Intratumoral peptide injection enhances tumor cell antigenicity recognized by cytotoxic T lymphocytes: a potential option for improvement in antigen-specific cancer immunotherapy

Antigen-specific cancer immunotherapy is a promising strategy for improving cancer treatment. Recently, many tumor-associated antigens and their epitopes recognized by cytotoxic T lymphocytes (CTLs) have been identified. However, the density of endogenously presented antigen-derived peptides on tumor cells is generally sparse, resulting in the inability of antigen-specific CTLs to work effectively. We hypothesize that increasing the density of an antigen-derived peptide would enhance antigen-specific cancer immunotherapy. Here, we demonstrated that intratumoral peptide injection leads to additional peptide loading onto major histocompatibility complex class I molecules of tumor cells, enhancing tumor cell recognition by antigen-specific CTLs. In in vitro studies, human leukocyte antigen (HLA)-A*02:01-restricted glypican-3144–152 (FVGEFFTDV) and cytomegalovirus495–503 (NLVPMVATV) peptide-specific CTLs showed strong activity against all peptide-pulsed cell lines, regardless of whether the tumor cells expressed the antigen. In in vivo studies using immunodeficient mice, glypican-3144–152 and cytomegalovirus495–503 peptides injected into a solid mass were loaded onto HLA class I molecules of tumor cells. In a peptide vaccine model and an adoptive cell transfer model using C57BL/6 mice, intratumoral injection of ovalbumin257–264 peptide (SIINFEKL) was effective for tumor growth inhibition and survival against ovalbumin-negative tumors without adverse reactions. Moreover, we demonstrated an antigen-spreading effect that occurred after intratumoral peptide injection. Intratumoral peptide injection enhances tumor cell antigenicity and may be a useful option for improvement in antigen-specific cancer immunotherapy against solid tumors.

patients has gradually improved; however, these therapies remain far from being satisfactory in most cancers [1,2]. Therefore, the development of novel treatment modalities, including antigen-specific cancer immunotherapies with peptide vaccines, dendritic cell vaccines, and adoptive cell transfer therapies, is critical for advancing effective cancer treatments [3][4][5]. While many tumor-associated antigens and epitopes recognized by cytotoxic T lymphocytes (CTLs) have been explored as possible antigen-specific cancer immunotherapies, the results of several anticancer immunotherapy clinical trials have been disappointing [6,7]. We conducted a clinical trial using the glypican-3 (GPC3) peptide vaccine in advanced hepatocellular carcinoma (HCC) patients. While this carcinoembryonic antigen overexpressed in HCC seemed to be an ideal target for anticancer immunotherapy [8][9][10][11][12][13][14][15], only immunological efficacy was apparent [16], whereas the clinical benefit was limited in patients [17]. Therefore, the establishment of an innovative strategy to link the antitumor immune response with the clinical response and to enhance the power of antigen-specific cancer immunotherapy is urgently required.
In the antigen-specific cancer immunotherapy concept, antigen-specific CTLs recognize and destroy tumor cells that present antigen-derived peptides using cell surface major histocompatibility complex (MHC) class I molecules. However, the density of the antigen-derived peptide endogenously presented on tumor cells is generally low, resulting in the ineffectiveness of antigen-specific CTLs [18]. This low density of presented antigen is one reason why antigen-specific cancer immunotherapy has been ineffective in clinical settings. One solution for overcoming this critical problem is to induce high-avidity CTLs. Such CTLs can recognize a smaller number of peptide-MHC class I complexes and would contribute to a better outcome [19]. Another solution is to enhance tumor cell antigenicity by means of additional peptide loading onto MHC class I molecules. Increasing the density of antigen-derived peptide would facilitate CTL recognition and destruction of the tumor cells.
In this study, we investigated whether intratumoral peptide injection would induce additional peptide loading onto tumor cells, and, if so, whether increased presentation would enhance antigen-specific CTL tumor cell recognition. Moreover, we evaluated whether intratumoral peptide injection could be a useful option for improvement in antigen-specific cancer immunotherapy against solid tumors.

