Purpose: Antitumor activity of cancer immunotherapies may elicit immune responses to nontargeted (secondary) tumor antigens, or antigen spread. We evaluated humoral antigen spread after treatment with sipuleucel-T, an immunotherapy for asymptomatic or minimally symptomatic metastatic castration-resistant prostate cancer (mCRPC), designed to target prostatic acid phosphatase (PAP; primary antigen).

Experimental Design: Serum samples from patients with mCRPC enrolled in the placebo-controlled phase III IMPACT study (evaluable n = 142) were used to assess humoral antigen spread after treatment with sipuleucel-T. Immunoglobulin G (IgG) responses to self-antigens (including tumor antigens) were surveyed using protein microarrays and confirmed using Luminex xMAP. IgG responses were subsequently validated in ProACT (n = 33), an independent phase II study of sipuleucel-T. Association of IgG responses with overall survival (OS) was assessed using multivariate Cox models adjusted for baseline prostate-specific antigen (PSA) and lactate dehydrogenase levels.

Results: In patients from IMPACT and ProACT, levels of IgG against multiple secondary antigens, including PSA, KLK2/hK2, K-Ras, E-Ras, LGALS8/PCTA-1/galectin-8, and LGALS3/galectin-3, were elevated after treatment with sipuleucel-T (P < 0.01), but not control. IgG responses (≥2-fold elevation posttreatment) occurred in ≥25% of patients, appeared by 2 weeks after sipuleucel-T treatment, and persisted for up to 6 months. IgG responses to PSA and LGALS3 were associated with improved OS in sipuleucel-T–treated patients from IMPACT (P ≤ 0.05).

Conclusions: Sipuleucel-T induced humoral antigen spread in patients with mCRPC. IgG responses were associated with improved OS in IMPACT. The methods and results reported may identify pharmacodynamic biomarkers of clinical outcome after sipuleucel-T treatment, and help in clinical assessments of other cancer immunotherapies. Clin Cancer Res; 21(16); 3619–30. ©2015 AACR.

See related commentary by Hellstrom and Hellstrom, p. 3581

This article is featured in Highlights of This Issue, p. 3573

Translational Relevance

Clinical benefits of cancer immunotherapies are often not evident with the commonly used radiographic measures of objective tumor response (e.g., RECIST or WHO criteria). Biomarkers that are indicative of immunologic antitumor activity may be useful in assessing the clinical benefit of this class of therapeutics. The development of immune responses to secondary tumor antigens not directly targeted by a therapy, or antigen spread, may indicate tumor cell destruction and provide biomarkers of clinical benefit. We show here that serum antibody (IgG) responses to secondary tumor antigens (humoral antigen spread) can be detected within weeks after treatment with sipuleucel-T, an immunotherapy for metastatic castration-resistant prostate cancer, and such responses can be associated with improved overall survival. These results may enable identification of pharmacodynamic biomarkers of clinical outcome after treatment with sipuleucel-T. More generally, they may constitute a useful tool for evaluating the efficacy of treatment regimens that include cancer immunotherapies.

Methods for the assessment of efficacy of cancer immunotherapies are critical in both clinical development and practice. Radiographic measures for objective responses (e.g., RECIST or WHO criteria) have limitations in their assessment of the effects of immunotherapies that are designed to stimulate immune responses against the tumor (1–8). Several clinical studies have now shown that immunotherapies can result in improved overall survival (OS), without frequent objective responses or improving disease progression as assessed by radiography (2, 3, 9). Therefore, appropriate modifications of existing methods or alternative biomarkers of clinical outcome are needed that are indicative of these agents' immunologic mechanism of action (2, 3, 6–8, 10).

Evidence of immune responses to nontargeted (secondary) antigens following treatment with an immunotherapy, referred to as antigen (or epitope) spread, may enable the identification of novel biomarkers of clinical outcome (11–17). Originally described in autoimmune diseases, antigen spread is believed to play an important role in the progression and pathogenesis of immune-related disorders (18–21) and in the protection against infectious diseases (22, 23). In the context of antitumor immune responses, antigen spread to tumor-associated antigens (TAA) may be indicative of tumor cell killing, antigen release, and subsequent priming of self-reactive T and/or B lymphocytes against TAAs (21, 24, 25). It has been suggested that treatment-induced antigen spread may be associated with improved clinical outcomes (11–17), but evidence from controlled clinical studies is currently lacking. Here, we report an investigation of antigen spread and its association with OS following treatment with sipuleucel-T, an autologous cellular immunotherapy for the treatment of patients with asymptomatic or minimally symptomatic metastatic castration-resistant prostate cancer (mCRPC; ref. 4).

