Materials and methods

Frozen tissue samples from 39 cases of PCBCL were retrieved. The corresponding histologic and immunohistochemistry sections were reviewed by three of the authors (FJC, EC, RRJ) and classified according to the REAL classification (^{Harris et al, 1994}). The diagnosis of PCBCL was based on diagnostic or compatible histology plus either light chain restriction or evidence of a clonal B cell population using southern blot IgH gene analysis. In all cases there had been no evidence of extra-cutaneous disease at presentation or for 6 mo following diagnosis, assessed by adequate staging procedures, which included a complete physical examination, thoraco-abdominal-pelvic CT scan, and bone marrow biopsy.

Immunohistochemical staining

Detailed immunophenotyping was performed on routinely fixed paraffin-embedded tissue sections using a panel of monoclonal antibodies that included anti-CD3, anti-CD20, anti-CD21, anti-CD79a, anti-, and anti-, and the monoclonal anti-bcl-2 protein (Dako, High Wycombe, U.K.).

DNA extraction

DNA was extracted from fresh skin biopsy specimens according to standard procedures by proteinase K digestion and phenol/chloroform extraction. Control DNA samples were extracted from the Daudi cell line (monoclonal control) and from six cases of benign cutaneous lymphoid hyperplasia (polyclonal controls). DNA was also extracted from involved skin of four cases of systemic B cell lymphoma that had presented with cutaneous lesions.

Southern blot analysis

DNA was digested separately with restriction enzymes Eco R1 and Hind III and subjected to southern blot analysis. Filters were hybridized with radiolabeled probes under high stringency conditions as previously described (^{Foroni et al, 1984}), using a genomic DNA probe for the JH region of the IgH gene.

IgH gene PCR amplification

Three different sets of consensus primers were used to amplify the CDR3 VDJ region of the IgH gene: (i) a V region consensus primer corresponding to the third framework region (FR3) (^{Deane and Norton, 1990a}) in combination with a JH consensus primer (^{Deane and Norton, 1990a}); (ii) a series of seven V region consensus primers corresponding to each FR1 region family (^{Deane and Norton, 1990b}) in combination with a JH consensus primer (each V region primer was used in a separate monoplex reaction); (iii) a series of six consensus primers corresponding to the leader segment 5' to each VH region family (^{Campbell et al, 1992}) in combination with a JH consensus primer (each leader segment primer was used in a monoplex reaction). The leader sequence primers are slightly longer than those published because this modification has produced more effective amplification.

In eight cases in which the V region/JH consensus primers had failed to detect a clonal IgH gene rearrangement, repeat analysis with V region primers and a mix of JH primers (JH1, JH2, JH4, JH5 (JH consensus), JH3, and JH6) was performed. Rarely, JH gene segments are not recognized by the JH consensus primer. False negative results may also occur due to somatic hypermutation in the JH region (^{Aubin et al, 1995}).

Figure 1 shows a schematic representation of the rearranged IgH gene and the relative positions of the amplification primers used in the three different strategies.

Figure 1.

Schematic representation of the rearranged IgH gene and the relative positions of the amplification primers used in the three different strategies.

Full figure and legend (14K)

The reaction mix (20 l final volume) consisted of 1 g template DNA, 1 PCR buffer (Amersham Pharmacia Biotech, Little Chalfont, U.K.), 2.5 mM MgCl₂, 200 M [³³P]dCTP, dNTPs, 0.01% Tween 20, 200 nM synthetic oligonucleotide consensus primers (Amersham Pharmacia Biotech) for the JH region and either FR3, FR1, or leader sequence regions, and 1 unit of Taq Polymerase (Amersham Pharmacia Biotech) in combination with taq start antibody (Clontech, Basingstoke, U.K.). A negative control reaction containing reaction mix but no template DNA was examined for each PCR assay performed. All PCR reactions were performed in duplicate for each DNA sample.