Mice
Female BALB/c nude, NOD/SCID, and C57BL/6 mice (6-8 weeks old) were purchased from Japan Charles River Laboratories (Yokohama, Japan). OT-I mice [20], which are CD8 ? T-cell TCR transgenic mice expressing the TCR a-chain recognizing OVA 257-264 peptide in H-2 K b , were kindly provided by Dr. Takashi Nishimura (Hokkaido University, Sapporo, Japan). All animal procedures were performed according to the guidelines for the Animal Research Committee of the National Cancer Center, Japan.

IFN-c ELISPOT assay
The BD TM ELISPOT set (BD Biosciences, San Jose, CA, USA) was used for an interferon (IFN)-c enzyme-linked immunospot (ELISPOT) assay. CTLs were used as effector cells, and tumor cell lines with or without a peptide pulse (10 lg/ml for 1 h) were used as target cells. Effector cells (1 9 10 3 /well) were incubated with target cells (1 9 10 4 /well) in 200 ll of RPMI 1640 medium supplemented with 10 % FBS, penicillin, and streptomycin for 20 h at 37°C in 5 % CO 2 . The number of spots, indicating an antigen-specific CTL response, was automatically counted using the Eliphoto system (Minerva Tech, Tokyo, Japan).

Cytotoxicity assay
The Terascan VPC system (Minerva Tech) was used for cytotoxicity assays. Target cells were labeled with Calcein-AM (Dojindo Laboratories, Kumamoto, Japan) solution for 30 min at 37°C, washed three times, distributed to 96-well culture plates in duplicate, and incubated with effector cells for 4 h. Fluorescence intensity was measured before and after the 4-h culture, and antigen-specific cytotoxic activity was calculated as described previously [16].

Intratumoral peptide injection
In in vivo studies, tumors implanted on the backs of mice were injected with 50 lg peptide mixed with an equal volume of incomplete Freund's adjuvant (IFA, Montanide ISA-51VG; SEPPIC, Paris, France). The total volume of solution injected was 100 ll in all experiments.

Tumor excision and isolation of tumor cells
To investigate whether the injected peptide was loaded onto HLA class I molecules of tumor cells in a solid mass, an IFN-c ELISPOT assay was performed using these isolated tumor cells as target cells. Mice were killed and their dorsal tumors were dissected, cut into small pieces, and digested with collagenase (1.5 mg/ml) for 20 min at 37°C.

In vivo tumor growth inhibition assay
In a peptide vaccine model, H-2 K b -restricted OVA 257-264 peptide emulsified with IFA (50 lg/100 ll) was intradermally injected at the base of the tail of C57BL/6 mice, five times at 7-day intervals as described previously [13]. After vaccination, the induction of H-2 K b -restricted OVA 257-264 peptide-specific CTLs was detected by IFN-c ELISPOT assay (data not shown). In an adoptive transfer model, activated OT-I CTL (1 9 10 7 cells/500 ll) was intravenously injected. SW620 cells (5 9 10 6 cells/100 ll) were subcutaneously implanted into the backs of BALB/c nude mice; SK-Hep-1/vec, SK-Hep-1/GPC3, or HepG2 cells (5 9 10 6 cells/100 ll) were implanted into NOD/SCID mice, and RMA cells (5 9 10 4 or 5 9 10 5 cells/100 ll) were implanted into C57BL/6 mice. Tumor volume was monitored twice a week and calculated using the following formula: tumor volume (mm 3 ) = a 9 b 2 9 0.5, where a is the longest diameter, b is the shortest diameter, and 0.5 is a constant to calculate the volume of an ellipsoid. Mortality and morbidity were checked daily, and the mice were maintained until each mouse showed signs of morbidity or the length or width of the tumors exceeded 30 mm, at which point they were killed for reasons of animal welfare.

Tetramer staining and flow cytometry analysis
For the analysis of local accumulation of antigen-specific CTLs, isolated tumor cells, including tumor-infiltrating lymphocytes, were stained with H-2 K b OVA Tetramer-PE (OVA 257-264 [SIINFEKL]; MBL, Nagoya, Japan) for 20 min at room temperature and anti-mouse CD8-FITC (rat monoclonal, clone KT15; MBL) for 20 min at 4°C. Flow cytometry analysis was carried out using a FACSCanto II flow cytometer (BD Biosciences).