Sipuleucel-T, designed to target the prostate antigen, prostatic acid phosphatase (PAP, the primary antigen), prolongs OS in patients with mCRPC, but without significant improvement in objective measures of disease progression (4). Immune responses to PAP have been shown to be associated with OS in patients who received sipuleucel-T (26), but antigen spread to TAAs following treatment may provide a more relevant measure of an effective antitumor immune response (13, 25). Here, we show that sipuleucel-T, but not control, elicited serum antibody [immunoglobulin G (IgG)] responses to nontargeted tumor antigens, including prostate-specific antigen (PSA; also known as KLK3), KLK2/hK2 (KLK2), K-Ras, E-Ras, LGALS8/PCTA-1/galectin-8 (LGALS8), and LGALS3/galectin-3 (LGALS3). These responses were observed at 2 weeks and up to 6 months after treatment with sipuleucel-T. Sipuleucel-T–induced IgG responses to PSA and LGALS3 were associated with improved OS in IMPACT.

These results further the understanding of the mechanism of action of sipuleucel-T and may help to identify biomarkers of clinical outcome for this therapy. The methods and results presented here may also facilitate the identification of serum biomarkers of clinical outcome for other cancer immunotherapies. Such easily accessible biomarkers of clinical outcome may meet a critical need for assessing the in vivo mechanism and efficacy of this class of cancer therapies (10, 11, 27). Antigen spread may also help in the identification of TAAs that can be targeted by immunotherapies in the future.

Treatments, clinical trials, and biologic specimens

Sipuleucel-T is prepared by culturing autologous peripheral blood mononuclear cells (PBMC) with PA2024, a fusion protein comprising PAP and granulocyte colony-stimulating factor (GM-CSF; ref. 4). Sipuleucel-T was administered as a total of three infusions (or doses) at approximately 2-week intervals (Supplementary Fig. S1). In IMPACT, three infusions of a control product were administered at the same time intervals as sipuleucel-T (4).

IMPACT (Clinicaltrials.gov identifier NCT00065442) was a multicenter, randomized, double-blind, phase III clinical trial of sipuleucel-T in 512 patients with asymptomatic or minimally symptomatic mCRPC (4), and randomized 2:1 to the sipuleucel-T and control arms. For the evaluation of immune responses, the original protocol was amended to allow collection of peripheral samples from patients who provided consent (4, 26). Median follow-up duration for the patients from whom serum samples were available was 33.7 months. ProACT (NCT00715078) is a multicenter, randomized, single-blind, phase II clinical trial of sipuleucel-T manufactured with three different concentrations of PA2024 and administered to a total of 122 patients with advanced mCRPC. All sipuleucel-T–treated patients included in the analyses reported here (IMPACT and ProACT) received sipuleucel-T manufactured using the antigen concentration approved by the FDA and EMA (European Medicines Agency).

Serum samples were collected from patients at time points specified under each study protocol (see Supplementary Methods). In both studies, a pretreatment or baseline sample was collected at the time of registration. In IMPACT, samples were collected at approximately 2, 10, and 22 weeks after treatment (6, 14, and 26 weeks after initiation of treatment, respectively). In ProACT, samples were collected at approximately 4, 12, and 20 weeks after treatment (8, 16, and 24 weeks after initiation of treatment). Other than pretreatment, all time points were posttreatment (e.g., “week 10” refers to week 10 after completion of treatment).

Pre- and posttreatment serum samples were available from a total of 224 patients from IMPACT (4, 26) who provided consent; 155 in the sipuleucel-T arm and 69 in the control arm. Serum samples were collected from patients only up to the time of objective disease progression (4); consequently, samples were available from more patients at earlier time points than at later time points. In IMPACT, pairs of serum samples were evaluable as follows: (i) pretreatment and week 2, n = 204 (142 sipuleucel-T and 62 control patients); (ii) pretreatment and week 10, n = 132 (93 sipuleucel-T, 39 control); and (iii) pretreatment and week 22, n = 76 (60 sipuleucel-T, 16 control).

