Molecular Analysis of Immunoglobulin Genes in Primary Intraocular Lymphoma

Abstract

purpose. To analyze somatic hypermutations in clonally rearranged IgH chain variable (V) genes of primary intraocular lymphoma (PIOL), to identify the differentiation stage of B-PIOL cells.

methods. Sixteen cases of PIOL were diagnosed on the basis of morphology and immunohistology. In six patients, simultaneous cerebral lymphomatous involvement was known; stereotactic biopsy specimens were investigated in three cases. A polymerase chain reaction (PCR) was performed on DNA extracted from vitrectomy specimens or from paraffin-embedded sections, to amplify the clonally rearranged heavy-chain immunoglobulin (IgH) genes. The isolated PCR products were sequenced and compared with published VH germ-line segments to determine the VH usage and number of somatic mutations in the complementarity-determining region (CDR)2 and framework region (FR)3.

results. All tumors exhibited clonal IgH rearrangements. Of the eight sequenceable cases, four had the VH4-34 gene segment, two the VH3-23, one the VH3-7, and another the VH3-30. The pattern of somatic mutations indicated selection of PIOL cells for expression of a functional antibody. The mean frequency of somatic mutations detected for the IgH gene was very high (14.5%). In three oculocerebral lymphomas, the identical B-cell clone was demonstrated in ocular and cerebral tissues.

conclusions. The data suggest that PIOLs (1) are derived from mature B-cells that have undergone the germinal center reaction and (2) are closely related to primary cerebral nervous system lymphoma (PCNSL), due to their high mutation frequency of VH region genes. The close relationship between PIOL and PCNSL is underlined by demonstration of the same VH segment (VH4-34) in three of six cases of oculocerebral lymphoma.

Primary intraocular lymphoma (PIOL) is a high-grade malignant non-Hodgkin lymphoma (NHL) usually of B-cell type, which affects the retina, vitreous and/or the optic nerve. ¹ PIOL is considered a subtype of primary central nervous system lymphoma (PCNSL), and when occurring simultaneously with PCNSL, it is termed oculocerebral lymphoma. Most PIOLs are diffuse large B-cell lymphomas (DLBCL), according to the World Health Organization (WHO) Classification for Hematologic and Lymphoid Neoplasms. ² In general, PIOL has become more frequent; there has been a threefold increase in the incidence of PCNSL since 1970 in both the immunocompromised and immunocompetent population. ^{3 4} PIOL usually presents in patients in the fifth or sixth decade as a nonspecific corticosteroid-resistant uveitis. It remains a diagnostic dilemma, with delays being common and multiple specimens often being required before unequivocal establishment of the diagnosis. ^{5 6 7 8 9 10} Compared with systemic DLBCL, PIOL—particularly the oculocerebral form—is associated with a poor prognosis, with a 2-year overall survival of 39%. ¹¹  

To define the B-cell differentiation stage of PIOL tumor cells a molecular study of immunoglobulin (Ig) gene rearrangements in association with sequence analysis of rearranged variable (V) region genes is a suitable tool. The generation of functional antigen receptors by somatic recombination of the immunoglobulin variable (V), diversity (D), and joining (J) gene segments is a unique feature of B-lymphocytes. ¹² In a normal immune response, B lymphocytes stimulated by antigen can undergo somatic hypermutation in their immunoglobulin (Ig) variable region genes in the germinal centers. ¹³ Somatic hypermutation includes point mutations in the nucleic acids coding for the rearranged IgH chain V regions and results in an adapted antigen affinity of antigen-specific antibodies. ^{14 15 16} B-cells with high affinity for antigen are selectively rescued from apoptosis, which occurs at a very high rate in the germinal centers, to differentiate into either circulating memory cells or antibody-producing plasma cells. ^{13 14} B-cell lymphomas arise from cells that represent a certain differentiation stage when neoplastic transformation took place. Analysis of the IgVH region genes has provided invaluable information on the status of tumors in B-cell differentiation and has added another dimension to classification of B-cell tumors at the molecular level. For example, naive pre–germinal-center B cells carry nonmutated VH genes, whereas germinal-center B cells and germinal-center–derived memory cells harbor mutated VH region genes. ^{14 15 17 18} Correspondingly, B-cell malignancies descended from pregerminal B cells (e.g., mantle cell lymphoma ¹⁹ ), and those from post–germinal-center B cells (e.g., marginal zone B-cell lymphomas ^{20 21} ), usually demonstrate few or several somatic mutations (up to 5%), respectively. ²² Because they arise from germinal center cells, follicular lymphomas usually demonstrate a high rate of somatic mutations (average, 10.5%). ²² Typical also for follicular lymphomas are so-called “ongoing mutations” or the continued accumulation of mutations, resulting in intraclonal heterogeneity, indicating that the mutation mechanism is still active within the tumor cell population. ²³  

