Koji Kawai MD, Department of Urology, Institute of Clinical Medicine, University of Tsukuba, 1-1-1 Tennodai, Tsukuba-City, Ibaraki 305, Japan. E-mail: firstname.lastname@example.org
The cytotoxic T lymphocyte (CTL) is a promising candidate for an effector cell in adoptive immunotherapy for renal cell carcinoma (RCC). Here we report the clinical course and in vivo immune responses of a RCC patient with bulky retroperitoneal lymph node (RPLN) metastases who received adoptive autologous CTL therapy. A 56-year-old woman diagnosed with RCC with multiple RPLN metastases underwent unilateral nephrectomy. Autologous RCC cells were primary-cultured from surgical specimens. Before addition of peripheral blood mononuclear cells (PBMC) for CTL induction, subconfluent RCC cells were irradiated with 50 Gy. The PBMCs were then cultured on RCC cells in the induction medium supplemented with four kinds of interleukins. The induced CTLs showed the potent killing activity against autologous RCC cells in a typical MHC-class I-restricted manner. The patient received three courses of CTL therapy with a total of 10·2 × 109 cells, and the RPLN mass decreased markedly in size after the second course. Eosinophilia and enhanced CTL inducibility from peripheral blood were observed after CTL administrations. The patient was progression free without further treatment; however, she developed rapidly progressive glomerulonephritis more than 1 year after the last treatment. The patient died of newly developed metastases 27 months after the start of CTL therapy. At autopsy, viable RCC cells were found in multiple metastatic sites. However, only diffuse fibrous tissue was observed in the responding RPLN mass. Apparent histological divergence was observed between primary and metastatic sites.
Since the establishment of methods to isolate genes encoding tumour antigens were recognized by the cytotoxic T lymphocyte (CTL), numerous tumour-associated antigens (TAAs) have been identified in melanoma and various other types of cancer [1,2]. CTLs are known to have high and specific cytotoxic activity against cancer cells, recognizing TAAs in a major histocompatibility complex (MHC)-restricted manner. CTLs are also considered to play a significant role in tumour regression in melanoma and RCC patients receiving high-dose interleukin (IL)-2 therapy or tumour infiltrating lymphocyte (TIL) therapy [3,4]. If autologous CTLs can be generated, they should be potent effector cells for adoptive immunotherapy. Like melanoma, the expression of TAAs such as RAGE-1, mutated HLA-A2 protein or other melanoma-associated antigens were observed in RCC specimens and RCC cell lines [5–7]. However, only limited data are available regarding clinical efficacy of CTLs induced by those known TAAs. Thus, until more effective TAAs have been identified for RCC, an individual's tumour is the only source of TAAs for generating a tumour-specific immune reaction against RCC. Recently, we reported that human CTLs can be efficiently generated in vitro by co-culturing peripheral blood mononuclear cells (PBMCs) with primary cultured tumour cells or with tumour sections using a medium containing IL-1, -2, -4 and -6 [8–10]. This technique allows highly reproducible induction of autologous CTLs against malignant glioma cells and RCC cells, and induced CTLs show potent killing activity against autologous cancer cells . Based on this background, pilot studies of adoptive CTL therapy for patients with metastatic RCC and malignant glioma  are now ongoing at Tsukuba University. In a RCC case presented here, autologous CTL therapy induced remarkable and durable regression in bulky lymph node metastases. Along with the clinical response, the CTL inducibility from peripheral blood were enhanced markedly after the CTL therapy.
SUBJECTS AND METHODS
A 56-year-old woman with microhaematuria was diagnosed with left renal tumour. Abdominal computerized tomography (CT) showed a 10-cm left renal tumour and multiple retroperitoneal lymphadenopathies up to 5 cm in diameter. There was no distant metastasis. She underwent left nephrectomy and retroperitoneal lymph node dissection in October 1997. The pathological diagnosis of the surgical specimens was renal cell carcinoma with lymph node metastases (granular cell type, pT2pN2M0, UICC classification 2002). Although the patient received adjuvant inteferon-α administrations, the follow-up CT at 4 months after nephrectomy revealed recurrent retroperitoneal lymph node (RPLN) metastases. The patient was treated with a combination of cimetidine and interferon-α, but the therapy was discontinued due to rapid growth of metastases. After obtaining written informed consent, the patient was enrolled in the present pilot study, which had been approved by the ethical committee of Tsukuba University.
