Immunophenotypic identification of possible therapeutic targets in paediatric non-Hodgkin lymphomas: a children's oncology group report


  • This work was presented in part at the 2nd International Symposium on Childhood and Adolescent non-Hodgkin Lymphoma in New York, NY, May 18–20, 2006.

Sherrie L. Perkins, MD, PhD, Department of Pathology, University of Utah Health Sciences, 50 N. Medical Dr., Salt Lake City, UT 84132, USA. E-mail:


Immunophenotypic analysis can identify protein epitopes in non-Hodgkin lymphomas (NHL) that may respond to targeted immunotherapies, such as anti-CD20 and anti-CD52. Recent studies suggest additional targets may provide therapeutic benefits in NHL. This study evaluated protein expression of CD25, CD52, CD74 and CD80 in paediatric NHL to determine possible targets for immune-based therapeutic approaches. Patient samples were derived from paediatric NHL clinical trials sponsored by the Children's Cancer Group (CCG, now the Children's Oncology Group, COG) and included Burkitt lymphoma (BL), diffuse large B-cell lymphoma (DLBCL), disseminated T- and B-cell lymphoblastic lymphoma (T-LBL and B-LBL) and anaplastic large cell (ALCL). Immunophenotypic studies were performed on formalin-fixed, paraffin-embedded diagnostic tissues. CD25 was expressed in 8% of T-LBL and 75% of ALCL cases, but not in BL, DLBCL, or B-LBL. CD52 was expressed in 99% of cases of paediatric NHL of all subtypes. CD74 was expressed in 100% of B-LBL, BL and DLBCL, but was absent in ALCL and T-LBL. CD80 was expressed in 12% of B-LBL, 6% of BL and 10% of DLBCL cases studied, but was not detected in T-cell NHL. These expression patterns suggest that CD25, CD52 and CD74 may represent potential new therapeutic targets in paediatric NHL.

Paediatric non-Hodgkin lymphomas (NHL) generally include four major subtypes of disease: Burkitt and atypical Burkitt lymphomas (BL), diffuse large B-cell lymphomas (DLBCL), precursor T- or B-cell lymphoblastic lymphomas (T-LBL or B-LBL) and anaplastic large cell lymphomas (ALCL) (Sandlund et al, 1996; Cairo et al, 2006). These subtypes have been well described in currently utilised lymphoma classifications, such as the World Health Organization (WHO) (Jaffe et al, 2001) and the revised European and American lymphoma (REAL) (Harris et al, 1994) classifications. Using current schemes of multi-agent chemotherapy, children have excellent responses to therapy with cure rates in excess of 90% for limited stage mature B-cell lymphomas (such as BL and DLBCL) (Reiter et al, 1999; Patte et al, 2001, 2003; Cairo et al, 2002, 2003a; Gerrard et al, 2003; Cairo et al, 2006) and from 80 to 85% for non-leukaemic lymphoblastic disease (Meadows et al, 1989; Link et al, 1997; Reiter et al, 2000; Cairo et al, 2006). ALCL often shows initial response to therapy in >80% of patients, but relapse is a significant clinical problem (Sandlund et al, 1994, 1996; Seidemann et al, 2001; Perkins et al, 2003; Cairo et al, 2006). Most therapeutic approaches in paediatric NHL have not utilised immune-based therapies as a part of the initial treatment, in contrast to the relatively common use of immune-based therapies in adults, where response and cure rates are lower.

Immunophenotypic analysis of NHL on paraffin-embedded tissues continues to expand to include both phenotyping to define tumour cell type and diagnosis as well as identification of prognostic markers and possible therapeutic targets for immunotherapy. Perhaps the best known target for immune-based therapy is the CD20 antigen found on most mature B-cell NHL that provides a target for an anti-CD20 monoclonal antibody, rituximab [for review (Maloney, 2005)]. However, a variety of other lymphoma cell antigens that may provide targets for immune-based therapy are being identified as new drugs become available. These antigens include B-cell specific antigens (such as CD20, CD22 and CD74), T-cell antigens (CD25) and less specific antigens, such as CD80, CD40 and CD52, that may be expressed by both B- and T-cell NHL (Maloney, 2005).

