Induction of Antigen-Specific Immunity with a Vaccine Targeting NY-ESO-1 to the Dendritic Cell Receptor DEC-205

NY-ESO-1 expression screening

Immunohistochemistry (IHC) and polymerase chain reaction (PCR) testing of archived tumor tissue for 280 screened patients showed that melanoma and synovial sarcoma frequently expressed NY-ESO-1, and results were highly concordant between the two assays (Table 1). However, colorectal, ovarian, breast, and bladder cancers were rarely strongly positive and showed poor correlation between the assays for the target antigen based on the relatively small number of samples tested.

Patient enrollment and disposition

A total of 45 patients were enrolled and treated on study. Figure 1 illustrates the allocation of patients to each treatment cohort. In cohorts 1 to 3, 23 patients received escalating doses of CDX-1401 (0.1, 1, or 3 mg) in combination with topical resiquimod. A dose of 1 mg was subsequently chosen for cohorts 4 to 6, in which 23 patients received CDX-1401 with subcutaneously administered adjuvant, either poly-ICLC, resiquimod, or both.

As shown in Table 2, nearly half of the enrolled patients had melanoma. Additional cancer types included ovarian, sarcoma, non–small cell lung, and colorectal cancers. Distant metastases were noted in 87% of the treated patients, consistent with the presence of advanced disease, as required for enrollment. Expression of the NY-ESO-1 antigen in tumor specimen was confirmed by IHC or PCR in 27 of 42 (64%) of the treated patients with available tissue. In cohorts 1 to 4, 12 of 30 (40%) had confirmed tumor NY-ESO-1 expression, whereas revised entry criteria required that all patients in cohorts 5 and 6 had confirmed tumor NY-ESO-1 expression.

Dosing, toxicity, and pharmacokinetics

Dose escalation proceeded as planned with no dose-limiting toxicities (DLTs). Of the 45 enrolled patients, 41 completed at least one cycle of CDX-1401 treatment, defined as a 6-week course of treatment with CDX-1401 and applicable adjuvant(s). Ten patients with stable disease were re-treated, with a median of 10 CDX-1401 doses (range, 6 to 20).

There were no cases of treatment discontinuation due to toxicity, and treatment-related toxicities were all of grade 1 or 2 severity. As shown in Table 3, the most frequent treatment-related adverse events were administration site reaction, fatigue, nausea, and chills.

Table 3. Adverse events considered related to study treatment (CDX-1401, resiquimod, and/or poly-ICLC).

As expected from these relatively small doses (up to 3 mg), pharmacokinetic analysis revealed no detectable levels of circulating CDX-1401 (limit of detection, 0.08 μg/ml) in samples obtained at 2, 4, or 24 hours after administration of CDX-1401.

Induction of humoral immunity to NY-ESO-1

As shown in Fig. 2, most patients (79%) had NY-ESO-1–specific immunoglobulin G (IgG) titers after vaccination, with high titers (≥1:10,000) in 52% and very high titers (≥1:100,000) in 33% of patients. Similarly, strong humoral immunity developed in each cohort and in patients with or without confirmed NY-ESO-1 expression in their tumor. Patients with NY-ESO-1⁺ tumors frequently (54%) had anti–NY-ESO-1 antibodies present at baseline, including 5 patients with titers above 1:100,000. Isotype analysis of patient’s samples with anti–NY-ESO-1 responses after treatment revealed a predominant IgG1 response, with several patients also showing IgG2, IgG3, and IgG4 anti–NY-ESO-1 antibodies (table S1).