Immunohistochemistry
To investigate whether CD8 ? T-cells infiltrated normal tissues due to intratumoral peptide injection in a murine adoptive cell transfer model, we performed immunohistochemical staining of CD8 in tissue specimens from C57BL/ 6 mice using monoclonal anti-CD8 antibody (dilution 1:20, BioLegend, San Diego, CA, USA).

Statistical analysis
Comparisons of spot numbers and tumor volume at the last time point were performed using the Mann-Whitney U test. Survival was analyzed according to the Kaplan-Meier estimate, and differences between groups were compared using the log-rank test. Differences were considered significant at P \ 0.05. Data were analyzed with the statistical package, Dr. SPSS II (SPSS Japan, Tokyo, Japan).

In vitro CTL activity against peptide-pulsed targets
To evaluate the antigen-specific CTL response in vitro, IFN-c ELISPOT and cytotoxicity assays were performed. In both assays, the two types of effector cells were the HLA-A*02:01-restricted GPC3 144-152 peptide-specific CTL clone, which was established from peripheral blood mononuclear cells (PBMCs) of an HCC patient who had received the GPC3 144-152 peptide vaccine [16], and the HLA-A*02:01-restricted CMV 495-503 peptide-specific CTL clone, which was established from PBMCs of a healthy volunteer. The target cells were tumor cell lines with or without antigenic peptide pulses.
The peptide-specific CTLs showed strong activity against all peptide-pulsed cell lines, regardless of whether the tumor cells expressed the antigen. The density of the HLA-A*02:01-restricted GPC3 144-152 peptide endogenously presented on tumor cells was not enough to induce strong CTL activity.
Loading of injected peptide onto HLA class I molecules of tumor cells in vivo As shown in Fig. 2a, BALB/c nude mice were inoculated subcutaneously on their backs with SW620 (GPC3 -) tumor cells. When tumor diameters reached 5-7 mm, 50 lg HLA-A*02:01-restricted GPC3 144-152 peptide was injected into the tumor. After 2-96 h, the tumors were dissected, cut into small pieces, and digested with collagenase (1.5 mg/ ml) for 20 min at 37°C. To investigate whether the injected HLA-A*02:01-restricted GPC3 144-152 peptide was loaded onto HLA class I molecules of tumor cells in a solid mass, an IFN-c ELISPOT assay was performed in duplicate using these isolated tumor cells as target cells and HLA-A*02:01-restricted GPC3 144-152 peptide-specific CTLs as effector cells.
Loading of HLA-A*02:01-restricted GPC3 144-152 peptide onto HLA class I of tumor cells was detected (Fig. 2b). Without IFA, the density of loaded peptide gradually decreased after intratumoral peptide injection, whereas the loaded peptide density remained for 96 h after injection with IFA, suggesting that IFA is required for long-term stability of the injected peptide (Fig. 2c). Similar data were obtained with a combination of the HLA-A*02:01-restricted CMV 495-503 peptide and its specific CTLs (data not shown).