Monoplex PCR was performed using a Perkin Elmer thermal cycler (model 9700, Warrington, U.K.). The annealing temperatures used were as follows: FR3 primer, 58°C; FR1 primers, 60°C-62°C; leader sequence primers, 65°C-66°C. The PCR reactions were performed over 30–33 cycles.

The PCR products with appropriate molecular weight markers were loaded onto 2% agarose gels, subjected to electrophoresis at 100 V, and stained with ethidium bromide. If amplification was successful, an aliquot of the radiolabeled PCR products was run on denaturing 6% polyacrylamide gels. Formamide (95% formamide, 20 mmol per l ethylenediamine tetraacetic acid, 0.05% xylene cyanol FF, 0.05% bromophenol blue) loading dye was added to each amplified product and denatured at 94°C for 5 min. The samples were put on ice prior to loading onto denaturing polyacrylamide gels [6% acrylamide (Gibco BRL, Paisley, U.K.), 7 M urea (United States Biochemicals, Cleveland, OH)]. The gel was run at 30 W for 2–7 h, depending on the size of the PCR product, at room temperature in 0.5 TBE buffer. Autoradiography was carried out at room temperature for 24–72 h with an intensifying screen (Appligene Oncor, Watford, U.K.).

Sequence analysis

Sequence analysis was performed on selected products to confirm amplification of a clonal immunoglobulin rearrangement and to identify whether mutations at FR3 primer binding sites had occurred in cases in which a clonal rearrangement was demonstrated with leader or FR1 primers but in which a polyclonal result was obtained with the FR3 primer. Clonal bands were excised from the gel and the product was re-amplified using a biotinylated JH consensus primer together with the appropriate 5' primer. Amplification was determined by electrophoresis and the remaining PCR product was passed through a microspin column (Amersham Pharmacia Biotech) to remove unincorporated primers. The amplification products were then immobilized on the surface of streptavidin-coated microbeads (Dynabeads M-280, Dynal, Oslo, Norway). The sequenase chain termination DNA sequencing method was performed using the Sequenase Version 2.0 DNA Sequencing Kit (United States Biochemicals). Products were analyzed by denaturing polyacrylamide gel electrophoresis and visualized by autoradiography.

Sensitivity of technique

To determine the sensitivity of our technique, DNA obtained from the Daudi B cell line (ECACC no. 89120702) (known to harbour a clonal VH3 family IgH gene rearrangement) was serially diluted with DNA obtained from a polyclonal control. PCR/denaturing polyacrylamide gel electrophoresis was performed using the FR1/VH3 primer in combination with the JH consensus primer.

Results

Histologic diagnosis of lymphoma

Of the 39 cases that made up the study group, 10 were classified as diffuse large cell lymphoma (high-grade) and 29 were low-grade PCBCL (21 marginal zone lymphoma, five follicle center lymphoma, three definite lymphoma but difficult to classify). The high-grade lymphomas and all but seven low-grade lymphomas showed diagnostic histology. Of the seven cases in which histology was compatible but not diagnostic, three showed a clonal IgH population by southern blot analysis and the remaining four cases demonstrated light chain restriction.

Demonstration of monoclonality using the different consensus primer pairs Table I

FR3 + JH consensus primers Figure 2a

Seventeen of 38 cases (45%) demonstrated a monoclonal rearrangement of the IgH gene. The group consisted of seven out of 10 (70%) high-grade lymphomas but only 10 of 28 (36%) low-grade lymphomas. Only 38 cases were analyzed using the FR3 primer as DNA was not available in one case.

FR1 + JH consensus primers Figure 2b

Twenty-five out of 39 (64%) showed a monoclonal rearrangement using the FR1 series of consensus primers in combination with the JH consensus primer. This included six of 10 (60%) high-grade lesions and 19 out of 29 (66%) low-grade lymphomas.

Leader sequences + JH consensus primers Figure 2c

Twenty-nine of 39 (74%) showed a monoclonal rearrangement using the series of leader sequence consensus primers in combination with the JH primer, consisting of eight out of 10 (80%) high-grade and 21 of 29 (72%) low-grade lymphomas.