Immunophenotypic expression patterns of NHL in children have not been as extensively studied as in adult disease, and there is a relative lack of immune-targeted therapies used in paediatric patients. We have previously reported ≥95% expression of CD20 and CD22 in paediatric mature B-cell NHL (Perkins et al, 2003). Although immune-based approaches have not been used as front-line therapy in children, based on our previous results, we have initiated a clinical investigation of rituximab in combination with the French–American–British/Lymphome malins de Burkit (FAB/LMB96) chemotherapy protocol in children with advanced mature B-cell NHL [Children's Oncology Group (COG) ANHLO1P1 trial] (Goldman et al, 2006). Furthermore, immune target therapies may have an important role in second line or salvage paediatric NHL therapies. The present study aimed to document and identify possible immune targets in the cohort of cases that represent the major subtypes of paediatric NHL.


Patient selection

All cases were derived from children entered on co-operative clinical trials through COG (formerly Children's Cancer Group or CCG). Patient samples were chosen at random from the CCG 5961 (Cairo et al, 2003a; Gerrard et al, 2003; Patte et al, 2003) (BL and DLBCL), CCG 5971 (Gaynon et al, 2000; Reiter et al, 2000) (disseminated T-LBL and B-LBL) and CCG 5941 (Abromowitch et al, 2006) (T-LBL and ALCL) paediatric NHL clinical trials sponsored by the CCG, now the COG). The subset of cases studied showed no distinguishing features other than availability of tissue. Appropriate Institutional Review Board approval of the protocol and written informed consent from all patients and/or parents/guardians were obtained by participating institutions prior to patient enrolment.

Pathological materials

Samples were derived from diagnostic tissue from patients enroled in one of three clinical trials. The pathological diagnosis was confirmed as a part of the clinical trial by central pathology review by at least two pathologists with appropriate morphological and phenotypic data to allow for definitive diagnosis of BL (CCG 5961), DLBCL (CCG 5961), precursor B or T-LBL (CCG 5971 and CCG 5941) and ALCL (CCG 5941). All immunophenotypic studies were performed on formalin-fixed, paraffin-embedded tissues derived from diagnostic biopsy materials of paediatric NHL. Immunohistochemistry (IHC) for CD25 and CD52 was performed on a total of 81 cases, including 25 BL, 19 T-LBL, 9 B-LBL, 15 DLBCL and 20 ALCL. IHC for CD74 and CD80 was performed on a total of 66 cases, including 16 BL, 17 B-LBL, 10 DLBCL, 11 ALCL and 12 T-LBL.

Immunohistochemical staining

Specific immune targets of interest for paediatric NHL were tested for using standard immunohistochemical staining techniques and automated staining on the Ventana ES platform (Ventana Medical Systems, Tucson, AZ, USA). The clone, dilution and manufacturer for each antibody are as follows: CD25, 4C9 mouse anti-human monoclonal, 1:300, Novocastra, Norwell, MA, USA; CD52, YTH34.5 rat anti-human monoclonal, 1:600, Serotec, Raleigh, NC, USA; CD74, LN2, 1:320, Novocastra; and CD80, 3H5, 1:30, Pharmingen, Palo Alto, CA, USA. Enzyme-induced epitope retrieval using Protease 2 (Ventana) for 12 min was performed for CD52 staining. Heat-induced epitope retrieval was performed in BORG buffer for CD80 and CD25 (pH 9.5; BioCare Medical, Walnut Creek, CA, USA) or citrate buffer for CD74 (pH 6.0) for 3 min in an electric pressure cooker (BioCare Medical). The Ventana amplification kit was utilised for CD25 and CD80 staining. Signal was detected using the IView diaminobenzidine (DAB) detection kit (Ventana), and slides were counterstained with haematoxylin.

Evaluation of staining

All antibody targets showed appropriate staining of known positive and negative control tissues. Staining of the tumour tissues was evaluated semi-quantitatively: 0 = no staining of tumour cells; 1+ = 1–25%; 2+ = 26–50%; 3+ = 51–75% and 4+ = 76–100% of tumour cells staining. The pattern of staining (nuclear, membranous or cytoplasmic) as well as intensity of staining (weak seen at 400× magnification; moderate seen at 200× magnification and strong seen at 100×) was also noted. Samples were reviewed independently by two pathologists, and discrepant results were resolved by re-evaluation at a multiheaded microscope.