Additional information can be obtained from the analysis of the VH genes with regard to a putative antigen selection process before or during lymphoma development. In the absence of negative or positive selective pressure, a random mutational process of genes would result in an even distribution of nucleotide changes yielding amino acid replacements (R mutations) and nucleotide changes not yielding amino acid replacements (S, or silent mutations) throughout the coding sequence. Antigen-selected antibodies, however, have been shown to include a higher frequency of R mutations in the IgV-gene complementarity-determining regions (CDRs)—thought to constitute the antigen-binding site ²⁴ —than in the framework regions (FRs), where the proportion of S mutations may be greater. The high R-to-S mutation ratio characteristic of affinity mature IgV-gene CDR sequences is consistent with their products’ being under positive pressure to mutate to provide the best-fit antigen. In contrast, the low R-to-S mutation ratio characteristic of the FR sequences is consistent with their need to be preserved in structure, to provide the skeleton for the antigen-contacting CDRs (so-called counterselection). ^{25 26}  

Finally, analysis of VH gene segments used by lymphomas and a comparison with germ-line segment banks allows for insight with regard to an overrepresentation of one or more of the seven germ-line segment families (VH1–VH7). For example, a biased V4-34 gene of the VH4 family has been observed in PCNSL, ²⁷ whereas an overrepresentation of VH gene families in peripheral DLBCL remains controversial. ²⁸ A preferential VH-gene usage may implicate a role of a superantigen, which may drive B-cell proliferation by binding to the surface Ig receptor via an unmutated FR. 

Previous reports have described the use of PCR examining for monoclonal rearrangements of immunoglobulin heavy (IgH) or light (IgL) chains in B-cell lymphoma or T-cell receptor genes in T-cell lymphoma as an adjunctive diagnostic tool in the evaluation of vitreous specimens for PIOL. ^{6 29 30 31 32} Recently, we reported the findings of an analysis of the somatic mutations in the VH region of a patient with oculocerebral lymphoma. ³³ To gain further insight into the nature of the IgVH genes in PIOL, we investigated 16 cases of PIOL using IgH-PCR. It was possible to sequence the VH genes of eight of these cases. We have found a frequent usage of VH4 and VH3 gene segments and observed that the VH genes in PIOL are extensively mutated in most cases. 

Materials and Methods

Patients and Samples

Sixteen cases of PIOL were taken from the files of the Department of General Pathology and Reference Center for Hematopathology, Charité-University Medicine Berlin. These cases consisted of nine chorioretinal biopsy specimens and seven vitrectomy specimens obtained via pars plana vitrectomy (PPV). CNSL involvement was seen in 6 of the 14 cases; in 3 cases, paraffin-embedded blocks of the cerebral biopsy specimens were also examined. The specimens had been fixed in 4% buffered formalin and embedded in paraffin. Subtyping of the lymphomas followed conventional histologic staining (hematoxylin and eosin, Giemsa, and periodic-acid Schiff) and immunohistochemistry, according to the WHO Classification for Hematologic and Lymphoid Neoplasms. ² Cytospin specimens of the unfixed (n = 6) and HOPE (HEPES-glutamic acid buffer-mediated organic solvent protection effect)-fixed (n = 3) ³⁴ vitrectomy specimens were obtained by spinning the vitreous biopsy specimens at 500 rpm for 5 minutes and concentrating the cells onto glass slides. These were subsequently air dried and stained using conventional stains (e.g., May-Grünwald-Giemsa) and immunocytology. 