Induction of autologous CTL
A RCC cell line was established from nephrectomy specimens, and maintained in RPMI-1640 medium with 10% fetal bovine serum (FBS). Induction of CTL was carried out as described previously . Briefly, PBMCs were separated by the conventional Ficoll-Paque method (Lymphosepal, Tokyo, Japan) from 30 ml of heparinized peripheral blood and suspended in CTL-induction medium, i.e. RHAM-α medium  supplemented with autologous plasma (5%) or plasma protein fraction (5%), IL-1 (Genzyme Co., Cambridge, MA, USA, 167 U/ml), IL-2 (Shionogi Co., Ltd, Osaka, Japan, 67 U/ml), IL-4 (Genzyme Co., 67 U/ml) and IL-6 (Genzyme Co., 134 U/ml). The PBMCs (effector) were seeded onto 24-well culture-plates that contained a confluent monolayer of the autologous RCC cells (target) irradiated previously with 50 Gy. The effector/target (E/T) ratio was adjusted at 10 : 1. Half the culture medium was changed every other day until the lymphocytes began to grow. After complete lysis of irradiated target cells (usually 2 weeks), the lymphocyte preparations were transferred to six-well culture plates and restimulated with irradiated target cells at an E/T ratio of 10 : 1. Approximately 5–7 days after restimulation, the lymphocyte preparation was harvested. Before administration to the patient, lymphocytes were pelleted and washed three times with saline containing 1% human serum albumin (HSA) by centrifugation at 1400 r.p.m. (250 g) for 5 min, then resuspended in 100 ml of saline containing 4% HSA and 67 U/ml of IL-2 (Shionogi & Co.). This preparation was administered as CTLs. The first and second treatment course consisted of three administrations with 7-day intervals. A CTL induction culture was performed for administration. The third treatment course consisted of only one CTL administration.
CTL induction assay
A CTL induction assay was carried out in attempt to monitor CTL inducibility from PBMCs of the patient during CTL therapy. The culture method was almost the same as the CTL induction described above. Briefly, PBMCs were collected from 30 ml of the patient's peripheral blood, divided among 60 wells of a 96-well microplate, and co-cultured with irradiated autologous RCC cells at an E/T ratio of 2 : 1. Restimulation of lymphocytes was performed at 2 weeks and 4 weeks after the initiation of mini-scale culture by the addition of irradiated autologous RCC cells at an E/T ratio of 10 : 1. The cytotoxic activity of the proliferated lymphocytes was determined at 5–7 days after the second restimulation.
The non-radioisotope crystal violet (CV) staining assay was adopted for the determination of cytotoxic activity of CTLs as described previously . This assay is consistent with the standard 51Cr-release assay at an effector/target (E/T) ratio of 10 : 1 or lower (see Fig. 1b in ). Briefly, the target cells, 1 × 104 cells/well in 200 µl culture medium, were seeded in each well of 96-well plates and precultured overnight. After washing the culture plate with PBS, the cultured lymphocytes, suspended in 200 µl of CTL culture medium, were added as effector cells to each well at the indicated E/T ratio. The cells were co-cultured for 24 h, and were then washed once gently with an appropriate amount of PBS to remove effector cells. Adherent target cells were fixed for 1 h with 10% (v/v) formalin (200 µl/well) and then stained with crystal violet solution (0·4% in water, 100 µl/well) for 30 min at room temperature. The plate was washed with tap water and dried at room temperature. To each well, 200 µl of 80% methanol was added, and the OD 570 of each well was determined. The percentage of surviving target cells was expressed as follows:
Surviving target cells (%) = (B-C)/A × 100
where A is the absorbance of control target cells precultured in a separate plate just before the addition of the effector cells, B is the absorbance of remaining target cells to which effector cells were added and C is the absorbance of effector cells only. Each value in the figures represents the mean of triplicates. Note that some results showed more than 100% of surviving due to growth of target cells during the 24 h incubation.