Induction of cellular immunity to NY-ESO-1

Cellular immunity to NY-ESO-1 was determined by interferon-γ (IFN-γ) enzyme-linked immunospot (ELISpot) on purified blood mononuclear cells from patient samples collected before and after CDX-1401 administration. Patient samples were qualified by demonstrating responsiveness to CD3 stimulation and to a pool of major histocompatibility complex class I peptides derived from human viruses (fig. S1). We detected NY-ESO-1–specific T cell responses in 56% patients with evaluable pre- and posttreatment samples (Fig. 3A). As seen with humoral immunity, significant NY-ESO-1–specific T cell responses were observed in most cohorts, suggesting that even low doses of CDX-1401 were immunogenic and both poly-ICLC and resiquimod can provide appropriate immune activation to drive T and B cell responses. Durability of the T cell response was demonstrated in two patients from whom samples from additional cycles of CDX-1401 treatment were available. In these patients, the induction of NY-ESO-1–specific T cells was maintained through three cycles (about 7 months) of treatment (Fig. 3B).

In a few patients with sufficient lymphocytes, it was possible to further characterize NY-ESO-1 T cell responses by intracellular cytokine and HLA pentamer staining. Figure 4A shows a patient sample with an NY-ESO-1–specific response by both CD4⁺ and CD8⁺ cells as demonstrated by intracellular cytokine staining for IFN-γ and TNF-α (tumor necrosis factor–α). Of the five patients tested, three had both CD8⁺ and CD4⁺ NY-ESO-1 T cell responses, whereas two patients only showed a dominant CD4⁺ response (Fig. 4B).

Peripheral blood mononuclear cells (PBMCs) from HLA-A2.1⁺ patients were stained directly (that is, no in vitro stimulations) with an NY-ESO-1 peptide–specific pentamer or irrelevant control pentamer. Of the 16 patients stained with the pentamers, only 3 showed NY-ESO-1 pentamer–positive CD8 T cells that increased in numbers after vaccination (fig. S2), suggesting a low abundance of circulating T cells reactive to the well-known immunodominant epitope, NY-ESO_157–165, and possible reactivity against epitopes presented by other HLA class I/II alleles.

Clinical activity

Stable disease was noted for 13 patients, distributed among treatment cohorts (Fig. 2). Seven of these patients had melanoma, two had colorectal cancer, and the remaining four had myeloma, bladder cancer, cholangiocarcinoma, and non–small cell lung cancer. Median duration of stable disease was 6.7 months (range, 2.4+ to 13.4 months). Remarkably, two of the melanoma patients were also observed to have some tumor shrinkage after one cycle of CDX-1401 treatment, with both patients having about 20% shrinkage in the sum of the diameters of target lesions. One patient with regression of a tonsil lesion subsequently showed clinical progression before additional treatment with CDX-1401, but was still alive at the 2-year follow-up. The second patient had regression in a lung lesion and received a second cycle of treatment before progression indicated by the appearance of a new lesion.

The detection of NY-ESO-1 expression in tumor tissue did not appear to correlate with patient outcome. Stable disease was seen in 7 of 27 (26%) of the patients with NY-ESO-1 expression, and in 5 of 15 (33%) of those lacking NY-ESO-1 expression. Note that these tested tissue samples represented archival specimens obtained at a median of 17.5 months (range, 2.3 months to 10.2 years) before study entry. Similarly, the rate of disease stabilization was similar for most patients who developed or maintained NY-ESO-1–specific humoral response (10 of 34, 29%) and for the few patients who did not (3 of 8, 38%).

However, the proportion of patients with stable disease was higher for those who maintained or developed NY-ESO-1–specific T cell responses (Fig. 3). Stable disease was seen in three of six (50%) patients who entered the study with preexisting cellular immunity to NY-ESO-1, and all three had increased responses while on study. For the remaining 13 patients who developed cellular immunity while on treatment, 6 (46%) experienced stable disease. In contrast, stable disease was seen in only 2 (13%) of the 15 patients who did not develop cellular immunity to NY-ESO-1. Four of six (67%) patients who displayed the strongest responses (>50 IFN-γ spots per 2 × 10⁵ PBMCs) also experienced stable disease. The association of cellular response and stable disease does not appear to be a consequence of extended duration of therapy. Peak responses were observed in the first treatment cycle for seven of the nine patients with stable disease who developed cellular immunity.