Immunocytology and Immunohistochemistry

The primary monoclonal antibodies used for immunocytology included CD79a and CD3 (Table 1) . Polyclonal antibodies were used to test the expression of the Ig chains κ and λ. The visualization of the antibodies was obtained using the alkaline phosphatase–antialkaline phosphatase (APAAP) ³⁵ and peroxidase–antiperoxidase (PAP) methods. ³⁶  

For immunostaining of paraffin-embedded tissues, sections (4 μm) were cut from paraffin blocks, dewaxed, rehydrated, and subjected to heat-induced epitope retrieval methods before incubation with the appropriate antibodies. Sections were immersed in sodium citrate buffer solution at pH 6.0 and were subsequently heated in a pressure cooker for antigen retrieval. ³⁷ After they were rinsed in running water and Tris-buffered saline, the sections were incubated with the primary antibodies listed in Table 1 . Polyclonal antibodies were applied to examine the expression of IgH and IgL. The antibodies were made visible with an indirect immunoperoxidase method, ³⁶ for the antibodies to the heavy and light chains, whereas the APAAP method was used to demonstrate the binding of the remaining antibodies. ³⁵ Appropriate negative and positive tissue control experiments were performed with each investigation. 

The tumor cells of PIOLs were considered (1) positive (+) when >75% of the tumor cells were immunoreactive for the examined antibody; (2) positive/negative (+/−) when 50% to 74% of the tumor cells were immunoreactive; (3) negative/positive (−/+) when 25% to 49% of the tumor cells showed immunoreactivity; and negative (−) when <24% of the tumor cells demonstrated immunoreactivity. The number of Ki-67-positive cells was determined by counting the number of cells with clear nuclear positivity for this marker per 100 cells in three high-power fields (HPFs; 40× objective, BH2; Olympus, Tokyo, Japan). The resultant mean of the three HPFs is expressed as a percentage, representing the growth fraction of the tumor. 

Clinical outcome data were available for all samples (Table 2) . An informed consents were obtained according to the Declaration of Helsinki. The research was approved by the institutional review board. 

PCR Amplification Method

DNA was extracted after dewaxing of 20-μm-thick paraffin-embedded sections, as well as from the centrifuged vitrectomy specimens, with a DNA minikit (QIAmp; Qiagen, Hilden, Germany) according to the manufacturer’s recommendations. The IgH PCR was performed with the primer sets FR1, FR2, and FR3 (BioMed-2) in three separate single-step PCRs without reamplification, each consisting of 50 cycles (96°C for 15 seconds, 60°C for 40 seconds, and 72°C for 60 [FR1], 45 [FR2], and 30 [FR3] seconds), as described previously. ³⁸  

Reactive tonsils and reactive vitritis specimens were used as the polyclonal control, whereas DNA from the B-cell lines Raji and Daudi served as the monoclonal control. Sufficient negative control samples were included. Reproducible dominant PCR products of the same size indicated monoclonality. The PCR products were analyzed on an automated DNA sequencer (model 310A; Applied Biosystems, Inc. [ABI], Foster City, CA) to allow a highly precise dose-by-dose separation of the amplificates. For this purpose, fluorescence (FAM)-labeled JH (JH22) oligonucleotides were used. 

DNA Sequence Analysis

For DNA sequencing of the dominant IgH PCR products, a second PCR was performed in which the fluorescence-labeled JH22 primer was replaced by an unlabeled primer of the same sequence. The resultant unlabeled amplificates were isolated after gel electrophoresis by diffusion into distilled water overnight. The isolated PCR products were directly subjected to DNA sequencing with the two respective IgH PCR primers in two separate sequencing reactions. The sequencing reactions were run on an automated DNA sequencer (model 377; ABI) and analyzed. Sequences were compared with published VH germ line sequences with the help of the IMGT blast search (http://imgt.cines.fr:8104). Codons were numbered according to Kabat et al. ³⁹ The putative antigen selection in the tumor cell IgH rearrangements was determined by the ratio of replacement to silent mutations (R/S) in the CDR and FR regions. A rearrangement was considered to be antigen selected when the R-to-S ratio in the CDR2 region was higher than 2.9 and the R-to-S ratio in the FR3 region was lower than 1.5. ⁴⁰  