In ProACT (all treated with sipuleucel-T), pairs of serum samples were evaluable as follows: (i) pretreatment and week 4, n = 33; (ii) pretreatment and week 12, n = 26; and (iii) pretreatment and week 20, n = 19.

Baseline demographic and clinical characteristics of the evaluated patients from IMPACT and ProACT are provided in Supplementary Tables S1 and S2 respectively.

Assessment of serum IgG responses using high-content protein microarray (ProtoArray)

ProtoArray v5.0 (Life Technologies Corporation) contains roughly 9,000 recombinant human proteins produced using a baculovirus expression system and printed on a glass slide under nondenaturing conditions (15, 28–30). ProtoArray was used to measure levels of IgG to these proteins in pretreatment and posttreatment serum samples from patients in IMPACT (Fig. 1 and Table 1). IgGs bound to the proteins on the array were detected using anti-human IgG conjugated to Alexa Fluor 647. Serum samples (30 μL) were assayed by Life Technologies Corporation at 1:500 dilution (see Supplementary Methods for experimental details, data acquisition, and normalization). Normalized signal intensities were log2-transformed before analysis.

Figure 1.

Schematic for identification of serum IgG responses to secondary antigens in IMPACT, using ProtoArray. IgG levels in pre- and posttreatment (Tx) serum samples were compared with identify IgG responses against specific proteins (antigens) on the ProtoArray (≥2-fold average increase in serum IgG level posttreatment relative to pretreatment, with FDR ≤ 10%). The number of antigens against which IgG responses were observed (y-axis) at the three posttreatment time points (x-axis) are shown for patients in the sipuleucel-T and control arms.

Figure 1.

Schematic for identification of serum IgG responses to secondary antigens in IMPACT, using ProtoArray. IgG levels in pre- and posttreatment (Tx) serum samples were compared with identify IgG responses against specific proteins (antigens) on the ProtoArray (≥2-fold average increase in serum IgG level posttreatment relative to pretreatment, with FDR ≤ 10%). The number of antigens against which IgG responses were observed (y-axis) at the three posttreatment time points (x-axis) are shown for patients in the sipuleucel-T and control arms.

Close modal
Table 1.

Secondary antigen candidates identified in patients from IMPACT, using ProtoArray

Antigen annotationProtoArray data
Antigen symbolAntigen nameAverage fold changePFDR (%)Rank by average fold change
LGALS3 Galectin-3 4.94 2.84E−10 2.05E−04 
CACNG1 Calcium channel, voltage-dependent, gamma subunit 1 4.83 3.60E−07 2.69E−02 
ANPEP Aminopeptidase N (CD13) 4.73 5.69E−08 6.85E−03 
FBXO6 F-box protein 6 3.88 1.62E−08 2.68E−03 
ECE1 Endothelin converting enzyme 1 3.42 3.73E−06 3.26E−02 
ERAS Embryonic stem-cell–expressed Ras 2.72 4.16E−06 3.26E−02 15 
TSPAN13 Tetraspanin 13 2.66 1.73E−06 3.26E−02 17 
PAP Prostatic acid phosphatase 2.55 1.18E−06 3.15E−02 23 
LGALS8/PCTA-1 Galectin-8/prostate carcinoma tumor antigen 1 2.19 2.89E−05 5.09E−02 68 
KRAS Kirsten rat sarcoma viral oncogene homolog 2.10 3.20E−06 3.26E−02 99 
KLK2/hK2 Kallikrien-related peptidase 2 2.04 3.67E−05 5.50E−02 138 
Antigen annotationProtoArray data
Antigen symbolAntigen nameAverage fold changePFDR (%)Rank by average fold change
LGALS3 Galectin-3 4.94 2.84E−10 2.05E−04 
CACNG1 Calcium channel, voltage-dependent, gamma subunit 1 4.83 3.60E−07 2.69E−02 
ANPEP Aminopeptidase N (CD13) 4.73 5.69E−08 6.85E−03 
FBXO6 F-box protein 6 3.88 1.62E−08 2.68E−03 
ECE1 Endothelin converting enzyme 1 3.42 3.73E−06 3.26E−02 
ERAS Embryonic stem-cell–expressed Ras 2.72 4.16E−06 3.26E−02 15 
TSPAN13 Tetraspanin 13 2.66 1.73E−06 3.26E−02 17 
PAP Prostatic acid phosphatase 2.55 1.18E−06 3.15E−02 23 
LGALS8/PCTA-1 Galectin-8/prostate carcinoma tumor antigen 1 2.19 2.89E−05 5.09E−02 68 
KRAS Kirsten rat sarcoma viral oncogene homolog 2.10 3.20E−06 3.26E−02 99 
KLK2/hK2 Kallikrien-related peptidase 2 2.04 3.67E−05 5.50E−02 138 