Inhibition of cytotoxicity by monoclonal antibodies
Target cells were precultured overnight in 96-well plates at 1 × 104 cells per well. Effector cells were pretreated with monoclonal antibodies against CD3 (Nichirei Co., Tokyo, Japan, OKT-3), CD8 (Nichirei Co., OKT-8) and CD4 (Nichirei Co., OKT-4) at 37°C for 4 h. Target cells were pretreated with antibodies against HLA-I (Dako Japan Inc., Kyoto, Japan, w6/32), HLA-II (Cosmo Bio Co., Tokyo, Japan, 88·12·2 clone), HLA-A24 and HLA-31 (Cosmo Bio Co., SPVL3 clone) at 37°C for 4 h and then incubated with effector cells at 37°C for 6 h. The killing activities at the E/T ratio of 8 : 1 were measured by 24 h CV-staining assay. The killing inhibition by antibodies was expressed as follows:
Killing inhibition (%) = (A-B)/(T-B) × d100
where at the end of the assay, A is the absorbance of target cells to which CTL and antibodies were added, B is the absorbance of target cells to which only CTLs were added and T is the absorbance of target cells to which neither CTL nor antibodies were added.
Flow cytometry and HLA-typing
Lymphocytes were characterized by staining with FITC-labelled monoclonal antibodies (Becton-Dickinson, Moutain View, CA, USA) against CD3, CD4, CD8 and CD56 surface antigens. Flow cytometry was performed with a FACScan (Becton-Dickinson). Serological HLA-typing was performed using Terasaki HLA-class I tissue typing trays (One Lambda, Canoga Park, CA, USA).
Cytotoxic activity of CTLs
The percentage of CD3+ CD8+ cells of the induced CTLs ranged from 60% to 90% with the remainder consisting mainly of CD3+ CD4+ in each induction. As shown in Fig. 1a, CTLs had high cytotoxic activity against autologous cancer cells. As the HLA-subtype of the patient was A24, A31, CTL activity was characterized using antibodies including anti-HLA-A24 and anti-HLA-A31. The cytotoxic activity of CTLs was inhibited by anti-CD3 and -CD8 antibodies, but not by anti-CD4 antibody. Both anti-HLA class I and HLA A24 antibody inhibited CTL activity. Neither anti-HLA class II and nor -HLA A31 antibody inhibited CTL activity (Fig. 1b). The CTLs did not kill an allogeneic RCC cell line and allogeneic malignant glioma cell lines sharing HLA-24 (data not shown).
These results are consistent with the typical characteristics of a CD8+ and MHC-class I-restricted CTL population.
Adoptive transfer of CTL and clinical course
In May 1998, the patient received the first course of CTL therapy. The CTLs were transferred adoptively to the patient by weekly intravenous infusions for 3 consecutive weeks. In this study, systemic IL-2 was not administered to the patient. The number of CTLs infused was 2·2 × 109 cells in this course. There were no episodes of toxicity associated with the treatment. Laboratory tests revealed no remarkable changes except eosinophilia, up to 500 cells/mm3, which returned to normal 3 months after the treatment. Figure 2 illustrates the size of the RPLN mass as evaluated by CT. After starting the therapy, the lymph node mass stopped growing, and thereafter decreased gradually in size. The patient maintained good performance status without development of new lesions. In January 1999, the second course of CTL therapy was performed. A total of 7·0 × 109 CTL cells was infused in divided doses over 3 weeks. At this time the patient experienced low-grade fever, but it resolved spontaneously within 72 h after CTL infusions ended. The eosinophilia developed again and continued for 2 months. One month after the CTL infusion, a CT scan revealed a marked decrease in the size of the RPLN mass, as shown in Fig. 3a and b. The patient stayed in partial remission for 8 months, and received additional CTL therapy with a single injection of 1 × 109 cells in October 1999. Although no further regression of the size of the RPLN mass was observed, it remained stable for 1 year without further treatment. The patient was progression-free and doing well until she noticed pretibial oedema in December 2000. Laboratory testing revealed acute renal dysfunction and proteinuria. The clinical course was compatible with rapidly progressive glomerulonephritis (RPGN). The histology of an open renal biopsy showed crescentic glomerulonephritis with granular deposition of immunoglobulins in the glomerular capillary walls. Plasmapheresis was initiated to remove immune complex. The renal function was improved slightly, and proteinuria was reduced after starting plasmapheresis. Despite this, the patient's general condition deteriorated progressively. Subsequently, the patient started to suffer from repeated episodes of gastrointestinal bleeding. An abdominal CT scan revealed a newly developed large tumour in the ascending colon. In contrast, there was no increase in the size of the RPLN mass. The patient died of panperitonitis 27 months after the start of the CTL therapy.