Results

Patient Group

The patient group examined consisted of eight women and eight men with an age range of 40 to 84 years (mean, 65.0; Table 2 ). The clinical symptoms, treatment, and current status of the patients are summarized in Table 2 . In seven patients, a vitreous aspirate only was performed to reach diagnosis. In eight patients, both a vitreous aspirate and a chorioretinal biopsy were performed within the same surgical procedure. In one patient, a chorioretinal biopsy only was performed (Table 3) . 

Conventional Histology and Immunohistochemistry

The 15 vitreous aspirates consisted of mature inflammatory cells, medium to large neoplastic lymphocytes, and lytic cells. The cells expressed the B-cell antigens CD79a, CD20, and PAX5/BSAP. Conventional histology of the nine chorioretinal biopsy specimens demonstrated an infiltrate of atypical lymphocytes in the retina and adjacent choroid. The neoplastic cells were medium to large, with basophilic cytoplasm, oval-shaped nuclei, and conspicuous nucleoli (Fig. 1) . The tumor cells were positive for the above-mentioned B-cell antigens as well as for BCL2 (partial), BCL-6 (partial) and MUM1/IRF4 (Fig. 1) . The tumor cells of three PIOLs expressed only CD10 (Table 3) . A monotypical expression of an IgL and/or IgH was demonstrated in most cases, and a large growth fraction (average, 84%) using the MIB1 antibody was determined (Fig. 1 ; Table 3 ). The reactions against CD30 and Epstein-Barr virus (LMP1) were negative. 

The three examined stereotactically obtained cerebral biopsy specimens (patients 8, 9, and 10) displayed diffuse and perivascular neoplastic infiltrates consisting of medium-sized blasts (Fig. 1D) . These showed an immunophenotype similar to that of the respective PIOL infiltrates. 

PCR and Sequence Analysis

In eight of the 16 PIOL samples, sufficient amount of PCR products was obtained allowing for sequencing. In the samples where monoclonal products were obtained with FR3 only, no sequencing was performed because of limited VH sequence information. In three of six PIOL cases in which oculocerebral lymphoma was known (patients 8, 9, and 10), identical clonal B-cell populations were demonstrated (Figs. 2A 2B) . Furthermore, in one patient, it was possible to demonstrate the identical B-cell clone in vitreous specimens taken from both eyes and in the cerebral biopsy (patient 9; Figs. 2B 2C ). 

Results of the sequence analysis are summarized in Table 4 . The lymphoma VH gene segments showed the highest homology to the germ-line genes of the VH4 and VH3 gene families. While the VH4-34 (also called HV-21 or DP63) gene of the VH4 gene family was used in four cases, the VH3-23 gene segment was used in two cases, the VH3-7 and the VH3-30 gene in one case each, respectively (Table 4) . A high frequency of somatic mutation, ranging from 16 to 65 (average 38.57) was seen in PIOL. The mean mutation frequency was 14.5% (range, 6.5%–24.3%) of the VH genes. 

The R-to-S ratio in FR1-3 was calculated for all in-frame IgH genes. This parameter indicates whether the corresponding B lymphocyte was selected for expression of a functional antibody. The R-to-S ratios for the FRs were determined as 1.3 (range, 0.2–2.3) for the IgH genes. According to this calculation, the ratios clearly indicate that the tumor cells or their precursors were selected for antibody expression, because selected B cells usually show R-to-S ratios below 1.5 for the FRs. ⁴⁰  

Discussion

The purpose of the present study was to investigate the differentiation stage of the neoplastic B-cells of PIOL, which develop in an environment that normally contains a very low number of lymphocytes and actively suppresses the proliferation of lymphocytes. To investigate, we performed IgH-PCR on tumor samples from 13 patients. All PIOLs of our series carried monoclonally rearranged immunoglobulin heavy chain (IgH) genes. The DNA amplificates of eight cases could be sequenced and analyses demonstrated the introduction of numerous somatic mutations into the rearranged VH genes. These data suggest that PIOLs are derived from neoplastic B cells that have undergone a prolonged interaction in the microenvironment of the germinal center. 

