Immune response to the ALK oncogenic tyrosine kinase in patients with anaplastic large-cell lymphoma

In recent years, there has been renewed interest in the topic of tumor antigens since the immune recognition of such tumor-associated antigens has opened up the prospect of new therapeutic strategies.1,2 Examples of such tumor-associated antigens recognized by both the cell-mediated and humoral arms of the immune system are Her-2/neu or c-erbB-2, MAGE-1, NY-ESO-1, and immunoglobulin idiotypes.3-9 

Anaplastic lymphoma kinase (ALK)–positive anaplastic large-cell lymphoma (ALCL) or “ALKoma” has been recently identified as a distinct lymphoma entity of T- or null-cell origin.10 This tumor is associated with the (2;5) translocation that results in the expression by the neoplastic cells of a hybrid oncogenic tyrosine kinase consisting of the amino terminus of the nuclear protein nucleophosmin (NPM) fused to the intracytoplasmic region of the ALK receptor tyrosine kinase.11,12 Variants of the (2;5) translocation, fusing ALK to partners other than NPM, are present in a minority of ALKomas.10 

Although NPM is ubiquitously expressed, its partner protein, ALK, is absent from normal cells (with the exception of scattered cells in the central nervous system).13 There is, therefore, a strong possibility that the NPM-ALK protein would be the target of an immune response, and this might account for the paradox that ALK-positive ALCL is highly aggressive histologically but has a better prognosis than other large-cell lymphomas.14-16 It would also open up the possibility of new therapeutic options in cases of this disease that are refractory to conventional treatment.

The present report describes a simple immunocytochemical assay to study the presence of an antibody response to ALK proteins. These results were confirmed by means of an immunochemical technique based on an in vitro kinase assay.

Plasma was obtained from 11 ALK-positive ALCL patients attending the John Radcliffe Hospital, Oxford, UK (4); the Park Hospital, Slough, UK (1); Milton Keynes General Hospital (1); and the Hematology Institute, Perugia, Italy (5). ALCL was diagnosed according to the REAL classification system. The ALK-positive ALCL patients presented with differing stages of the disease, and plasma samples were obtained at varying times after diagnosis (Table 1).

Control plasma samples were collected from 5 normal subjects, 5 carcinoma patients attending the Oncology Clinic at the Churchill Hospital, Oxford, and 3 patients with ALK-negative ALCL attending the Department of Haematology, John Radcliffe Hospital, Oxford (2), and the Department of Internal Medicine, Hospital Merkur, Zagreb, Croatia (1).

The ALCL-derived t(2;5)–positive SU-DHL-1 cell line and the COS-1 monkey epithelial cell line were cultured, and cytocentrifuge preparations were made as previously described.13 

Cryostat sections (6 μm) were cut from fresh tonsil obtained from the Ear, Nose and Throat Department, Radcliffe Infirmary, Oxford. All tissue sections and cytocentrifuge preparations were fixed in acetone for 10 minutes and stored at −70°C.

COS-1 cells were transiently transfected with the use of the diethylaminoethyl (DEAE)–Dextran method17 with one of the following plasmids: pcDNA3-ALK (encoding full-length ALK), pcDNA3-NPM-ALK (encoding NPM-ALK protein), and the pcDNA3 expression vector alone. Cytocentrifuge preparations were made from cells harvested after 72 hours of culture.

Cytocentrifuge preparations of transfected COS-1 cells were incubated with plasma (diluted between 1:100 and 1:6750) for 30 minutes, washed in phosphate-buffered saline, and then incubated with horseradish-peroxidase (HRP)–conjugated rabbit antihuman immunoglobulin (Ig)G (DAKO/AS, Glostrup, Denmark). Transfectants labeled with the use of monoclonal anti-ALK13 followed by HRP-conjugated goat antimouse immunoglobulin (DAKO/AS) were used as a positive control.

The in vitro kinase assay was performed as previously described18 but with the use of 5μL aliquots of human plasma to immunoprecipitate proteins from cytocentrifuge preparations of SU-DHL-1 cells or from tonsil sections.

Circulating antibodies recognizing the NPM-ALK kinase protein were detected in all 11 ALCL patients studied, irrespective of their disease status. Antibodies were detected initially by immunostaining COS-1 cells transfected with NPM-ALKcomplementary DNA (cDNA) (Table 1, Figure1A), and their specificity was then confirmed by an in vitro kinase assay of proteins immunoprecipitated from the t(2;5)–positive SU-DHL-1 cell line. Plasma from 10 patients tested immunoprecipitated an autophosphorylated 80-kd protein, a pattern identical to that obtained with a control monoclonal anti-ALK antibody. No proteins of similar size were observed in the immunoprecipitates from lysates of normal tonsil cells (Table 1, Figure 1B).

The antibodies detected by these techniques could potentially recognize epitopes present in either the NPM or the ALK portions of NPM-ALK (or in both moieties). Patients' plasma was therefore tested against transfected COS-1 cells expressing full-length ALK protein (Table 1, Figure 1A), and in 10 cases positive reactivity indicated the presence of antibodies recognizing the ALK moiety of NPM-ALK. In the remaining patient (number 2), it is possible that anti-ALK antibodies were present at too low a titer to be detected by immunostaining. This result could, however, also be explained by the presence of antibodies recognizing only novel epitopes spanning the NPM-ALK junction and/or epitopes present in the amino terminus of NPM (the region of NPM retained in the NPM-ALK protein).

No immunocytochemical evidence was found for circulating antibodies to NPM-ALK or ALK protein in any of the control subjects (Figure 1A). This was confirmed by the kinase assay in 7 of the control subjects (Figure 1B). However, plasma from one normal subject, one carcinoma patient, and one ALK-negative ALCL patient (13, 17, and 24, respectively) recognized a weakly phosphorylated 80-kd protein, corresponding in size to NPM-ALK (Figure 1B). Autoantibodies to normal NPM have been described previously19 and may account for this weak reactivity with NPM-ALK, but it is also possible that antibodies recognizing NPM-ALK were present at levels too low to detect by the immunocytochemical technique.

Controversy surrounds the possible antitumor effect of antibodies to tumor-associated molecules in patients with malignant disease.20,21 There is, however, evidence that boosting antibody responses to tumor-associated antigens is clinically beneficial in melanomas22-24 and B-cell lymphomas.7-9 25 In the present preliminary study, we could draw no conclusions concerning the clinical significance of anti–NPM-ALK antibodies although it was clear that they differed in titer, with a tendency to higher levels in pretreatment blood samples. However, it is of interest that circulating antibodies to NPM-ALK and ALK proteins persisted in a patient (number 5) who had been in complete remission for more than 9 years. Further studies are therefore necessary to establish whether anti–NPM-ALK titers are relevant to clinical prognosis.

In conclusion, the present study describes for the first time a specific immune response to the NPM-ALK fusion kinase in patients with ALK-positive lymphoma. Whether this contributes to the relatively good prognostic outlook of ALK-positive ALCL requires further study, but the present study also indicates that a simple immunocytochemical technique can be used to detect and quantitate the B-cell immune response to a tumor-associated protein.

Dr R. Kusec, Department of Internal Medicine, Hospital Merkur, Zagreb, Croatia, and Robin Roberts-Gant, Medical Informatics Unit, John Radcliffe Hospital, Oxford, UK.