Eleven patients completed study follow-up at 2 years, whereas one patient remains in follow-up. During study follow-up, six melanoma patients went on to receive anti–CTLA-4 mAb within 3 months of the last CDX-1401 dose, and four of these patients were reported to experience a partial response (PR) or complete response (CR) by the Response Evaluation Criteria in Solid Tumors version 1.1 (RECIST 1.1) (32) or irResponse (33) criteria (table S2). In addition, two non–small cell lung cancer patients who received investigational checkpoint blockade within 2 months of discontinuing CDX-1401 were also reported to experience PR. All six of these responding patients had tumors confirmed to express NY-ESO-1. Five also developed NY-ESO-1–specific cellular response, whereas four also developed or maintained NY-ESO-1–specific humoral response, by the end of treatment with CDX-1401.

Patient population

The study was open to men and women at least 18 years of age, with cancer types known to express NY-ESO-1, including, but not limited to, bladder, breast, ovary, non–small cell lung, myeloma, sarcoma, and melanoma. Eligible patients had advanced malignancies that had progressed after any standard curative or salvage therapies. Confirmation of tumor NY-ESO-1 expression (>5% positive cells) by PCR and/or IHC at a central laboratory was required for cohorts 5 and 6. For cohorts 1 to 4, this analysis was performed retrospectively. Exclusionary conditions included Eastern Cooperative Oncology Group (ECOG) performance status >1, active central nervous system metastases or potential alternate malignancy, autoimmune disease, and concurrent treatment with immunosuppressive or immunomodulatory agents. Before the conduct of any protocol-specific procedures, all patients signed written informed consent after the nature and possible consequences of the studies were explained.

Study design and procedures

This phase 1 study was designed to examine the safety and tolerability profile of CDX-1401 in combination with the TLR agonists, resiquimod and/or poly-ICLC, in patients with malignancies known to express NY-ESO-1. All patients who received at least one dose of study drug were included in safety analyses. Additional objectives of this study included an evaluation of CDX-1401–induced immune responses, antitumor activity, and pharmacokinetics.

The study consisted of six sequentially enrolled (nonrandomized) open-label treatment cohorts (Fig. 1). An initial dose-escalation phase (cohorts 1 to 3; CDX-1401 in combination with topical resiquimod) was conducted to select a CDX-1401 dose level of interest for further study. Subsequent cohorts evaluated CDX-1401 in combination with poly-ICLC, escalating doses of parenteral resiquimod, or both to assess an optimal adjuvant regimen.

Safety parameters assessed at baseline and study visits included physical examination, vital signs, ECOG performance status, hematology, blood chemistry, urinalysis, thyroid function, and the incidence and severity of adverse events. Toxicity was graded according to the National Cancer Institute Common Terminology Criteria for Adverse Events version 3.0. DLT was defined as any treatment-related grade ≥3 nonhematologic adverse event excluding alopecia, or persistent grade 4 hematologic adverse event. Assessment of immunological activity included measurements of cellular and humoral response to NY-ESO-1. Standard disease-specific tumor evaluation was performed according to RECIST 1.1 within 28 days before each re-treatment cycle and at a minimum frequency of 12 weeks. Patients were followed for up to 2 years for assessment of safety, response to subsequent therapy, and survival.

Study drugs and dosing

In all cohorts, a treatment cycle was defined as a 6-week course of treatment with CDX-1401 and applicable adjuvant(s), followed by a 6-week rest. Additional 12-week cycles were allowed for patients who continued to meet all eligibility criteria and who had not experienced DLT/intolerance or progression of disease.