CTL induction assay
Before and 1 month after the CTL infusion of the second treatment course, CTL induction assays were performed as described in Subjects and methods. Before the CTL infusions, lymphocyte proliferation was observed in 12 of 60 wells (20%), as shown in Fig. 4. The proliferated lymphocytes in each well, which were enriched with CD3+ CD8+ cells, were harvested and their cytotoxic activity against autologous RCC cells determined. The lymphocytes from four wells killed over 50% of target cells at an E/T ratio of 8 : 1. In contrast, a marked lymphocyte proliferation was observed in 30 of 60 wells (50%) at 1 month after CTL infusions. Moreover, lymphocytes from 19 wells showed high killing activity, as shown in Fig. 3. The inducibility of the CTLs was somewhat decreased at 7 months after the second treatment, but a significant increase in the induction rate was observed again after the CTL administration at the third treatment course, as shown in Fig. 4.
Autopsy and histological findings
With permission from the family, an autopsy was performed 2 h after death. Grossly, massive exudative ascites and subphrenic abscess were found. Because the ascending colon tumour had apparently penetrated the colon wall, it was considered to be responsible for the peritonitis. Widespread lymph node metastases were also seen in parapancreatic, para-aortic and mesenteric regions. In contrast, diffuse fine fibrous tissues were found around the inferior vena cava and the abdominal aorta, which corresponded to the shrunken RPLN mass. Histological examination of the ascending colon tumour revealed that it consisted of round and pleomorphic RCC cells with severely atypical nuclei, which was in sharp contrast to the histology of the primary site, where granular cell carcinoma was dominant (Fig. 5a,b). In addition, sarcomatoid changes were observed in the colon and lymph node tumours. RCC cells were morphologically viable in most of the metastatic sites. In contrast, the histology of the RPLN mass, the lesion that responded to CTL therapy, showed only diffuse fibrous tissue without viable RCC cells (Fig. 5c). Immunohistochemical staining revealed the marked infiltration of both T and B lymphocytes in the fibrous tissue. Among T lymphocytes, proportionately more CD8+ cells than CD4+ cells were observed in the lesion. A similar CD8+ T lymphocyte accumulation was not seen in viable metastatic sites. In the right kidney, cellular and fibrocellular crescent formations were found in about 70% of the glomeruli. There was no apparent lymphocyte infiltration in glomeruli. In contrast, predominately CD8+ T cell infiltration was found in the interstinalis. Careful histological examination revealed no vasculitis or any other changes associated with active autoimmune disease.