NOTE: Increase in serum levels of IgGs against 10 candidate secondary antigens at week 10 after completion of sipuleucel-T treatment, as measured by ProtoArray. Data for PAP (primary antigen) are shown for reference. Columns indicate (for serum level of anti-antigen IgG posttreatment): average fold increase posttreatment, P value from moderated paired t test (limma), estimated FDR (% FDR, Benjamini and Hochberg procedure), and rank of antigens by average fold increase in corresponding serum IgG level.

At the time of the ProtoArray analysis, serum samples from pretreatment and all three posttreatment time points were available from 47 patients in the sipuleucel-T arm and 13 in the control arm in IMPACT. ProtoArray analyses were performed on samples from all 13 patients from the control arm, and a randomly chosen subset of 28 patients from the sipuleucel-T arm (to ensure successful analysis of roughly 25 samples from each time point). After eliminating failed assays and array quality control, ProtoArray data were available for analysis from the following patients: (i) pretreatment and 2 weeks, 25 sipuleucel-T and 12 control, (ii) pretreatment and 10 weeks, 24 sipuleucel-T and 11 control, (iii) pretreatment and 22 weeks, 24 sipuleucel-T and 13 control.

Assessment of serum IgG responses posttreatment using Luminex xMAP

IgGs levels against candidate antigens were evaluated by Life Technologies Corporation using Luminex xMAP, which uses multiplexed antigen-conjugated, spectrally distinguishable, fluorescent (31). All available pre- and posttreatment serum sample pairs from IMPACT and ProACT patients were evaluated. Samples (30 μL) were assessed at a 1:200 dilution (see Supplementary Methods). Normalized signal intensities were log2-transformed before analysis.

Statistical analyses

Statistical analyses were performed in the “R” computing environment (http://cran.us.r-project.org/). Statistical tests were two-sided, unless stated otherwise. Changes in serum IgG levels are reported relative to pretreatment (e.g., “fold change in IgG level at week 10” refers to the ratio of IgG level at week 10 to that at pretreatment).

Comparing pre- and posttreatment serum IgG levels from ProtoArray.

To assess the statistical significance of pre- to posttreatment changes in levels of IgGs, normalized signal intensities (log2) from ProtoArray assays were tested using a moderated paired t test (limma, R/Bioconductor; ref. 32). The Benjamini and Hochberg procedure was used to perform multiple testing adjustment of P values and obtain estimated false discovery rates (FDR; percentage false discoveries estimated at a certain P value; ref. 33).

Comparing pre- and posttreatment serum IgG levels from Luminex xMAP.

To evaluate increases in IgG levels after treatment, pre- and posttreatment signal intensities (log2) from Luminex xMAP were compared using a one-sided paired Wilcoxon signed-rank test. To evaluate whether the fold changes in IgG levels after treatment were higher in the sipuleucel-T group than in the control group, the values from the two groups were compared using a one-sided Wilcoxon rank-sum test. “IgG response” to an antigen was defined as ≥2-fold increase in signal intensity posttreatment compared with pretreatment (ad-hoc threshold).

Evaluating association of posttreatment changes in serum IgG levels (or IgG responses) with OS in IMPACT.

The association between posttreatment changes in IgG levels (or IgG responses) and OS was evaluated in the sipuleucel-T–treated patients from IMPACT using a two-sided Wald test on the basis of Cox models. Associations were evaluated with: (i) fold change (log2) in serum IgG level posttreatment, and (ii) IgG response status (yes/no), using univariate as well as multivariate models adjusted for baseline PSA and lactate dehydrogenase (LDH) levels. Baseline PSA (log10) and LDH (log10) were selected for use in multivariate models following the modeling approach used in other analyses from the IMPACT study (4, 26). Unless stated otherwise, P values from multivariate models are given in the text, with additional information available in Supplementary Results. No adjustments to P values were made for multiplicity of variables or posttreatment time points.