In the majority of PIOLs an intermediate to large number of mutations in the VH region, ranging from 16 to 65 (average, 37.38; 14.5%) were observed (Table 4) . This average frequency significantly exceeds that of normal B cells (2%–3%) ²² and that of most B-cell lymphomas (average, 7%). ²² A similar high mutation frequency was reported for VH region genes in PCNSL, ^{27 41 42 43 44} but differs from systemic DLBCL, the same histologic subtype as PIOL, which demonstrates variable mutation frequencies. ^{28 45} The reasons for the differences between PIOL/PCNSL and systemic DLBCL remain unclear. The current findings of a high frequency of somatic mutations in the VH genes of PIOLs, together with the tumor cell phenotype (PAX5/BSAP⁺, CD20⁺, BCL-6^+/−, CD10^+/−, MUM1/IRF4⁺) ^{6 46} provide further evidence for the hypothesis that PIOLs are derived from mature B cells that have undergone a prolonged interaction in the microenvironment of the germinal center and are either at the late germinal center stage of differentiation or are early post–germinal-center B cells. This may explain the extensive loss of CD10 and the variable expression of BCL-6, ⁴⁶ both of which are normally expressed by germinal-center cells but not by post–germinal-center lymphocytes, as well as the positivity of MUM1/IRF4 (post–germinal-center marker) in PIOLs. 

A limited germ line VH gene usage was seen in the development of PIOLs, with preferential expression of the VH3 and VH4 germ-line families. In particular, a biased inclusion of the VH4-34 gene segment by PIOLs became apparent. The VH4-34 gene is present in 4% to 7% of normal adult peripheral B cells in healthy adults ⁴⁷ and appears to be overexpressed in some autoimmune diseases and B-cell lymphomas. ^{48 49} For example, the VH4-34 gene has been demonstrated to be mandatory for encoding the IgM proteins of cold agglutinin disease. ⁵⁰ In this case, the red blood cell I/I antigens bind to the FR1 of the immunoglobulin. This binding outside the CDRs is an indication of a B-cell superantigen. Preferential use of the VH4-34 gene has been reported in PCNSL, ⁴³ as well as in other lymphoma entities, such as high-grade mucosa-associated lymphoid tissue lymphoma, mantle cell lymphoma, and Burkitt lymphoma. ^{19 51 52 53} Overrepresentation of the VH4-34 gene in peripheral DLBCL is controversial, however, with one study suggesting a biased presence of this gene, ⁵⁴ but the second not confirming the finding. ^{28 45}  

The relatively high usage frequency of the VH4-34 gene in PIOLs suggests a functional role of the Ig encoded by this gene in its development. Either of two mechanisms may underlie its biased usage: an antigen-driven expansion or a superantigen expansion of VH4-34-encoded Ig, which may have autoreactive properties. ^{48 50 53} The notion of an antigen-selected maturation of the tumor cells is supported by the R-to-S ratio’s being >2.9 in the CDR regions and <1.5 in the FR3 region, which reflects counterselection of R mutations in the FRs and indicates that the tumor cells have been, if only temporarily, antigen selected for expression of a functional antibody. Evidence for such an antigen-selection process has been reported previously in PCNSL. ^{41 43}  

