Human Dendritic Cells Transfected with Renal Tumor RNA Stimulate Polyclonal T-Cell Responses against Antigens Expressed by Primaryand Metastatic Tumors¹

The incidence of RCC³ is rising, with approximately 30,000 new cases detected annually in the United States accounting for 11,600 deaths/year (1). About one-third of RCC patients will present with metastatic disease, and one-third of the remainder will eventually develop distant metastases for which no effective standard treatment exists. Like melanoma, RCC has been shown to respond to immunotherapeutic intervention, suggesting that these tumors express specific rejection antigens, which are recognized by the immune system. However, widely expressed and clinically effective renal tumor antigens shared among many RCC patients have not yet been identified. Realizing the limitations of immunization with defined antigens, several studies have provided preliminary evidence that DCs loaded with tumor-derived antigenic mixtures in the form of tumor extracts (2) or cell lysates (3) may represent an efficient strategy to induce antitumor immunity in RCC patients. Although this is promising, it can be predicted that a generalized and widespread application of these approaches will be hampered by the requirement of large amounts of tumor tissue needed for antigen preparation and DC pulsing. This limitation becomes even more relevant because it has been argued that continuous boosting will be necessary to maintain effective antitumor immune responses (4).

Alternatively, it was recently shown that DCs transfected with antigens encoded in tumor mRNA are capable of inducing potent T-cell responses against multiple, tumor-specific epitopes while obviating the need to identify the antigens involved (5). Because functionally intact RNA can be amplified using PCR technology, nonlimiting amounts of antigen could be generated, even from small amounts of tumor tissue, for DC pulsing (6).

The primary objective of this study was to determine whether DCs transfected with autologous RCC RNA are able to stimulate CTL responses in vitro against a broad repertoire of as yet unidentified RCC-specific antigens. We further sought (a) to investigate whether the CTLs generated by stimulation with tumor cell-derived antigenic mixture in the form of RNA would also induce autoimmune reactivities against nontumor cellular antigens and (b) to define whether primary renal tumor-specific CTLs also recognize antigens expressed by metastatic lesions.

Here we show that DCs transfected with RNA isolated from renal cancer tissue are remarkably effective in stimulating tumor-specific T-cell response in vitro but do not cross-react with antigens from normal tissues including antigens expressed by renal parenchyma. Conversely, the tumor-specific CTLs lysed allogeneic tumor but not allogeneic nonmalignant targets, suggesting the presence of shared antigens among renal carcinomas. T-cell responses against TERT accounted in part for the responses against the allogeneic tumor targets because tumor RNA-transfected DCs reproducibly stimulated the formation of TERT-specific CTLs as a component of the polyclonal antitumor response. Although our studies suggest that distinct differences exist in the antigenic repertoire between the primary renal tumor and metastases, total tumor RNA-transfected DCs were capable of stimulating CTLs specific for the primary tumor as well as CTLs against metastatic targets, providing a scientific foundation for further clinical investigation of this approach.

PBMCs, renal tumor, and nonmalignant control tissues were collected after obtaining informed consent from human subjects treated on protocols approved by the institutional review board. All primary tumor material was histologically classified as clear cell carcinoma by an experienced pathologist. Tissue samples were transported to the laboratory at room temperature in RNA preservation solution (RNAlater; Ambion, Austin, TX) and then processed for RNA generation. Control RNA was extracted from three histologically distinct, nonmalignant control tissues, which included benign RE, ureteral SM, and PBMCs. Autologous renal tumor cells were generated from short-term primary cultures. Tissues were minced with opposing scalpels, and the resulting tissue fragments were cultured in RPMI 1640 supplemented with 10% FCS using a humidified incubator. The epithelial nature of these cells was confirmed by flow cytometry. Early passage cells were used as targets in CTL assays. Total cellular RNA was generated using the RNeasy kit (Qiagen, Valencia, CA) following the manufacturer’s protocol. To attest to their intactness, all RNA preparations were subjected to electrophoresis in 1.2% agarose gel under denaturing conditions with clear visualization of intact 18S and 28S ribosomal bands after ethidium bromide staining. Furthermore, reverse transcription-PCR for β2-microglobulin was performed with successful amplification of the gene product visualized as a 240-bp fragment on agarose gel (data not shown).