Immunohistochemical staining of sections from formalin-fixed, paraffin-embedded tissues was successfully performed using commercially available antibodies and automated methods following appropriate antigen epitope unmasking by either heat (microwave or pressure cooker) or enzymatic methods. Staining is shown for each antigen tested in a representative case of ALCL (Fig 1A–E), T-LBL (Fig 1F–J), BL (Fig 1K–O), DLBCL (Fig 1P–T) and B-LBL (Fig 1U–Y). Positive immunostaining of tumour cells for CD25, CD74 and CD80 was relatively uniform within each case and between cases, however, the positive staining for CD52 varied among the cases in both extent and intensity. Therefore, the extent of staining and intensity of staining is reported for CD52 but not for the other antigens tested in Table I.

Figure 1.

 Classical type ALCL (A–E) showing large, multinucleated tumours cells [A, Hamatoxylin & Eosin (H&E), 1000×] that shows strong diffuse cytoplasmic and partial membrane expression of CD52 (B, 1000×). CD25 shows membrane and cytoplasmic staining of the tumour cells (C), while CD74 (D) and CD80 (E) show no tumour cell staining. T-LBL (F–J). H&E stain showing the diffuse effacement of nodal architecture by blast-like cells (F, 1000×). The tumour cells show strong, diffuse staining for CD52 (G). The CD25 immunostain shows staining of reactive T cells but no staining of the neoplastic T-lymphoblasts (H, 1000×). CD74 (I) and CD80 (J) showed staining of background lymphocytes but not tumour cells. BL (K–O). K and L demonstrate a BL with moderately intense granular cytoplasmic staining by CD52. (K, H&E; L, CD52, 1000×). CD25 stains non-neoplastic lymphocytes, but the tumour cells are negative (M). CD74 stains >50% of the cells (N, 400×), and CD80 shows weak staining on <50% of tumour cells (O, 1000×). DLBCL (P-T). DLBCL (P, H&E stain) with strong, diffuse cytoplasmic staining with CD52 (Q) but no tumour cell staining for CD25 (R). Staining for CD74 is strong and diffuse (S, 400×), while CD80 is negative (T, 400×). B-LBL (U–Y). B-LBL (U, H&E) showing strong, diffuse staining for CD52 (V) but no tumour cell staining for or CD25 (W). Staining for CD74 is strong and diffuse (×, 400×) and staining for CD80 is weak and diffuse (Y, 1000×).

Table I.   Extent and intensity of CD52 staining in each tumour type. The extent and intensity of staining is reported for each subtype of paediatric NHL.
NHL subtypeExtent (% of tumour cells positive)Intensity
4+ Staining (76–100%)3+ Staining (51–75%)2+ Staining (26–50%)1+ Staining (1–26%)StrongModerateWeak
  1. The extent refers to the percentage of tumour cells that demonstrated staining, and the intensity describes the strength of the staining as explained in the Methods section.

  2. NHL, non-Hodgkin lymphoma; DLBCL, diffuse large B-cell lymphoma; ALCL, anaplastic large cell lymphoma; LBL, lymphoblastic lymphoma.

Burkitt lymphoma2050012/2512/251/25
Precursor-T LBL1621010/195/194/19
Precursor-B LBL61101/94/93/9

CD52 was expressed to some degree in 80/81 (99%) cases of paediatric NHL (Table I). For each tumour type, the majority of positive cases (80% of BL, 84% of T-LBL, 67% of B-LBL, 93% of DLBCL and 93% of ALCL) demonstrated staining in >75% of tumour cells (Table I). The remaining positive cases showed variable staining, ranging from 26 to 75% of the tumour cells expressing the epitope. Most of the cases demonstrated moderate to strong intensity of staining [44/81 (54%) strong, 25/81 (30%) moderate and 21/81 (26%) weak]. Most cases showed cytoplasmic staining, with partial membrane expression detectable in a minority of cases studied.

CD25 was expressed in 1/12 (8·5%) cases of T-LBL and 17/20 (85%) cases of ALCL (Table II). Sixteen of the positive ALCL cases showed strong, cytoplasmic staining and one was weak in intensity. All cases demonstrated staining in at least 75% of the tumour cells. CD25 was not expressed in BL (0/25), DLBCL (0/15), or B-LBL (0/9).

Table II.   Expression of CD25, CD52, CD74 and CD80 in paediatric NHL.
  1. For each subtype of paediatric NHL, the number of cases (and percentage) that stained positively is reported for each target antigen.

  2. BL, Burkitt lymphoma; DLBCL, diffuse large B-cell lymphoma; ALCL, anaplastic large cell lymphoma; LBL, lymphoblastic lymphoma.