Identification of treatment-induced increases in serum IgGs to secondary antigens using protein microarray

We used an unbiased human protein microarray platform (ProtoArray v5.0, ≈9,000 proteins; refs. 15, 28–30, 34), representing a diversity of signaling pathways, oncogenes, and tumor antigens, to identify treatment-induced IgGs against self-antigens in IMPACT (Fig. 1; ref. 4). Because it was not established when the increases in levels of IgGs to secondary antigens may appear after treatment, we compared the pretreatment profiles to those from all available posttreatment time points in IMPACT, that is, weeks 2, 10, and 22 after completion of treatment (Supplementary Fig. S1; ref. 4). ProtoArray data from serum samples from 25 sipuleucel-T arm and 13 control arm patients were available for analysis (see Methods).

In the sipuleucel-T arm, the numbers of unique antigens on the ProtoArray with corresponding increases in IgG levels (with ≥2-fold average increase across the sample set, and estimated FDR ≤ 10%) were: 56 at week 2, 162 at week 10, and 23 at week 22 (Fig. 1). In the control arm, using the same thresholds, no increases in levels of IgGs against any of the antigens were observed on the ProtoArray at any of these time points. In the sipuleucel-T–treated patients, IgG against PAP (primary antigen) was highly induced at all the posttreatment time points (Table 1 and Supplementary Table S3). Because the number of antigens against which IgGs were elevated was highest at week 10, we focus the discussions below on further evaluation of the IgGs from this time point; IgGs from weeks 2 and 22 were also evaluated but are discussed only briefly (with supporting information in Supplementary Results).

We next examined whether the sipuleucel-T–induced IgGs were targeted against antigens that are overexpressed in prostate tumors. Target antigens of the 100 IgGs most highly induced at week 10 are encoded by genes that overlapped significantly (P = 0.01) with genes found to be overexpressed in prostate tumors, as identified by the largest study of gene expression in prostate tumor and normal tissues (see Supplementary Results for details of the analysis; ref. 35). This suggested that after treatment with sipuleucel-T, IgGs were induced against antigens that are overexpressed in prostate tumors.

Confirmation of serum IgG responses to secondary antigens with Luminex xMAP

Discoveries from high-content platforms (such as microarrays) can often contain false positives. Therefore, IgG responses to a subset of the secondary antigens identified using ProtoArray were confirmed using an independent analytic platform, Luminex xMAP (31). We chose Luminex xMAP because of its low sample volume requirements, high reported sensitivity, wide linear range, and capability for multiplex IgG detection. From the 162 secondary antigens to which IgG responses were observed with ProtoArray at week 10, we selected 10 for confirmation (Table 1). Of these 10 antigens, five corresponded to the IgGs that exhibited the highest-fold increases in level posttreatment (LGALS3, ANPEP, ECE1, FBXO6, and CACNG1); LGALS3 (36–38), ANPEP (39–41), ECE1 (42–45) are known to be expressed at high levels in prostate tumors or to have functional roles in prostate cancer development. We selected the remaining five antigens based on reported functional relevance in cancer and/or increased expression in prostate tumors, viz., KLK2 (46–48), E-Ras (49), K-Ras (35), TSPAN13 (50), and LGALS8 (36, 51). Descriptions of the 10 candidate antigens are provided in Supplementary Results. ProtoArray data showed that levels of IgGs to most of these 10 candidate secondary antigens were also elevated at the earlier (week 2) and later (week 22) time points (Supplementary Table S3).

Some well-known prostate tumor antigens were not present on the ProtoArray; consequently, no data were available from that platform regarding IgG responses to these antigens. We included two such antigens in the evaluations with Luminex xMAP: PSA and prostate-specific membrane antigen (PSMA [FOLH1]). Three additional antigens were evaluated for control purposes: tetanus toxoid (against which a majority of Americans are immunized; ref. 52), and PA2024 and PAP, the primary antigens in sipuleucel-T.

To consider IgG response to a candidate antigen as confirmed by Luminex xMAP in IMPACT, we required the following criteria to be satisfied (Table 2):

Table 2.

