1B). As the absolute numbers of IELs were very variable, we used the proportion of CD8 or TCR-β positive IELs as a more reliable factor. The proportions of CD8 and TCR-β positive IELs were significantly decreased in patients in the EITCL group and in those with refractory sprue of uncertain cause in comparison with patients with coeliac disease. The proportion of TCR-β positive IELs in patients in the EITCL group was decreased compared with refractory sprue patients of uncertain cause (fig 1F, fig2). One patient with coeliac disease had low numbers (<50%) of CD8 positive IELs but responded to a gluten free diet. Histological examination indicated an improvement after a gluten free diet. This patient is clinically well five years after the initial diagnosis. In three patients with EITCL, TCR-β staining could not be evaluated because the initial staining failed and further tissue was not available. CD4 was expressed only in a small proportion of IELs in all groups of patients (medians (minimum-maximum): EITCL 3 (0–9); refractory sprue of uncertain cause 6 (4–9); refractory sprue due to defined disorders 3 (1–17); and coeliac disease 3 (1–23); p=0.22).

• Download figure
• Open in new tab
• Download powerpoint

Figure 2

Haematoxylin-eosin staining (A) and immunohistochemistry (B–D) of a duodenal biopsy of a patient with an enteropathy-type intestinal T cell lymphoma (EITCL) showing an increased number of intraepithelial T lymphocytes (IELs) in conventional histology (A) that were mostly positive for CD3 (B) but negative for CD8 (C) and T cell receptor (TCR)-β (D), and a patient with coeliac disease (haematoxylin-eosin (E); CD3 (F); CD8 (G); TCR-β (H)). IELs in the duodenal biopsy of the patient with coeliac disease were CD3, CD8, and TCR-β positive (F–H). Original magnification 100×.

Patients with collagenous sprue had CD3 positive IELs exceeding 40 per 100 epithelial cells. The proportion of CD8 positive IELs was above 50% but TCR-β was only detected in a small percentage of IELs (see fig 1D, F) in both.

Lymphoma tissue of patients with overt EITCL and refractory sprue evolving into EITCL was negative for CD8 in 8/12 cases and negative for the TCR-β in 9/10 cases (in two cases staining for TCR-β was not possible). There was no correlation between the phenotype of the tumour cells and the phenotype of IELs in duodenal biopsies (data not shown).

REARRANGEMENT OF THE TCR-γ GENE

Investigation of TCR-γ gene rearrangement for clonality in duodenal biopsies of patients with coeliac disease and of patients with sprue syndromes due to defined disorders revealed polyclonal results in all cases (fig 1). In duodenal biopsies of 7/13 patients in the EITCL group, amplification revealed clonal rearrangements (p<0.005 compared with the non-EITCL group). Rearrangements of the tumour specimens were clonal with the same length and sequence of amplificate (table 1, fig3). Both patients with ulcerative jejunitis had a clonal rearrangement in their duodenal biopsy as well as in their whole bowel wall specimen. Two of the three patients who had refractory sprue before developing overt EITCL showed the same clonal amplificate in their duodenal biopsy as that of the subsequent tumour specimen. All patients in the EITCL group with normal CD8 positive IEL counts (>50%) had a polyclonal TCR-γ gene rearrangement in their duodenal biopsies (medians for CD8 positive IELs per 100 epithelial cells: polyclonal 68.5; monoclonal 6 (p<0.05); medians for TCR-β positive IELs: polyclonal 17; monoclonal 13.5 (p=0.22)). The exception was patient EITCL2 whose CD8 positive IELs were mainly CD56 positive and whose duodenal biopsy was monoclonal for the TCR-γ gene rearrangement.

• Download figure
• Open in new tab
• Download powerpoint

Figure 3

GeneScan analysis of polymerase chain reaction (PCR) products obtained by amplification of DNA from the T cell line PEER (A), of polyclonal tonsillar DNA (B), of DNA from a patient with coeliac disease showing polyclonal amplification (C), of DNA from a duodenal biopsy of patient EITCL7 with a reduced number of CD8 and T cell receptor (TCR) β positive intraepithelial T lymphocytes (IELs) on immunohistology (D), and of DNA from the lymphoma of the same patient showing the same length of monoclonal amplificate in both cases (E).

Amplification of the TCR-γ gene revealed clonal rearrangement in only 1/6 patients with refractory sprue of uncertain cause. An abnormal phenotype of IELs—CD8 and TCR-β negativity—was demonstrated in the duodenal biopsy of the monoclonal case (RS6); all other cases in this group were mostly CD8 or TCR-β positive (table 1, fig 1). In general, clonality was associated with decreased expression of CD8 and TCR-β to less than 50% of IELs in 7/8 cases (fig 1).

SPECIFICITY AND SENSITIVITY

The specificity for detection of an EITCL in comparison with coeliac disease and with refractory sprue due to defined disorders was 77% with a cut off at 50% of positive IELs for CD8 and TCR-β. Sensitivity for detection of an EITCL was 85% (table 2). With a cut off at 25% of positive IELs for CD8 and TCR-β, specificity was still 89%, but sensitivity was only 31%. Immunohistology for TCR-β was more specific than for CD8 (table 2, fig 1).

The specificity of TCR-γ gene clonality for detection of an EITCL in comparison with coeliac disease and refractory sprue due to defined disorders was 100% but sensitivity was only 62% (table 2).