Antitumor effect of intratumoral peptide injection in an immunodeficient mouse model
We planned and executed the experimental schedule shown in Fig. 3a. Four tumors were implanted per mouse, and each tumor received a different combination of injections, as shown in Fig. 3b. From 5-7 days after tumor inoculation, mice were treated two or three times in 5-day intervals. The treatment regime was as follows: HLA-A*02:01-restricted GPC3 144-152 or CMV 495-503 peptide emulsified with IFA (50 lg/100 ll) was injected into a tumor, and, 2 h later, HLA-A*02:01-restricted GPC3 144-152 or CMV 495-503 peptide-specific human CTLs (1 9 10 7 cells/100 ll) were injected into the tumor.
Therapeutic advantage of intratumoral peptide injection as an option for antigen-specific cancer immunotherapy After the induction of OVA 257-264 peptide-specific CTLs by peptide vaccination (Fig. 4a) or after the adoptive transfer of OVA 257-264 peptide-specific CTLs (Fig. 4c), intratumoral injection of OVA 257-264 peptide was effective against RMA cells, which are OVA-negative tumor cells.
The RMA tumors cells that were injected intratumorally with OVA 257-264 peptide demonstrated significant tumor growth inhibition, compared with mice without intratumoral injection of OVA 257-264 peptide (P \ 0.05). The survival rate in the treatment group was significantly better than that in the control groups (P \ 0.05) (Fig. 4b, d). The group that did not receive OVA 257-264 peptide vaccine but that received intratumoral peptide injection showed a partial treatment effect (Fig. 4b).
To obtain direct evidence that intratumoral peptide injection leads to local accumulation of antigen-specific CTLs, an OVA tetramer assay was performed using an adoptive cell transfer model (Fig. 4e). Two RMA tumors were bilaterally implanted per mouse. One tumor was injected with the OVA 257-264 peptide plus IFA, and the other tumor with IFA alone (Fig. 4f). As shown in Fig. 4g, the tumor that underwent both adoptive cell transfer of activated OT-I CTLs and intratumoral injection of the OVA peptide contained more OVA-specific CTLs than the other tumors. Local accumulation of OVA-specific CTLs after intratumoral injection of the OVA 257-264 peptide was confirmed by OVA tetramer assay.
Neither toxic signs nor death due to intratumoral injection of the OVA 257-264 peptide was observed. Moreover, to evaluate the risk of autoaggression by intratumoral peptide injection, the tissues of treated mice in an adoptive cell transfer model were pathologically examined. The spleen, brain, lung, heart, liver, kidney, and tumor were critically scrutinized, and the findings were compared with those from  Tx-1 Tx-2 Tx-3 h Immunohistochemical staining of CD8 in tumor and normal tissues. Spleen was used as positive control. Scale bars, 50 lm CD8 ? T-cells was not observed (Fig. 4h). These results suggest that peptide from intratumoral injection did not spread into normal tissues.
The effect of antigen spreading to another tumor after intratumoral peptide injection Using an adoptive cell transfer model, we assessed the possibility of antigen-spreading effect after intratumoral peptide injection, as depicted in Fig. 5a. Two RMA tumors were bilaterally and metachronously implanted per mouse, and only the first tumors received intratumoral injection of the OVA 257-264 peptide. The sizes of the second tumors were compared with those from mice that received intratumoral injection of IFA alone (Fig. 5b). Whereas the second tumors were established 14 days after the second tumor inoculation in three out of four control mice, all four peptide-loaded mice that had received intratumoral OVA 257-264 peptide injection into their first tumor completely rejected the challenge of the second tumor, which did not receive intratumoral OVA 257-264 peptide injection itself (Fig. 5c).
To confirm the hypothesis of antigen spreading, an IFN-c ELISPOT assay was performed. RMA tumor-bearing f e  C57BL/6 mice that had received adoptive transfer of OT-I CTLs and intratumoral injection of OVA 257-264 peptide were killed, and their spleens were obtained 21 days after adoptive transfer and 7 days after the last intratumoral injection. CD8 ? T-cells, isolated from the spleen cells using anti-CD8a magnetic beads, were incubated with irradiated RMA cells for 3 days. CD8 ? T-cells were separated from RMA cells using anti-CD8a magnetic beads before the assay. An IFN-c ELISPOT assay was performed in duplicate using CD8 ? T-cells as effector cells and RMA cells as target cells (Fig. 5d). The mice that had received intratumoral injection of OVA 257-264 peptide showed a significant response to OVA-negative RMA tumor cells compared with control mice that had received intratumoral injection of IFA alone (P \ 0.05). The observed induction of RMAderived antigen-specific CTLs provides evidence that antigen spreading occurred by treatment with intratumoral OVA 257-264 peptide and intravenous OT-I CTLs (Fig. 5e).  