Confirmation of serum IgG responses to candidate secondary antigens at week 10 in IMPACT, using Luminex xMAP

Antigens testedSipuleucel-T (n = 93)Control (n = 39)Sipuleucel-T vs. Control
Selection sourceName or symbolP (pre vs. post)Patients with ≥2-fold upreg. (%)Patients with ≥5-fold upreg. (%)P (pre vs. post)Patients with ≥2-fold upreg. (%)Patients with ≥5-fold upreg. (%)P (fold change)
Controls PAP 8.46E−16 69 (74.2) 55 (59.1) 0.107 4 (10.3) 0 (0) 2.38E−12 
 PA2024 2.83E−17 86 (92.5) 75 (80.6) 0.256 10 (25.6) 2 (5.1) 4.81E−17 
 Tetanus Toxoid 5.71E−05 10 (10.8) 0 (0) 0.100 6 (15.4) 1 (2.6) 1.93E−01 
Known PCa antigens PSA 1.42E−10 36 (38.7) 13 (14) 0.066 5 (12.8) 1 (2.6) 2.15E−04 
 PSMA 1.48E−05 18 (19.4) 4 (4.3) 0.293 4 (10.3) 2 (5.1) 2.48E−02 
ProtoArray candidates LGALS3 2.83E−10 26 (28) 3 (3.2) 0.152 4 (10.3) 2 (5.1) 4.72E−04 
 CACNG1 4.23E−04 9 (9.7) 3 (3.2) 0.425 2 (5.1) 0 (0) 3.07E−02 
 ANPEP 8.14E−06 6 (6.5) 1 (1.1) 0.753 1 (2.6) 0 (0) 7.06E−04 
 FBXO6 4.69E−03 12 (12.9) 4 (4.3) 0.846 1 (2.6) 0 (0) 1.36E−02 
 ECE1 2.22E−04 25 (26.9) 6 (6.5) 0.302 4 (10.3) 3 (7.7) 5.22E−02 
 ERAS 2.97E−10 39 (41.9) 11 (11.8) 0.148 5 (12.8) 2 (5.1) 1.92E−04 
 TSPAN13 1.41E−02 11 (11.8) 4 (4.3) 0.779 2 (5.1) 0 (0) 4.74E−02 
 LGALS8 3.57E−11 23 (24.7) 5 (5.4) 0.034 3 (7.7) 0 (0) 3.06E−04 
 KRAS 1.82E−10 37 (39.8) 14 (15.1) 0.208 5 (12.8) 0 (0) 4.71E−05 
 KLK2 1.73E−09 41 (44.1) 9 (9.7) 0.079 5 (12.8) 1 (2.6) 3.94E−04 
Antigens testedSipuleucel-T (n = 93)Control (n = 39)Sipuleucel-T vs. Control
Selection sourceName or symbolP (pre vs. post)Patients with ≥2-fold upreg. (%)Patients with ≥5-fold upreg. (%)P (pre vs. post)Patients with ≥2-fold upreg. (%)Patients with ≥5-fold upreg. (%)P (fold change)
Controls PAP 8.46E−16 69 (74.2) 55 (59.1) 0.107 4 (10.3) 0 (0) 2.38E−12 
 PA2024 2.83E−17 86 (92.5) 75 (80.6) 0.256 10 (25.6) 2 (5.1) 4.81E−17 
 Tetanus Toxoid 5.71E−05 10 (10.8) 0 (0) 0.100 6 (15.4) 1 (2.6) 1.93E−01 
Known PCa antigens PSA 1.42E−10 36 (38.7) 13 (14) 0.066 5 (12.8) 1 (2.6) 2.15E−04 
 PSMA 1.48E−05 18 (19.4) 4 (4.3) 0.293 4 (10.3) 2 (5.1) 2.48E−02 
ProtoArray candidates LGALS3 2.83E−10 26 (28) 3 (3.2) 0.152 4 (10.3) 2 (5.1) 4.72E−04 
 CACNG1 4.23E−04 9 (9.7) 3 (3.2) 0.425 2 (5.1) 0 (0) 3.07E−02 
 ANPEP 8.14E−06 6 (6.5) 1 (1.1) 0.753 1 (2.6) 0 (0) 7.06E−04 
 FBXO6 4.69E−03 12 (12.9) 4 (4.3) 0.846 1 (2.6) 0 (0) 1.36E−02 
 ECE1 2.22E−04 25 (26.9) 6 (6.5) 0.302 4 (10.3) 3 (7.7) 5.22E−02 
 ERAS 2.97E−10 39 (41.9) 11 (11.8) 0.148 5 (12.8) 2 (5.1) 1.92E−04 
 TSPAN13 1.41E−02 11 (11.8) 4 (4.3) 0.779 2 (5.1) 0 (0) 4.74E−02 
 LGALS8 3.57E−11 23 (24.7) 5 (5.4) 0.034 3 (7.7) 0 (0) 3.06E−04 
 KRAS 1.82E−10 37 (39.8) 14 (15.1) 0.208 5 (12.8) 0 (0) 4.71E−05 
 KLK2 1.73E−09 41 (44.1) 9 (9.7) 0.079 5 (12.8) 1 (2.6) 3.94E−04 