To generate in vitro transcribed TERT RNA, the vector pGEM4Z/TERT/A64 was constructed. In brief, the EcoRI fragment of pGRN145 (Geron Corp., Menlo Park, CA) was cloned into the EcoRI site of pGEM4Z/A64. To remove a SpeI site upstream of the poly(A) stretch in the vector, the 3400-bp SaII-SpeI fragment was cloned into the SaII and XbaI sites of pGEM4Z/A64. The in vitro transcribed RNA contains 61 nt from the polylinker of pGEM4Z, followed by 10 nt of the hTERT 5′-untranslated region, 3393 nt of the hTERT open reading frame, 24 nt of hTERT untranslated region, 27 nt of polylinker, 64 A residues, and a SpeI half site. GFP RNA was generated by in vitro transcription of the plasmid pGEM4Z/GFP/A64 as described previously (7).

For DC culture, we used techniques previously described by us (8) to allow processing of clinical-grade cellular material using defined serum free-media and supplements. Briefly, a concentrated leukocyte fraction was generated through a 2-h restricted peripheral blood leukapheresis processing 6–8 liters of blood with each collection. The leukapheresis product was further separated by density gradient centrifugation over polysucrose/sodium diatrizoate (HISTOPAQUE; Sigma Diagnostics, St. Louis, MO), and cells were resuspended in serum-free AIM-V medium (Life Technologies, Inc., Grand Island, NY). PBMCs were incubated in a humidified incubator for 2 h at 37°C to allow plastic adherence. The adherent cell fraction was used for DC culture by incubation in serum-free AIM-V medium supplemented with recombinant human IL-4 (500 units/ml) and recombinant human granulocyte macrophage colony-stimulating factor (800 units/ml; R&D Systems, Minneapolis, MN). After 7 days of culture, cells were harvested and phenotypically characterized to assure that they met the typical phenotype of immature DCs: CD3^neg; CD14^neg; CD16^neg/CD56^neg; CD19^neg; MHC I^pos; MHC II^pos; CD40^low; CD86^low; CD80^low; and CD83^low (9). DC preparations fulfilling these phenotypic criteria were used for subsequent RNA transfection.

RNA loading of autologous DCs was performed by simple coincubation of DCs in RNA solution without any transfection reagent. In brief, DCs were washed twice in PBS, counted, and spun at 300 × g for 10 min. They were resuspended in AIM-V medium and incubated in RNA-containing solution (3.0 μg RNA/2 × 10⁶ DCs) for 45 min in a humidified incubator at 37°C/5% CO₂. RNA-transfected DCs were used as stimulators or targets in cytotoxicity assays.

The antigen-presenting function of the RNA-pulsed DCs was assessed by measuring the induction of primary CTL responses in a standard ⁵¹Cr cytotoxicity assay. The T-cell-enriched nonadherent fraction of PBMCs obtained after the DC plastic adherence step was used for CTL generation. Nonadherent PBMCs were cultured in RPMI 1640 supplemented with 20 units/ml human IL-2 and 10 ng/ml human IL-7 (R&D Systems, Minneapolis, MN). Cells were stimulated twice (8 days apart) with autologous DCs transfected with RCC RNA at a stimulator:effector ratio of 1:10. After 16 days of culture, effector cells were harvested without further separation for cytotoxicity assays. The purity of CD8+ T cells was routinely at least 45% by flow cytometric analysis. Target cells included DCs pulsed with renal tumor or benign tissue-derived RNA or DCs pulsed with an irrelevant control RNA encoding the jelly fish GFP. Target cells were labeled with ⁵¹sodium chromate in saline solution (Na[⁵¹Cr]O₄; NEN Life Science Products, Boston, MA) by incubation of 2 × 10⁶ transfected DCs in 1 ml of RPMI 1640 with 100 μCi of ⁵¹Cr for 1 h at 37°C/5% CO₂. After three washes, 5 × 10^{3 51}Cr-labeled targets and serial dilutions of effector cells at various E:T ratios were incubated in 200 μl of RPMI 1640 in 96-well U-bottomed plates. These plates were incubated for 6 h at 37°C/5% CO₂. Fifty μl of the supernatant were harvested, and released ⁵¹Cr was measured with a scintillation counter. Spontaneous release was less than 15% of the total release by detergent in all assays. SD of the means of triplicate wells was less than 5%. Monoclonal antibody blocking studies were performed by the addition of CD8, HLA class II, or isotype control antibodies (Research Diagnostics Inc., Flanders, NJ) during coincubation of tumor-specific effector cells with RCC RNA-transfected DC targets plated at E:T ratios of 40:1.