CD2517/20 (85)1/12 (8)0/9 (0)0/25 (0)0/15 (0)
CD5213/13 (100)19/19 (100)8/9 (89)25/25 (100)15/15 (100)
CD740/11 (0)0/12 (0)17/17 (100)16/16 (100)10/10 (100)
CD800/11 (0)0/12 (0)2/17 (12)1/16 (6)1/10 (10)

CD74 and CD80 protein expression was evaluated in 11 cases of ALCL and 12 cases of T-LBL, however neither antigen was significantly expressed in any case of either disease (Table II). CD74 was expressed in 17/17 (100%) cases of B-LBL, 16/16 (100%) cases of BL and 10/10 (100%) cases of DLBCL (Table II). The staining was strong in intensity and highlighted greater than 75% of tumour cells in the majority of cases in each disease group. CD80 was expressed in 2/17 (12%) cases of B-LBL, 1/16 (6%) cases of BL and 1/10 (10%) cases of DLBCL (Table II). CD80 staining was weak in each positive case, and the percentage of tumour cells stained ranged from 26 to 75%.


Paediatric NHL appears to be relatively responsive to conventional multi-agent chemotherapy, with cure rates ranging from >90% in non-disseminated mature B-cell lymphomas (Reiter et al, 1999; Patte et al, 2001, 2003; Cairo et al, 2002, 2003a; Gerrard et al, 2003) to >80% in disseminated lymphoblastic lymphomas (Meadows et al, 1989; Link et al, 1997; Reiter et al, 2000). However, some entities, such as ALCL, have less successful long-term results (Sandlund et al, 1994; Brugieres et al, 1998; Seidemann et al, 2001). In addition, those patients with paediatric NHL that fail front-line therapy or relapse have an extremely poor outcome (Anderson et al, 1993; Brugieres et al, 2000; Kobrinsky et al, 2001; Cairo et al, 2003b,c), making identification of effective salvage therapies an important goal. Thus, alternative approaches to treatment, including the use of immune-targeted therapies, are of particular interest.

CD25 is the low affinity portion of the naturally occurring interleukin (IL)-2 receptor. Recently, a fusion protein composed of diphtheria toxin fragments and IL-2 (denileukin diftitox) has been developed for the treatment of cutaneous T-cell lymphomas in adults (Williams et al, 1987; Hesketh et al, 1993). A recent study in adults with relapsed/refractory NHL treated with denileukin diftitox (18 mg/kg/d × 5 d) has demonstrated a 70% overall response rate in subsets of adult NHL (Dang et al, 2004). In the present study, expression of CD25 was identified in the majority of paediatric ALCL cases but was uncommonly expressed in T-LBL or B-cell NHL. These findings suggest that therapeutic agents directed at CD25 may have a role in the treatment of paediatric ALCL, but a role for this in T-LBL or B-cell lymphomas appears less likely (Satwani et al, 2004).

CD52 has been identified in high antigen density in both neoplastic and non-neoplastic B- and T-cells (Treumann et al, 1995; Hale, 2001), and a humanised monoclonal antibody has been developed that targets a CD52 epitope (Dyer et al, 1989). This antibody, alemtuzumab, induces antibody-dependent cellular cytotoxicity, complement-dependent cellular cytotoxicity and apoptosis, and has been approved for treatment of CLL in adults and as a relapse agent in both B- and T-cell adult lymphomas (Osterborg et al, 1997; Keating et al, 2002a,b). Our results demonstrated that nearly all cases of paediatric BL, T-LBL, B-LBL, DLBCL and ALCL express CD52 by IHC. These findings differ somewhat from a recent study that reported CD52 expression in 75% of BL, 7% of T-LBL, 89% B-LBL, 75% DLBCL and 0% of ALCL (Rodig et al, 2006). However, it should be noted that disparate staining results have been noted with CD52, with a recent review article noting almost ubiquitous staining of T-cell malignancies with CD52 (Dearden, 2006). These differences may be related to differences in the patient populations studied and/or immunohistochemical techniques utilised, including antigen retrieval methods, choice of antigen, epitope amplification techniques and staining instrumentation. Targeted therapies directed at CD52 may have a role in each of these forms of paediatric NHL, but the implications of the variable extent and intensity of staining are not known. We have currently initiated a trial of alemtuzumab in childhood ALL in second relapse (COG ADVL 0222, personal communication: Anne L. Angiolillo, MD). Further studies correlating antigen expression levels and therapeutic response are clearly warranted in childhood NHL.