NOTE: Columns indicate the antigens evaluated, antigen selection source (i.e., control antigens, known PCa antigens, and antigens identified from ProtoArray), P value for serum IgG level increase at week 10 vs. pretreatment, number (%) of patients in the sipuleucel-T and control arms with ≥2- or ≥5-fold increase in serum IgG level after treatment; the rightmost column lists P values for fold change in serum IgG levels with sipuleucel-T vs. control using a one-sided Wilcoxon rank-sum test (HA: fold change in serum IgG level posttreatment is higher in the sipuleucel-T arm). IgG responses against antigens that were confirmed by Luminex xMAP are in bold (PAP, PA2024, PSA, LGALS3, E-RAS, LGALS8, K-RAS and KLK2). See Supplementary Table S4 for information on the protein reagents used for Luminex assays.

Abbreviation: PCa, prostate cancer.

  • In a comparison of patients from sipuleucel-T and control arms, higher fold change in IgG level posttreatment with sipuleucel-T (n = 93) than with control (n = 39), with P ≤ 0.01 for the comparison.

  • In sipuleucel-T–treated patients (n = 93), a significant increase in IgG levels after treatment compared with pretreatment (P ≤ 0.01) and ≥10% of patients, demonstrating an IgG response (defined as ≥2-fold increase in IgG level posttreatment compared with pretreatment).

Sipuleucel-T–induced increases in levels of IgGs to the following antigens were confirmed by Luminex xMAP: PSA, KLK2, K-Ras, E-Ras, LGALS8, and LGALS3 (Table 2 and Fig. 2A). As expected, increases in levels of IgGs to PAP and PA2024 were also confirmed. IgG responses to several antigens, for example, PSA, KLK2, K-Ras, E-Ras, and LGALS3, were observed in ≥25% (range, 28%–44%) of the sipuleucel-T–treated patients from IMPACT. A 5-fold increase in anti-PSA, anti-K-Ras, or anti-E-Ras IgG level was observed in ≥10% (range, 12%–15%) of sipuleucel-T–treated patients. Levels of IgGs to these antigens were not significantly elevated after treatment in the control patients (P > 0.01; Table 2).

Figure 2.

Confirmation of serum IgG responses to secondary antigens in patients from IMPACT using Luminex xMAP, and analyses of overlaps between IgG responders. A, confirmation of IgG responses in IMPACT, 10 weeks after treatment, using Luminex xMAP. Log2 of serum IgG levels (y-axes) pre- and posttreatment in the control and sipuleucel-T arms are shown. Data for three secondary antigens (PSA, K-Ras, and LGALS3) are shown. Serum levels of IgG against PAP (primary antigen) are shown for reference. See Table 2 for details. B, overlap of IgG responses to different antigens in patients from the sipuleucel-T arm shown using Venn diagrams. The numbers of patients with overlapping IgG responses to different antigens at week 10 are shown. Left, overlaps include IgG responses to PAP (primary antigen). Right, overlaps include IgG responses to secondary antigens only. IgG responders are defined as patients with ≥2-fold increase in serum IgG level posttreatment versus pretreatment. Information on the significance of the overlaps between IgG responders to all antigens is given in Supplementary Table S5; representative examples are shown here.

Figure 2.