CDX-1401 (Celldex Therapeutics) was administered via intracutaneous injection (a combination of intradermal and subcutaneous injection, with about half of the planned dose administered by each route) within a 5 χ 5–cm area on the extremities or the abdomen. CDX-1401 was given biweekly for a total of four administration sessions over the ~6-week treatment period (Fig. 1). Dose levels were 0.1, 1.0, and 3.0 mg. Topical resiquimod (3M Pharmaceuticals; also known as R-848; S-28463; 4-amino-2-ethoxymethyl-α, α-dimethyl-1H-imidazo[4,5-C]quinoline-1-ethanol) was applied topically at a dose of 500 μg (250 mg of 0.2% gel) immediately after CDX-1401 vaccination and again about 24 hours later. Each topical resiquimod application was to remain on the treatment site for about 8 hours, unless local toxicity warrants premature removal. Resiquimod for injection (0.02 to 0.5 mg in a 0.25-ml volume) was administered as a subcutaneous injection after CDX-1401 vaccination. Poly-ICLC (Oncovir; Hiltonol) was administered subcutaneously at a dose of 2 mg (1.0 ml) after CDX-1401 vaccination and again 24 hours later.

Pharmacokinetics

Serum samples for assessment of circulating CDX-1401 were obtained before vaccination and at 2, 4, and 24 hours after the first vaccination, and before vaccination for all other vaccinations in the first cycle. A functional ELISA was developed to test serum and plasma samples on a 96-well microtiter plate coated with a fragment of DEC-205. Bound CDX-1401 was detected with anti–NY-ESO-1 antibody followed by goat anti-human IgG-HRP (horseradish peroxidase) and visualized with a 3,3’,5,5’-tetramethylbenzidine (TMB) substrate. The assay has a lower detection limit of 0.08 μg/ml.

NY-ESO-1 expression

The IHC method was optimized, and expression of NY-ESO-1 was verified with control sections from breast cancer, melanoma, and ovarian carcinoma sections at Clarient Inc. Briefly, 4- to 5-μm cuts of standard formalin-fixed, paraffin-embedded tissue specimens were mounted on charged glass slides, deparaffinized, and washed per standard methods. After heat-induced antigen retrieval at 100°C, samples were stained for 30 min with a murine mAb specific for NY-ESO-1 (E978, Invitrogen). Known positive cell line HT-1080 fibrosarcoma and negative cell line SiHa squamous cell carcinoma were used for staining verification. Expression of NY-ESO-1 was visualized with the Leica Bond Refine-HRP detection system and DAB as chromagen and hematoxylin as counterstain. Tumors were manually evaluated on the basis of the percentage of positive cells with a standard bright-field microscope (0 to 100%, intensity 0 to 3+). A sample with greater than or equal to 5% positive cells was considered positive for NY-ESO-1 protein.

For reverse transcription PCR (RT-PCR), total RNA was prepared from 10-μm-thick, formalin-fixed, paraffin-embedded tumor sections and from control tumor cell lines with the in-house methods (Clarient Inc.) Quantitative RT-PCR was performed with a TaqMan assay and an ABI Prism 7900 HT system (Applied Biosystems) with a TaqMan RNA-to-C_T 1-Step kit (Life Technologies). TaqMan gene expression primers and probes for NY-ESO-1 (Hs00265824_m1) and human GUSB (HS99999908_m1) were purchased from Life Technologies. As a positive control, total RNA extracted from HT-1080 cells transformed with NY-ESO-1, and a negative control from HL-60 cell line extracts, determined not to express the target, were used in every run, as well as a negative template control to test for contamination. GUSB expression was included as a sample amplification control. For inclusion in the trial CDX1401-01, any sample with an NY-ESO-1 signal greater than the threshold was considered positive.