CD74 is an epitope of the invariant chain of human leucocyte antigen (HLA) class II. It is expressed in association with HLA-DR and participates in antigen trafficking and loading (Lamb & Cresswell, 1992; Cresswell, 1996). Expression of CD74 is restricted to plasma cells, dendritic cells, thymus, some endothelial cells and B-lymphocytes (Stein et al, 2004). Its expression has been demonstrated on B-cell (Stein et al, 2004) and T-cell (Ong et al, 1999) lymphoma cell lines as well as adult B-cell NHL (Stein et al, 2004). CD74 is expressed on the cell surface and rapidly internalised (Hansen et al, 1996; Ong et al, 1999). We and others have demonstrated that anti-CD74 monoclonal antibodies are rapidly internalised and have been shown to inhibit growth and induce antibody-dependent cellular cytotoxicity in pre-B ALL and B-NHL cell lines (Stein et al, 2004; Roman et al, 2005). Our study demonstrated that CD74 was highly expressed on tumour cells from paediatric patients with B-LBL, BL and DLBCL in all cases tested, which suggests that CD74 may be a possible therapeutic target in these diseases in children. However, CD74 expression was not detected in either T-LBL or ALCL, suggesting possible limited therapeutic utility in T-cell NHL. Future studies are warranted to determine the possible efficacy of anti-CD74 antibodies in poor risk paediatric B-NHL.

CD80 (B7–1) is a membrane-bound co-stimulatory protein for T-cells, which induces proliferation, cytokine production and effector functions (Coyle & Gutierrez-Ramos, 2001). Monoclonal antibodies directed at CD80 inhibit B-cell proliferation and differentiation, and induce apoptosis in B-cell lymphoma cells (Suvas et al, 2002). Furthermore, early studies with monoclonal anti-CD80 (galiximab) have shown clinical efficacy in relapsed/refractory follicular lymphoma in adults (Younes et al, 2003; Czuczman et al, 2005). In the present study, weak expression of CD80 was detected in a small minority of cases of paediatric B-LBL, BL and DLBCL, but was not detected in T-LBL or ALCL. The lack of CD80 staining despite clear expression in non-neoplastic control B-cells may indicate that this epitope is expressed at lower levels in tumour cells, as previously suggested (Dorfman et al, 1997). A different approach to identification of tumours that will be sensitive to anti-CD80 therapy may be warranted as IHC may lack the sensitivity to detect low level expression of the epitope in tumour cells.

In a previous COG study, CD20 and CD22 were expressed in almost all the cases of BL and high-grade B-cell (Burkitt-like) lymphoma, as well as DLBCL (Perkins et al, 2003). The current study showed that CD25, CD52, CD74 and CD80 expression can be evaluated in archival paraffin embedded tissue from paediatric NHL patients using IHC. CD25, CD52 and CD74 appear to be potential targets for immunotherapeutic study in all or more important subsets of paediatric NHL. The expression patterns identified for these markers suggest therapeutic approaches for targeted immunotherapies in paediatric NHL. However, additional studies of antigen intensity and surface expression by flow cytometry are warranted as clinical trials are developed. In this study, all tumour samples came from primary biopsies prior to treatment so we were unable discern antigenic drifts or modulation in response to treatment, an issue that would be worth pursing as an explanation of lack effect of immunotherapy or development of immunotherapy treatment resistance. We are currently testing rituximab in newly diagnosed advanced paediatric mature B-NHL (Goldman et al, 2006), and future studies are warranted with other antibodies or immunotoxins against CD25, CD52 and/or CD74 in poor risk paediatric NHL.


The authors would like to acknowledge the clinical assistance of Erin Morris, R.N. and Donna Correia in the preparation of this manuscript, as well as all of the COG/CCG investigators who submitted specimens used for this research. Dr Miles is supported by the College of American Pathologists Foundation Scholars Research Program.

This study was supported in part by the Pediatric Cancer Research Foundation (MSC), Marisa Fund (MSC), Sonia Scaramella Fund (MSC), Andrew Gargiso Foundation (MSC), ARUP Institute for Research and Development (SLP, ST), Division of Cancer Treatment, National Cancer Institute, National Institute of Health, Department of Health and Human Services (COG, MSC, RS, MA and SLP).