We first determined whether DCs generated from renal cancer patients and transfected with autologous renal tumor RNA are capable of stimulating CTL responses against an uncharacterized repertoire of renal tumor antigens in vitro. DCs were cultured from leukapheresis-derived monocytic precursors in the presence of IL-4 and granulocyte macrophage colony-stimulating factor as described previously (8). DCs of the immature phenotype (CD3^neg, CD14^neg, CD16^neg/CD56^neg, CD19^neg_, MHC I^pos, MHC II^pos, CD40^low, CD86^low, CD80^low, and CD83^low) were used for subsequent RNA transfection. The rationale for using immature rather than mature DCs for CTL priming is based on the demonstration that immature DCs are a prerequisite for successful RNA uptake and antigen processing (10). Autologous PBMCs (patient A) were stimulated twice, 8 days apart, with the renal tumor RNA-transfected DCs to generate CTLs. As shown previously, we assessed the induction of tumor-specific CTLs in cytotoxicity assays by measuring lytic activities against tumor RNA-transfected DC targets (11). We also tested whether these RNA-transfected DC targets faithfully represent the antigenic spectrum expressed by renal carcinoma cells by simultaneously measuring lytic activities against autologous renal tumor cells, which were obtained from early passage primary cultures. DCs transfected with RNA extracted from autologous PBMCs or DCs transfected in vitro transcribed GFP RNA (data not shown) served as control targets. As demonstrated in Fig. 1, upper panel, renal tumor RNA-transfected DCs were capable of stimulating CTLs that recognized and lysed both renal tumor RNA-transfected DCs and primary renal tumor cells with similar efficacy, whereas PBMCs and GFP RNA-transfected DCs were not lysed. To our knowledge, these data represent the first demonstration that CTLs directed against renal tumor antigens can be induced by stimulating PBMCs from RCC patients with renal tumor RNA-transfected DCs. They further suggest that RNA-transfected DCs can, in fact, substitute for tumor cells as targets in cytotoxicity assays, thus providing a general platform for monitoring patients for tumor-specific CTLs and obviating the need to culture and label autologous tumor cells from every individual patient.

As shown in Fig. 1, bottom left panel, the addition of anti-CD8 but not MHC class II or isotype control antibodies during coincubation of effector and target cells resulted in significant inhibition of lytic activities, suggesting that the observed responses are predominately CD8⁺ T-cell mediated. Furthermore, lytic activities against K562 targets were weak and significantly less than those against RCC RNA-transfected DC targets (Fig. 1, bottom right panel). Cumulatively, these findings suggest that immunization with RNA-encoded antigens is specifically geared for the induction of class I restricted CTL responses because the transfected RNA is translated into protein-antigen in the cytoplasm. It is well known that antigens that gain access to the cytoplasm are channeled preferentially into the endogenous class I presentation pathway (12).