Overall, a clonal IgH gene rearrangement was found in 31 of 39 (79%) PCBCL, which included nine of 10 (90%) high-grade and 22 of 29 (76%) low-grade lymphomas. In 14 cases (37%) a clonal rearrangement was identified with all three primer sets, and in six cases (15%) a clone was detected with only one primer set (five with the leader sequences and one with the FR3 primer). The FR1 primer set did not identify a clonal rearrangement in any cases where a clone was not detected with either the FR3 or leader sequence primers. In eight cases where no clonal IgH gene rearrangement was detected, repeat analysis with a mix of JH primers failed to detect any further clonal rearrangements.

A clonal IgH rearrangement was detected in all four cases of systemic B cell lymphoma with cutaneous presentation (three follicle center lymphoma, one unclassifiable), but with all three primer sets in only one case. In two cases a clone was detected with two primer sets (FR3 and FR1 in one case and FR1 and leader sequences in the other). In the final patient an IgH clone was detected with the leader sequences alone.

The six cases of reactive cutaneous lymphoid hyperplasia produced a polyclonal result with all three primer sets.

The sensitivity of this technique for the detection of a clonal IgH gene rearrangement was approximately 0.5% when DNA from the Daudi cell line (VH3 family monoclonal control) was serially diluted with DNA from a reactive B cell infiltrate.

Comparison with southern blot analysis

There were 16 patients with PCBCL showing diagnostic histology that had previously undergone southern blot analysis. Of the 12 patients who had a southern blot clonal rearrangement of the IgH chain gene, 11 were clonal by PCR and one was polyclonal. In the four patients with germline southern blot analysis, two were clonal by PCR and two were polyclonal (these patients had immuno phenotypic evidence of light chain restriction).

Sequence analysis

Sequence analysis of clonal rearrangements was performed in selected cases to confirm amplification of the correct IgH VDJ sequence and to determine whether somatic hypermutation of the IgH gene was the cause of false negative results. Discrete bands representing clonal IgH rearrangements obtained using FR1 or leader sequence primers were sequenced in nine cases (six with a polyclonal pattern and three with a clonal pattern using the FR3 primer). Sequences were compared with germline sequences using the Vbase database (^{Cook and Tomlinson, 1995}) and DNAplot (^{Giudicelli et al, 1997}) (http://www.mrc-cpe.cam.ac.uk/imt-doc).

Point mutations of the sequence corresponding to the FR3 primer were identified in the six cases where a polyclonal pattern had been obtained. The number of mutations varied between two and seven and in three cases involved the nucleotide at the 3' end of the sequence. In contrast, the three cases sequenced in which a clonal rearrangement had been identified showed either a single point mutation (two cases) or no mutations (one case) of the FR3 primer binding site Table II.

Table II - Sequence analysis of the FR3 region corresponding to the FR3 primer binding site in nine cases of PCBCL.

Full table

Discussion

The detection of clonal IgH gene rearrangements using either southern blot analysis or PCR is a valuable method for assessing clonality in cutaneous B cell infiltrates where histologic and immunophenotypic features are not diagnostic (^{Landa et al, 1993}). To obtain a correct diagnosis, however, highly sensitive and specific PCR methods are required. In this study we used a monoplex PCR method incorporating three different sets of IgH V region consensus primers in conjunction with an IgH J region consensus primer to assess clonality in a set of DNA samples obtained from lesions of PCBCL.

Overall, with this PCR methodology, we demonstrated that 31 of the 39 (79%) PCBCL showed a clonal IgH gene rearrangement. We were able to detect a clonal IgH rearrangement in nine of 10 (90%) high-grade and 22 of 29 (76%) low-grade PCBCL. Our results suggest that use of the leader sequence primers was the most sensitive detection method. Primers for the FR1 and FR3 regions were less sensitive, but were able to detect a small number of IgH clonal rearrangements that were not detected using the leader sequence primers. There were no cases, however, in which only the FR1 primers detected an IgH clone. There were only eight cases of low-grade PCBCL in which all three different primer sets detected a clonal rearrangement.