At present, the process of somatic hypermutation and antigen selection is believed to occur exclusively in germinal centers in peripheral lymphatic tissues (e.g., tonsil, mucosa-associated lymphatic tissue, and lymph nodes) only. Because intraocular and cerebral tissue is devoid of an organized germinal center structure, our data and others’ findings ^{41 43 55} imply that PIOLs and PCNSLs are tumors that arise initially in an extraneural germinal center environment. Subsequent homing of the neoplastic B cells to the retina and/or to the central nervous system (CNS) may involve the development of a neurotropic cellular phenotype, or it may be a result of the influence of chemokine receptors and their ligands. ^{56 57} In one of the examined cases (patient 9), there was evidence of intraclonal heterogeneity in the recurrent B-cell clone occurring in the contralateral eye, indicating that the mutation mechanism is possibly still active in the PIOLs/PCNSLs in this patient (Fig. 2C) . One may speculate that an antigen within the intraocular environment or in the CNS together with T cells and antigen-presenting cells supports ongoing mutations in these milieu. Although several viruses with the capacity to persist in cerebral and retinal tissue (e.g., polyomavirus and herpes viruses) ⁵⁸ have been thought to play a role in lymphomagenesis, an antigen has not been identified to date. Further research is necessary at the molecular level of PIOLs/PCNSLs to understand their pathogenesis, possibly with the purpose of designing therapeutics that exploit the apparent immunoglobulin gene usage, perhaps in the form of engineered anti-idiotypic antibodies or other compounds that specifically target autoantibody-producing B cells. 

Finally, in oculocerebral lymphoma, it is assumed on the basis of clinical, morphologic, and immunohistochemical findings that the cerebral and ocular infiltrations represent the same tumor. Recently, our group provided molecular biological evidence for the first time in the literature that the lymphomatous manifestations in oculocerebral lymphoma consist of the identical neoplastic B-cell population and that they derive from the same tumor precursor cell. ³³ In this study, we were able to demonstrate that the ocular and cerebral lymphomatous manifestations were derived from the same neoplastic clones in two additional oculocerebral cases of lymphoma (Fig. 2) . These results underline the close relationship between PIOL and PCNSL. 

In summary, the present study demonstrates that most PIOLs, similar to cerebral counterparts, are derived from mature B cells that have undergone the germinal center reaction. This detailed characterization of the molecular phenotype of PIOL could provide further insight into their development, possibly having an influence on their future treatment strategies. 

 Submitted for publication March 30, 2005; revised June 3, 2005; accepted August 17, 2005.

 Disclosure: S.E. Coupland, None; M. Hummel, None; H.-H. Müller, None; H. Stein, None

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Table 1.

 

Antibodies with Clones and Derivations Employed in the Investigations

Antibody	Clone	Company/Source	Cell Population(s)
CD79a	JCB117	DakoCytomation, Denmark	B-lymphocytes, plasma cells
CD20	L26	DakoCytomation	B-lymphocytes
BSAP/PAX5	2h	BD Biosciences-Pharmingen, San Diego, CA	Nonplasmacellular differentiated B-cells
CD3	F7.2.38	DakoCytomation	T-lymphocytes
CD10	56C6	Novocastra, Newcastle-upon-Tyne, UK	B-lymphocytes (centrocytes, centroblasts); granulocytes
BCL-2	124	DakoCytomation	B-lymphocytes (mantle and marginal zone B-cells); plasma cells; T-cells; melanocytes
BCL-6	PG-B6p	DakoCytomation	B-lymphocytes (centrocytes, centroblasts); extrafollicular blasts
1RF4	MUM1p	DakoCytomation	Plasmacellular differentiated B-cells; T-cells; melanocytes
IgL	Polyclonal	DakoCytomation	Plasma cells, plasmacellular differentiated B-cells
IgH	Polyclonal	DakoCytomation	Plasma cells, plasmacellular differentiated B-cells
MIB-1	Ki67	DakoCytomation	Proliferating cells

Table 2.