We next determined whether the tumor-specific CTLs stimulated by tumor RNA-transfected DCs can recognize antigens expressed by normal control tissues as well as tumor antigens. DCs from a second cancer patient (patient B) were transfected with autologous RNA from the following sources: (a) RCC; (b) RE; (c) PBMCs; (d) SM; and (e) in vitro transcribed RNA encoding for GFP. These RNA-transfected DCs were used for CTL priming and as target cells in cytotoxicity assays. As shown in Fig. 2,A, renal tumor RNA-transfected DCs stimulated a robust CTL response against their cognate DC targets (RCC RNA-transfected DCs), whereas no cross-reactivity was apparent against DC targets loaded with RNA isolated from normal tissues (RE, SM, and PBMCs) or GFP. Conversely, stimulation of PBMCs with GFP RNA-transfected DCs reproducibly induced CTLs that lysed GFP RNA-pulsed DCs, but not tumor RNA-pulsed DC targets (data not shown). As shown in Fig. 2, B–D, we attempted to induce CTLs against normal tissue antigen(s) including RE, SM, or PBMCs. CTLs were cultured using DCs transfected with the corresponding normal tissue RNA and analyzed for their ability to recognize and lyse their cognate normal or tumor RNA-transfected DC targets. Although weak but consistent cross-reactivities between RE RNA-derived CTLs and RCC RNA- or RE RNA-expressing DC targets were observed (Fig. 2,B), none of the normal tissue RNA-transfected DC preparations were capable of inducing measurable CTL activity against either normal or RCC antigens. Assuming that antigens that are shared between normal tissues and RCC exist (known as differentiation antigens in melanoma studies), our data indicate that those antigens are either too weak to be recognized by CTLs stimulated by RCC RNA-transfected DCs (Fig. 2,A) considerably less effective in stimulating CTL responses against the cognate target or RCC targets (Fig. 2, B–D).

We next investigated whether autologous renal tumor RNA-transfected DCs are capable of stimulating CTL responses against antigens shared among other renal tumors as well as against autologous tumor antigens. In the context of vaccination protocols, antigens in the form of allogeneic renal tumor RNA may represent an attractive alternative because well-characterized and generic vaccines could be generated. As seen in Fig. 3, left panel, DCs transfected with autologous RCC RNA (patient B) were capable of eliciting strong CTL responses against their cognate, but not against PBMC RNA-transfected DC targets. Interestingly, these CTLs (patient B) were also effective in recognizing and lysing autologous DC targets transfected with RCC RNA isolated from another RCC patient (patient A), whereas DC targets transfected with PBMC RNA from patient A were not lysed. Similar observations were made in another set of experiments (Fig. 3, right panel) in which renal tumor-specific CTLs were stimulated using DCs transfected with RCC RNA isolated from patient C. These CTLs lysed cognate DC targets and autologous DCs transfected with allogeneic RCC RNA (patients A and B) but did not lyse DCs transfected with (allogeneic) PBMC RNA extracted from these patients.

These experiments show that under the experimental conditions used, tumor RNA-encoded allogeneic MHC molecules did not stimulate an allogeneic CTL response. This observation can be explained as follows: first, it has been shown that expression levels of histocompatibility antigens by primary RCC are low (13). Thus, the copy number of mRNA encoding for MHC antigens within the allogeneic tumor RNA pool is expected to be minimal, and only insufficient quantities of allogeneic MHC antigens may be expressed as proteins and transported to the DC cell surface, precluding effective presentation. Second, allogeneic MHC-derived peptides are competing for presentation with those from degraded self-MHC, which are present in high numbers in the DC cytoplasm, making effective presentation unlikely. Therefore, the lysis seen with the allogeneic RCC can be solely attributed to renal tumor antigens common between the patients analyzed and not to an allogeneic CTL response.