A clonal IgH gene rearrangement was not detected in 21% of PCBCL with any of the primer sets used. Clonal IgH gene rearrangements may be undetectable, or detected with some primer pairs but not others, because of somatic hypermutation of rearranged VH gene segments, which may prevent annealing of the FR primers and to a lesser extent the JH primer (^{Diss et al, 1993;Segal et al, 1994b;Aubin et al, 1995;Lombardo et al, 1996}). The primers will anneal to DNA from the reactive B cell population, however, resulting in amplification of the polyclonal population alone. Cutaneous B cell lymphomas may also often consist of a large number of reactive B cells that can reduce the signal intensity of the monoclonal population, and a very small B cell clone may therefore be below the threshold for detection (^{Trainor et al, 1991;Ritter et al, 1997}).

We were able to demonstrate somatic hypermutation at primer binding sites by sequence analysis of six selected cases in which clonal rearrangements were detected with leader or FR1 primers but not the FR3 primer. Between two and seven point mutations were detected at the FR3 primer binding site. In contrast, three cases in which a clone was detected using the FR3 primer had either a single or no point mutation at the same site.

Comparison of the results obtained in our study with the southern blot analysis data, which were used as one of the "gold standards" for the diagnosis of PCBCL, demonstrated that 11 of 12 samples with a detectable southern blot rearrangement were clonal by PCR. The detection of a clonal rearrangement using southern blot analysis is not affected by somatic hypermutation, which is likely to account for the one negative PCR sample. Interestingly, of the four patients with germline southern blot analysis but diagnostic histology and immunophenotypic features, two were clonal by PCR techniques, confirming that PCR methods are more sensitive. Two cases with evidence of light chain restriction did not demonstrate clonality either by southern blot analysis or PCR methodology, however, presumably reflecting a lack of sensitivity and somatic hypermutation, respectively.

There are few previous PCR studies analysing the IgH gene in PCBCL. In 1997,^{Ritter et al} published data on 43 cases of PCBCL (23 small cell, 15 large cell) using the FR3 primer alone. A clonal IgH rearrangement was demonstrated in seven of 23 (30%) with small cell morphology and five of 12 (42%) with large cell morphology. Overall, these FR3 primer results are similar to our findings. This study was performed on paraffin-embedded tissue, and DNA degradation may have prevented amplification of the IgH gene in 30% of cases because only 70% of cases demonstrated an amplifiable -globin gene. More recently Nihal et al compared IgH PCR assays using multiplex FR1 and FR3 primers and DNA extracted from fresh tissue in 12 cases of cutaneous B cell lymphoma (nine PCBCL). In contrast to our findings a clonal IgH gene rearrangement was detected in 11 of 12 cases with the FR3 primer but only five of 12 cases with the FR1 primer set (^{Nihal et al, 2000}).

Previous studies (reviewed in Table III) have used different sets of primers for detection of clonal IgH gene rearrangements in systemic B cell lymphoma (^{McCarthy et al, 1990;Trainor et al, 1990,1991;Deane and Norton, 1991;Deane et al, 1991;Ramasamy et al, 1992;Diss et al, 1993;Segal et al, 1994a, b;Aubin et al, 1995;Lombardo et al, 1996;Essop et al, 1997;Derkson et al, 1999}) and the majority employ primers for the FR3 region (^{Trainor et al, 1990;1991;Diss et al, 1993;Segal et al, 1994a, b;Aubin et al, 1995;Lombardo et al, 1996;Essop et al, 1997;Derkson et al, 1999}). Some have compared results obtained using a consensus FR3 primer with those obtained using FR1 (^{Segal et al, 1994a, b;Aubin et al, 1995;Lombardo et al, 1996;Derkson et al, 1999}) or FR2 (^{Diss et al, 1993;Essop et al, 1997;Derkson et al, 1999}) consensus primers and have shown variable rates of detection, but when the results obtained with different primer sets are combined, the overall detection sensitivity is increased (^{Diss et al, 1993;Segal et al, 1994a, b;Lombardo et al, 1996;Essop et al, 1997;Derkson et al, 1999}), which concurs with our results. The detection sensitivities are much lower in those lymphomas derived from germinal center and post germinal center B cells (MALT, follicle center cell, and high-grade diffuse large cell lymphomas) that exhibit high degrees of somatic hypermutation (^{Diss et al, 1993;Segal et al, 1994b;Aubin et al, 1995;Lombardo et al, 1996;Essop et al, 1997;Ritter et al, 1997;Derkson et al, 1999}).