 

Clinical Details of the Investigated PIOLs

Case	Age	Gender	CNS Involvement	Period between CNS and Ocular Lymphoma Manifestations	Symptoms and Signs	Treatment	Current Status
1	55	M	Yes	8 Months(initial PCNSL)	Visual disturbances; decreased vision	HD MTX; Radiotherapy; HD-Ara-C	Deceased
2	71	F	Yes	20 Months (initial PCNSL)	Recurrent visual disturbances OD	Intraocular Triamcinolon; HD MTX; Topotecan	Alive in PR
3	82	F	No	Not applicable	Decreased vision OD; steroid-resistant uveitis	Radiotherapy	Deceased
4	76	F	No	Not applicable	Unilateral steroid-resistant uveitis	Radiotherapy	Deceased
5	73	M	No	Not applicable	Steroid-resistant uveitis	HD MTX	Deceased
6	77	M	No	Not applicable	Bilateral visual disturbances; vitritis; subretinal infiltrates	CHOP	Deceased
7	83	F	No	Not applicable	Pain OD; decreased vision	Trofofosfamid	Deceased
8	65	M	Yes	24 Months(initial PCNSL)	Bilateral vitreous infiltrates	MTX	Alive in CR
9	52	M	Yes	17 Months (LA); 27 months (RA) (initial PCNSL)	Decreased vision OS; followed 10 months later with decreased vision OD	HD MTX; Ifofosfamid	Alive in PR
10	81	F	Yes	35 Months(initial PCNSL)	Decreased bilateral vision	HD MTX	Alive in CR
11	76	M	No	Not applicable	Decreased bilateral vision; steroid-resistant uveitis	Trofofosfamid	Alive in CR
12	73	F	No	Not applicable	Decreased unilateral vision	Ifofosfamid	Alive in CR
13	71	M	Yes	20 Months(initial PCNSL)	Decreased vision; steroid-resistant uveitis	Radiotherapy	Alive in PR
14	84	F	No	Not applicable	Decreased vision	HD MTX	Alive in PR
15	58	M	No	Not applicable	Hypopyon; secondary glaucoma; pain	HD MTX	Alive in PR
16	40	F	No	Not applicable	Bilateral uveitis	HD MTX	Alive in PR

 HD MTX, high-dose methotrexate; CHOP, cyclophosphamide, hydroxydaunorubicin (adriamycin), vincristine (Oncovin) and prednisone; CR, complete remission; PR, partial remission.

Table 3.

 

Morphological Subtype and Immunophenotype of the Investigated PIOL

Case	Biopsy Source	CD79a	BSAP	CD20	IgM	BCL2	BCL6	IRF4	CD10	CD30	Ki67
1	R	Pos	Pos	Pos	Pos	Pos	Pos	Pos	Neg	Neg	80%
2	V+R	Pos	Pos	Pos	Pos	Pos	Pos	Pos	Neg	Neg	80%
3	V+R	Pos	Pos	Pos	Pos	Pos	Pos	Pos	Neg	Neg	80%
4	V+R	Pos	Pos	Pos	Pos	Pos	Pos	Pos	Neg	Neg	90%
5	V+R	Pos	Pos	Pos	Pos	Pos	Pos	Pos	Neg	Neg	90%
6	V+R	Pos	Pos	Pos	Pos	Pos	Pos	Pos	Neg	Neg	80%
7	V+R	Pos	Pos	Pos	Pos	Pos	Pos	Pos	Neg	Neg	90%
8	V+R	Pos	Pos	Pos	Pos	Pos	Pos	Pos	Pos	Neg	70%
9	V	Pos	Pos	Pos	Pos	Pos	Pos	Pos	Pos	Neg	90%
10	V+R	Pos	Pos	Pos	Pos	Pos	Pos	Pos	Neg	Neg	80%
11	V	Pos	NP	Pos	Pos	NP	NP	NP	NP	Neg	80%
12	V	Pos	NP	Pos	Pos	NP	NP	NP	NP	Neg	90%
13	V	Pos	NP	Pos	Pos	NP	NP	NP	NP	Neg	80%
14	V	Pos	Pos	Pos	Pos	Pos	Pos	Pos	Pos	Neg	80%
15	V	Pos	NP	Pos	Pos	NP	NP	NP	NP	Neg	90%
16	V	Pos	NP	Pos	Pos	NP	NP	NP	NP	Neg	90%

 All cases were WHO subtype DLBCL. V, vitreous biopsy; R, chorioretinal biopsy; V+R, combined vitrectomy and chorioretinal biopsy; NP, not performed (due to limited tissue).

Figure 1.