These experiments further suggest that these patients express, in addition to shared tumor antigens, patient-specific antigens unique to the parental tumor because the responses against the allogeneic targets (shared antigens only) were consistently less efficient than the responses against the autologous targets (patient specific and shared). At this point, it is unclear whether these antigens are shared among many or all RCC patients. The complexities inherent to human studies limited our observations to a small number of patients in whom these cross-reactivities between autologous RCC RNA vaccines and allogeneic RCC RNA-transfected DC targets were identified. On the other hand, similar observations were made by us in patients with prostate cancer in whom prostate tumor RNA-transfected DCs stimulated CTLs that were capable of recognizing tumor targets as well as (allogeneic) LNCaP prostate cancer cells.⁴

Unfortunately, unlike those seen in melanoma, attempts to identify and isolate widely expressed RCC-specific tumor antigens have not been very successful. However, it has recently been shown that the polypeptide component of telomerase (TERT) can be an attractive candidate for a broadly expressed rejection antigen for many cancer patients (14). Therefore, we sought to determine whether TERT-specific CTLs could account, at least in part, for the cross-reactivities observed among renal tumors. DCs were transfected with either renal tumor RNA derived from patient D or with TERT mRNA generated by in vitro transcription of the plasmid pGEM4Z/TERT/A64. Both DC preparations were used for CTL priming as well as targets in cytotoxicity assays. Fig. 4, left panel, shows that total renal tumor RNA-transfected DCs stimulated T-cell responses that recognized and lysed tumor targets, whereas DC targets transfected with control RNA species (PBMC RNA or GFP RNA) were not lysed. Renal tumor-specific CTLs were comparable to TERT-specific CTLs stimulated with TERT RNA-transfected DCs in recognizing and lysing TERT-expressing targets, indicating that the levels of TERT RNA in the total tumor RNA pool were sufficient to stimulate TERT-specific CTLs. These data also show that in addition to the characterized and single tumor antigen TERT, other as yet unidentified antigens are involved in this response. Most importantly, the tumor-specific CTLs were consistently superior to CTLs stimulated with TERT RNA-transfected DCs in recognizing and lysing tumor targets, suggesting that tumor-specific CTLs represent a polyclonal response providing more effective antitumor activity than T-cell responses directed against a single antigen in the form of TERT.

Novel vaccine protocols should be judged not only by their activity against the primary tumor but also, and more importantly, by their activity against distant metastases, which frequently escape immunological recognition (15). However, the preclinical assessment of lytic CTL activity against primary or metastatic tumor targets is hampered by the requirement to simultaneously culture both primary or metastatic tumor cells from the same patient. As a potential solution to this obstacle, we used tumor RNA-transfected DCs to assess the efficacy of vaccination with (RNA-encoded) primary tumor antigens against metastatic targets. Autologous DCs were transfected with cellular RNA extracted from the following sources and used to stimulate CTLs from PBMCs: (a) primary renal tumor; (b) regional LN metastasis; and (c) BM. These RNA-transfected DCs were used for CTL priming and as cellular targets. Additional targets consisted of DCs transfected with GFP RNA, TERT RNA, or autologous PBMC RNA. As shown in Fig. 5,A, RCC-transfected DCs stimulated CTLs that were not only effective in recognizing and lysing primary tumor targets (RCC-transfected DCs) but were also (albeit less effectively) effective in recognizing and lysing LN RNA-, BM RNA-, and TERT RNA-transfected DC targets, whereas control targets were not lysed. DCs transfected with LN RNA (Fig. 5,B) stimulated CTLs that were equally effective in recognizing and lysing primary renal tumor or both metastatic tumor targets, whereas BM RNA-transfected DCs (Fig. 5 C) lysed preferentially their cognate targets (BM RNA-transfected DCs) and were somewhat less effectively in lysing primary tumor, LN, and TERT targets. The likely interpretation of these observations is that, in fact, small, but distinct differences in the antigenic repertoire of primary renal tumors and distant metastases exist. Despite these differences, renal tumor RNA-transfected DCs are capable of stimulating CTL activities against antigens expressed by metastatic tumors as well as antigens present on primary tumors. Our findings also highlight the potential of this approach to generate potent anticancer vaccines from metastatic tumor tissue, if the primary tumor has been surgically removed, and metastatic tumor tissue represents the only source for vaccination, as is often the case in RCC.