Table III - Comparison of the published methods and detection sensitivities for clonal rearrangement of the IgH chain gene.

Full table

Various PCR methods for analysis of the IgH gene have been published. Many studies employ semi-nested PCR techniques (^{Trainor et al, 1991;Ramasamy et al, 1992;Diss et al, 1993;Essop et al, 1997;Ritter et al, 1997}) to increase sensitivity but specificity can be reduced. We used a single-round monoplex PCR reaction to increase specificity in combination with a highly sensitive electrophoretic technique. Some studies have only used agarose gels for analysis of PCR products (^{Trainor et al, 1990;Deane et al, 1991;Ramasamy et al, 1992;Segal et al, 1994a, b;Lombardo et al, 1996;Essop et al, 1997}) Table III. Although we used this method to confirm PCR amplification of the correct product, it was impossible to distinguish a monoclonal from a polyclonal product in all but four samples. In other studies nondenaturing polyacrylamide gel electrophoresis with ethidium bromide staining has been used (^{McCarthy et al, 1990;Trainor et al, 1991;Diss et al, 1993;Aubin et al, 1995;Ritter et al, 1997}) with a sensitivity of approximately 1%-10% (^{Aubin et al, 1995}). Although more time-consuming, radioactively labeled nucleotides and denaturing polyacrylamide gel electrophoresis increases the sensitivity to 0.1%-1% (^{Deane and Norton, 1991;Rosenquist et al, 1999}) as in our study. An increased sensitivity is important in PCBCL because the clonal population may be very small. We used a modification of the method employed by Deane et al (^{Deane and Norton, 1991;Chim et al, 1996}), which produces a fingerprint from polyclonal DNA against which a clonal rearrangement can be detected as a discrete band. The polyclonal fingerprint pattern serves as a convenient internal control for monitoring the efficacy of the PCR reaction and this confirmed the presence of amplifiable DNA in all 39 of our cases.

Analysis of fluorescent PCR products on automated sequencers has a similar sensitivity (10^-2-10^-3) (^{Landman-Parker et al, 1996}) to our method and is increasingly replacing other methods of detection. Recently,^{Derkson et al (1999)} compared different PCR-based approaches to assess clonal IgH gene rearrangements in B cell neoplasia. They used FR3/JH consensus primers and compared radiolabeled high resolution fingerprinting with nondenaturing polyacrylamide gel heteroduplex analysis and fluorescently labeled gene scanning using an automated sequencer. A clone was detected in 69% with gene scanning compared with 64% with high resolution fingerprinting and 60% with heteroduplex analysis, suggesting that gene scanning of fluorescent PCR products and automated sequence analysis is the most sensitive method of detection, but the difference was not statistically significant.

In summary, we have been able to detect a monoclonal B cell population in 79% of PCBCL using monoplex radiolabeled PCR and three different sets of consensus primers to amplify the IgH gene with high resolution fingerprint analysis. The leader sequence primers provided the greatest sensitivity. Sequence analysis confirmed hypermutation of FR3 primer binding sequences in cases that produced a false negative result with the FR3 primer. For diagnostic purposes, initial analysis with the FR3 primer may identify a clonal IgH rearrangement but, in cases that prove negative, consensus primers for the leader sequences must be employed.

References

References

Acknowledgments

Fiona J. Child was supported by The Special Trustees of Guy's Hospital and Dermatrust.