Adoptive cell transfer on the peptide density of tumor cells in an HLA class I-restricted manner. In other words, intratumoral peptide injection enhances the antigenicity of tumor cells, regardless of whether the tumor cells originally expressed the antigen. To the best of our knowledge, this is the first study to show the efficacy of intratumoral peptide injection in detail. A previous report demonstrated that peptide injection around a tumor assisted the activity of low-avidity CTLs in an immunodeficient mouse model [21]. In addition, we demonstrated the advantage as a therapeutic modality combined with antigen-specific cancer immunotherapy without any adverse reactions associated with this procedure in mice. Intratumoral peptide injection can strengthen the efficacy of every kind of antigen-specific cancer immunotherapy and may be a useful therapeutic option. This is the first study to describe anticancer treatment with CMV-derived peptide-specific CTLs. Virus-derived antigens, which are exogenous antigens, usually have stronger antigenicity than tumor-associated autoantigens. Therefore, virus-derived antigen-specific CTLs are easier to induce [22]. Theoretically, every kind of antigen is applicable to our procedure unless it is expressed in healthy human cells. However, it is unclear whether post-CMVinfected lesions are safe from CMV-specific CTL cytotoxicity. Further investigations are necessary regarding the possible clinical use of exogenous antigens, such as CMVderived peptides.
We used NaHCO 3 , which is known to have therapeutic effects against tumors [23,24], as a peptide diluent. However, our data demonstrated the efficacy of intratumoral peptide injection, because control animals which underwent intratumoral injection of IFA alone or IFA plus an irrelevant peptide also received NaHCO 3 .
In an in vivo tumor growth inhibition assay using a peptide vaccine model, the group that did not receive the OVA 257-264 peptide vaccine but that received intratumoral peptide injections showed a partial treatment effect. This indicates that intratumoral or peritumoral antigen-presenting cells recognized intratumorally injected OVA 257-264 peptide and induced OVA 257-264 peptide-specific CTLs after three intratumoral peptide injections. However, we showed in this study that intratumoral peptide injection attracted more OVA 257-264 peptide-specific CTLs and was more effective when combined with peptide vaccines or adoptive cell transfer therapies.
A limitation of intratumoral peptide injection is its delivery method. First, immunotherapy is expected to contribute toward cancer therapy especially in the early stages or in the prevention of recurrence, in which cancer sites, the so-called ''micro lesions,'' are undetectable by imaging modalities. However, intratumoral peptide injection must be limited to the tumors, which are detectable by imaging modalities, and can be approached with a needle. Second, it is difficult to spread the peptides over the whole tumor by intratumoral injection, especially against large tumors. Moreover, it is difficult to approach all of the multiple tumors. This procedure might limit the ability of immunotherapy as a systemic therapy. If a novel method of delivering peptides to tumor cells selectively through a systemic route is established in the future due to advances  in drug-delivery technologies, this method will become more suitable for clinical application.

Additional therapeutic efficacy
Another limitation is that it requires the presence of MHC class I molecules. The potential loss of MHC class I expression in tumors would lead theoretically to the failure of this approach. Previous reports have indicated that 61-85 % of breast cancers had loss of or decreased HLA class I expression [25][26][27]. On the other hand, the downregulation of HLA class I was less frequently observed in other cancers [27][28][29][30]. Before clinical application, it is necessary to select cancers in which HLA class I expression is sufficiently high.
Antigen-spreading effects have been observed following anticancer immunotherapy [31][32][33][34]. The second tumor challenge is easily rejected due to immunological memory. Therefore, we fixed the number of implanted tumor cells as the second tumors could be established. In this study, we report evidence of an antigen-spreading effect after intratumoral peptide injection. If this antigen-spreading effect is sufficiently steady and reliable, intratumoral peptide injection may even be effective against imaging-invisible or unapproachable tumors.
In conclusion, intratumoral peptide injection is an attractive strategy for enhancing tumor cell antigenicity. It can induce additional peptide loading onto tumor cells, making tumor cells more antigenic for antigen-specific CTL activity against tumor cells. Moreover, it may be a useful option for improvement in antigen-specific cancer immunotherapy against solid tumors (Fig. 6).