Preliminary clinical studies have shown the therapeutic benefit of using in vitro-activated T cells against several cancers such as glioma, malignant melanomas and RCC [14–16]. However, these reports also suggest that present adoptive immunotherapy is not powerful enough to improve the survival rate of patients with metastases . It is possible that suboptimal activity and a heterogeneous population of infused effector cells is one factor that may have limited the efficacy of adoptive T cell therapy . In this report, we describe the case of a RCC patient with bulky RPLN metastases that responded to adoptive CTL therapy. The patient received 10·2 × 109 CTL cells, the high killing activity of which were confirmed by cytotoxic assay using autologous RCC cells. The patient continued in partial remission for almost 2 years without additional treatment. During the treatment, two prominent in vivo immune responses were observed. First, the patient's routine haematological tests showed eosinophilia that was closely related to the CTL infusions. A significant increase in the number of peripheral blood eosinophils (but not other leucocytes) with a peak eosinophil count of 500/mm3 developed after the first and second treatment courses. The eosinophilia lasted 3 and 2 months, respectively. Although eosinophilia is well known to be associated with systemic IL-2 therapy for RCC , the patient had not received concomitant IL-2 administration. Eosinophil activation is known to be T cell-dependent in many immunotherapy model  systems, thus the eosinophilia observed here may reflect eosinophil mobilization induced by CTLs. It is possible that eosinophils play a role in regression of a tumour as well in acute rejection of transplanted allografts . Soiffer et al. reported that eosinophils were often found in association with damaged endothelium in the vasculature of the resected tumour of melanoma patients treated with granulocyte-macrophage colony-stimulating factor transduced tumour vaccine . In the present case, however, eosinophil infiltration was not observed in the shrunken RPLN mass. Hence, the significance of the eosinophilia remains to be clarified in the present case.
Secondly, CTL induction assay indicated clearly that the CTL inducibility from peripheral blood was enhanced by CTL administration. The enhanced inducibility persisted for 4 months and decreased somewhat at 7 months after CTL administration. The enhancement of the CTL inducibility was, however, observed again after the third treatment course. It is not clear whether the enhanced CTL inducibility directly indicates an increase in the frequency of CTL precursors. It might indicate the activation of immune cells in a helper arm for CTL induction. Nevertheless, we consider that the assay is a useful method to monitor the in vivo antitumour CTL response by directly measuring the killing activivity against autologous cancer cells. To our knowledge, the enhancement of CTL inducibility secondary to adoptive T cell therapy, especially that accompanied with a marked clinical response, has not been reported previously.
The clinical course of the present case was characterized by long-term remission of the bulky lymph node metastases and subsequent development of new metastases to the colon. The mixed character of the clinical response, with stabilization of one lesion while metastasis progresses in other sites, is often reported in clinical trials of T cell-based immunotherapy [22,23]. One possibility is the selective escape of phenotypically altered cancer cells from immunity induced by CTL. The CTLs were generated using RCC cells that originated from the primary nephrectomy site as antigen stimulators. Although the induced CTLs revealed polyclonal pattern by T cell receptor Vβ gene-segment repertoire analysis, according to the method of Genevee et al. , there was a tendency to distribute to a few prominent bands (Vβ 5,13 and 15; data not shown) compared to bulk PBMC. The finding suggested that the CTLs recognized a few specific antigens on RCC cells derived from the primary site.
At autopsy, there was an apparent difference in RCC histology between the primary nephrectomy site and the metastatic sites. The metastatic sites consisted of RCC cells that were considerably less differentiated than the RCC cells at the primary site. This may result from natural selection of the metastatic phenotype, because it is well known that metastatic sites do not always comprise the same histology as the primary site in RCC. Although it is not clear whether selective escape might be involved in the present case, the heterogeneous features of RCC strongly suggest the requirement of strategies to overcome the escape from immune recognition .
Finally, the patient suffered from RPGN. Both the open biopsy and autopsy histology were compatible with immune complex-associated crescentic glomerulonephritis. There was moderately diffuse interstitial infiltration with predominate CD8+ T cells; however, no abnormal T cell accumulation was found in the glomeruli. The question arises of whether the occurrence of RPGN in the present case is treatment-related or is a paraneoplastic feature of RCC. The answer to this question cannot be derived from this case. In the present case, RPGN and progression of RCC developed together more than 1 year after the last CTL administration. There have been several reports of RCC complicated with crescentic glomerulonephritis [26–28]. Taken together, the observations suggest that the RPGN in the present case is paraneoplastic rather than treatment-related. At the time of this report the trial is in progress, and three additional patients have been treated with no significant side-effects. Further clinical and immunological evaluation, including these patients, is required to ascertain the clinical benefits of CTL therapy in metastatic RCC.