 

(A) A cytospin specimen of the vitrectomy specimen from patient 10 showing blasts with large nuclei and prominent nucleoli (MGG staining). (B) CD20-positive tumor cells with erythrocytes in the background (APAAP staining). (C) Neoplastic B cells demonstrating a large growth fraction (APAAP staining, Ki-67 antigen). (D) Histologic section of a stereotactic biopsy specimen of the corresponding cerebral lymphoma in patient 10, demonstrating the typical perivascular infiltration of the lymphoid blasts (MGG staining). Original magnification: (A, D) ×40; (B) ×60; (C) ×20.

Table 4.

 

Details of the Sequencing of the Investigated PIOLs

Case	Primer Combinations Determining Monoclonality	Monoclonal Product Size (bp)	VH Gene*	Total Mutations (n)
1	BioMed-2 FR3 only	135	Not sequenceable	NA
2	BioMed-2 FR3 only	147	Not sequenceable	NA
3	BioMed-2 FR1; FR2; FR3	313; 247; 106	VH4-34/JH1	41
4	BioMed-2 FR3 only	125	Not sequenceable	NA
5	BioMed-2 FR1; FR2; FR3	321; 253; 118	VH3-23/JH5	31
6	BioMed-2 FR3 only	132	Not sequenceable	NA
7	BioMed-2 FR3 only	127	Not sequenceable	NA
8	BioMed-2 FR3 only	122	Not sequenceable	NA
9	BioMed-2 FR1; FR2; FR3	327; 257; 125	VH3-23/JH4	16
10	BioMed-2 FR1; FR2; FR3	319; 250; 109	VH4-34/JH6	26
11	BioMed-2 FR1; FR2; FR3	340; 270; 106	VH4-34/JH4	44
12	BioMed-2 FR1; FR2; FR3	333; 257; 146	VH3-30/JH5	47
13	BioMed-2 FR1; FR2; FR3	311; 232; 110	VH3-7/JH5	65
14	BioMed-2 FR3 only	147	Not sequenceable	NA
15	BioMed-2 FR3 only	125	Not sequenceable	NA
16	BioMed-2 FR1; FR2; FR3	311; 244; 103	VH4-34/JH5	29

 FR, framework; na, not applicable.

*  JH are the germ-line genes to which the lymphoma JH genes showed the highest homology.

Figure 2.

 

(A, top) Gene expression analysis after IgH-PCR (BioMed-2 FR1) of a vitreous biopsy specimen (C808-04) of patient 10, demonstrating a monoclonal 319-bp PCR product (blue). The smaller red peaks represent the control. (Bottom) Expression analysis after IgH-PCR (BioMed-2 FR1) of the corresponding paraffin-embedded cerebral biopsy specimen (H13700-04), with a monoclonal 319-bp PCR product. (B, top) Expression analysis after IgH-PCR (BioMed-2 FR1) of the vitreous biopsy specimen of the left eye (C1215-03) of patient 9, demonstrating a monoclonal 327-bp PCR product (blue). (Middle) Gene expression analysis after IgH-PCR (BioMed-2 FR1) of the corresponding paraffin-embedded cerebral biopsy (H21207-03), with a monoclonal 327-bp PCR product. (Bottom) Expression analysis after IgH-PCR (BioMed-2 FR1) of the vitreous biopsy specimen of the recurrent tumor in the right eye (C896-04), with a monoclonal 327-bp PCR product. The smaller red peaks represent the control. (C) Comparison of the CDR3 regions of the DNA sequence obtained from the vitreous of the left eye (C1215-03) of patient 9, with those obtained from the cerebral biopsy specimen (H21207-03) and from the tumor recurrence in the contralateral eye (C896-04). Top row: germ-line sequence with the greatest homology to the neoplastic DNA sequences, VH3-23. The nucleotide substations are marked in green. It can be seen that further single-nucleotide substitutions have occurred in the recurrent tumor (e.g., third row, fourth sequence between 82 and 126 bp; fourth row, fourth sequence between 127 and 161 bp). These data support the notion that ongoing mutations are occurring in the tumor clone.