In this study, we have shown that human renal tumor RNA-transfected DCs are capable of inducing potent CTL responses directed against a broad spectrum of unidentified primary and metastatic renal tumor antigens. The tumor-specific CTLs were capable of recognizing and lysing RNA-transfected DCs and primary cultured renal tumor cells with similar efficacy. Thus, tumor RNA-transfected DCs may not only represent a potent strategy for CTL priming but may also represent a general method for monitoring the presence of tumor-specific CTLs in cancer patients undergoing this form of therapy. Furthermore, as shown in this study, RNA-transfected DCs may serve as a common cellular platform allowing us to measure and compare CTL activities against a broad range of tissue antigens expressed by normal or malignant tissues.

One important question potentially limiting the application of tumor RNA-based DC vaccines is the induction of autoimmunity by introducing (RNA-encoded) normal tissue antigens present on tumor cells into the same antigen processing and presentation pathway as the tumor antigens (16). Here we show that tumor RNA-transfected DCs stimulated CTLs that are capable of lysing tumor cells but not normal tissue antigens and that DCs transfected with normal tissue RNA failed to stimulate CTLs that lysed neither tumor nor normal tissue targets, indicating that harmful autoimmunity with pathological consequences may not be a serious issue with this approach. A potential explanation for this phenomenon may be that the antigens shared between normal and tissue antigens (known as differentiation antigens from melanoma studies) are either too weak or are not present in sufficient quantities to elicit CTL responses against any cellular target tested. Although our studies failed to demonstrate any measurable CTL induction against normal tissue antigens, the question of whether or not this approach will induce harmful autoimmunity will need to be carefully addressed in a clinical setting.

Although we have not observed cross-reactivities between renal tumor and nonmalignant tissue antigens, we show that the CTLs generated with tumor RNA-transfected DCs cross-react with renal tumor antigens expressed by nonrelated renal cancers, suggesting the presence of shared antigens among RCC. CTLs stimulated against the allogeneic tumor RNA-encoded MHC molecules apparently did not contribute to the recognition of these targets because autologous DCs transfected with allogeneic RE or PBMC RNA were not lysed under these experimental conditions. An important novel aspect of this study is the demonstration that the antitumor response stimulated by tumor RNA-transfected DCs encompasses as a subcomponent a response against TERT, suggesting that telomerase, which is overexpressed in almost every renal cancer and particularly in renal metastases,⁵ may represent a potential candidate for antigen-specific immunotherapy for RCC patients. On the other hand, we show that the polyclonal responses stimulated by tumor RNA-transfected DCs are superior to the responses against the characterized and single antigen TERT in recognizing and lysing tumor targets. These polyclonal CTLs may not only be more potent but may also optimize protection against various subclones of the primary tumor that may appear as metastases during the course of disease.

Consistent with this expectation, we show that DCs transfected with RNA isolated from the primary renal tumor provide CTL reactivities not only against their cognate targets but also, and more importantly, against metastatic targets. The recognition of the metastatic tumor targets by the primary tumor-specific CTLs was somewhat less than that against primary RCC targets, suggesting that, in fact, minor but distinct differences in the antigenic repertoire between both tumor types exist. Consistent with these findings, CTLs stimulated by BM RNA-transfected DCs were more effective in recognizing and lysing their cognate metastatic targets than primary tumor targets, suggesting that an antigenic shift may occur during differentiation of primary tumors into the metastatic phenotype.

Our demonstrated ability to reproducibly stimulate tumor-specific CTLs in many RCC patients that cross-react against the antigens expressed by metastatic tumors may point to the conclusion that tumor RNA-transfected DCs precisely mirror the antigenic distribution expressed by individual tumors as well as account for their intrinsic heterogeneity with potential antigen shift or loss in some tumor cells. This feature may optimize vaccination against the many and unidentified antigens on renal tumor cells and thus may provide protection against primary and metastatic tumors in the tumor-bearing host.

We thank Dr. Eli Gilboa for helpful discussions and Kay Walker for secretarial assistance.