Humoral immune response analysis

The magnitude and isotype of the anti–NY-ESO-1 antibody response were determined from patient serum samples. Sample time points were day 0 (before dosing), day 28 (2 weeks after the second CDX-1401 dose in cycle 1 only), day 42 (2 weeks after the third CDX-1401 dose), and days 70 to 77 (about 1 month after the fourth CDX-1401 dose). NY-ESO-1–specific antibodies were detected by adding serially diluted serum to ELISA plates coated with NY-ESO-1 and developed with a goat anti-human IgG (H&L) peroxidase and TMB staining. A posttreatment sample was considered positive for the presence of anti–NY-ESO-1 antibodies if the mean absorbance value was greater than the mean value of the predose sample, plus 3 SDs or +0.1, whichever was greater. The titer of the sample was defined as the reciprocal of the dilution with an absorbance value closest to a preset background of 0.1. Anti–NY-ESO-1 antibody isotypes were determined with isotype-specific HRP-conjugated goat anti-human IgG1, IgG2, IgG3, IgG4, and IgM. Analysis was carried out on posttreatment samples (1:50 dilution) from all patients with positive NY-ESO-1 titers (n = 26).

Cellular immune response analysis of NY-ESO-1 synthetic peptide library

A library of HLA class I–binding peptides derived from NY-ESO-1 protein sequence previously described (29) was used to prepare high-quality and 95% pure synthetic peptides (ProImmune). Peptides were resuspended in anhydrous dimethyl sulfoxide at 10 mg/ml, aliquoted, flushed with nitrogen, sealed, and stored at −40°C until use.

ELISpot assay

ELISpot assay was used to screen the patient blood samples for immune responses. For every patient in the clinical trial, pre- and post-dose blood samples were collected, and multiple post-dose, in addition to the predose, samples were tested for NY-ESO-1–specific IFN-γ secretion. In each ELISpot assay, PBMCs were first presensitized in a 7-day in vitro sensitization culture in the presence of low-dose IL-2 (10 ng/ml). The NY-ESO-1 peptide pool comprised 28 overlapping peptides covering the full length of the NY-ESO-1 protein (29). In the assay, APCs (T-depleted PBMCs) were pulsed with NY-ESO-1 peptides and negative control peptides (WRKY47, ProImmune). Anti-CD3 mAb and CEF-I peptide pool served as positive controls in the assay. The CEF-I pool contains 32 HLA class I peptides from human cytomegalovirus, Epstein-Barr virus, and influenza virus (PANATecs GmbH). ELISpot plates were washed as per kit instructions. Spot counts were evaluated with the Zeiss system (ZellNet Consulting). Ambiguities arising from confluence issues were resolved with the following formula: total spot number = spot count + 2 × [spot count × % confluence/(100% − % confluence)].

Intracellular cytokine staining

Patient PBMCs were cultured for 7 days as described above and stained for intracellular cytokines. Briefly, presensitized cells were first restimulated with T-depleted APCs pulsed with various peptides and then treated with GolgiPlug (BD Biosciences). Cells were then stained with anti-CD3, anti-CD4, and anti-CD8 mAbs, fixed, and permeabilized followed by staining with fluorochrome-conjugated antibodies to IFN-γ (clone B27) and TNF-α (clone mAb 11) (BD Biosciences). Stained cells were then analyzed by flow cytometry with FACSCanto II (BD Biosciences).

HLA-A2–NY-ESO-1 pentamer staining

For pentamer staining, the A*0201/NY-ESO-1_157–165 (SLLMWITQV) pentamer-RPE (retinal pigment epithelial) and control A*0201/[negative] pentamer-RPE were purchased from ProImmune. The patient PBMCs were thawed and stained directly with the pentamers without culturing. About 2 χ 10⁶ cells were resuspended in 500 μl of phosphate-buffered saline and stained with 5 μl of the pentamers for 30 min and further stained for CD3 and CD8. A CD19-positive gate was used to exclude nonspecific binding by B cells with CD19 antibody (clone HIB19, BD Biosciences).

Statistical analysis

For this phase 1 trial, the statistical analysis consisted of descriptive summaries of the safety and immunologic data. Adverse events were tabulated on the basis of the standardized terms assigned by the Medical Dictionary for Regulatory Activities (MedDRA). Tumor responses were evaluated and reported in accordance with RECIST (version 1.1) (32).