Confirmation of serum IgG responses to secondary antigens in patients from IMPACT using Luminex xMAP, and analyses of overlaps between IgG responders. A, confirmation of IgG responses in IMPACT, 10 weeks after treatment, using Luminex xMAP. Log2 of serum IgG levels (y-axes) pre- and posttreatment in the control and sipuleucel-T arms are shown. Data for three secondary antigens (PSA, K-Ras, and LGALS3) are shown. Serum levels of IgG against PAP (primary antigen) are shown for reference. See Table 2 for details. B, overlap of IgG responses to different antigens in patients from the sipuleucel-T arm shown using Venn diagrams. The numbers of patients with overlapping IgG responses to different antigens at week 10 are shown. Left, overlaps include IgG responses to PAP (primary antigen). Right, overlaps include IgG responses to secondary antigens only. IgG responders are defined as patients with ≥2-fold increase in serum IgG level posttreatment versus pretreatment. Information on the significance of the overlaps between IgG responders to all antigens is given in Supplementary Table S5; representative examples are shown here.

Close modal

In the evaluable patients treated with sipuleucel-T in IMPACT, changes in levels of IgGs at week 10 against the confirmed secondary antigens were not associated (P > 0.01) with the prognostic baseline characteristics reported by Halabi and colleagues (baseline serum PSA, LDH, albumin, or hemoglobin, total Gleason score, metastatic disease site or Eastern Cooperative Oncology Group [ECOG] performance status; refs. 53, 54), or prior therapeutic interventions (chemotherapy, radiotherapy, radical prostatectomy, or bisphosphonates; data not shown).

Overlap among serum IgG responders in the sipuleucel-T arm from IMPACT

To determine whether sipuleucel-T–treated patients shared IgG responses to the same secondary antigens, we evaluated overlaps among IgG responders (see Venn diagrams in Fig. 2B for representative examples, and Supplementary Table S5 for details). The majority of the sipuleucel-T–treated patients who had IgG responses to secondary antigens overlapped with those who had IgG responses to PAP (Fig. 2B, left). Although there were differences in the patterns of IgG responses among the patients, significant overlaps were also observed among IgG responders to a number of different secondary antigens (e.g., E-Ras and KLK2, or LGALS3 and K-Ras; P ≤ 0.01, hypergeometric test). For example, when IgG responses to PSA, E-Ras, LGALS8, and LGALS3 were considered, 25% (23 of 93) of sipuleucel-T–treated patients exhibited responses to three or more of the same antigens, and 9% exhibited responses to all four of these antigens (Fig. 2B, right). In this case, depending on the antigen, a minority (≤30%) of the IgG responses were unique (i.e., did not overlap with IgG responses to other antigens).

Validation of serum IgG responses in ProACT using Luminex xMAP

To determine whether the sipuleucel-T–induced IgG responses observed in IMPACT also occurred in patients from an independent clinical study of sipuleucel-T, we assessed IgG responses to PSA, LGALS3, E-Ras, LGALS8, K-Ras, and KLK2 in a phase II study, ProACT. Serum samples were available from ProACT patients (n = 26) for the week 12 time point, which was comparable with the week 10 time point in IMPACT. In these patients, levels of IgGs to PSA, KLK2, K-Ras, E-Ras, and LGALS3 were significantly elevated versus pretreatment (P ≤ 0.01), with 2-fold increase in ≥10% (range, 15%–31%) of patients (Supplementary Table S6). These results provided evidence of sipuleucel-T–induced IgG responses to multiple secondary antigens in an independent cohort.

Association of sipuleucel-T–induced changes in serum IgG levels with OS in IMPACT

The association of sipuleucel-T–induced changes in IgG levels at week 10 with OS was evaluated in IMPACT using a multivariate Cox model, adjusted for baseline PSA and LDH levels. Sipuleucel-T–induced changes in levels of anti-PSA IgG [hazard ratio (HR), 0.63; 95% confidence interval (CI), 0.46–0.86; P < 0.01] and anti-LGALS3 IgG (HR, 0.60; 95% CI, 0.38–0.96; P = 0.04) were significantly associated with improved OS (Fig. 3 and Supplementary Table S7A). Sipuleucel-T–induced changes in levels of anti-KLK2 IgG and anti-E-Ras IgG showed a trend toward association with improved OS (0.05 < P < 0.1).

Figure 3.

Association of posttreatment changes in serum levels of IgG (log2) against specified secondary antigens with OS in the sipuleucel-T arm in IMPACT. The Kaplan–Meier plots for serum level of IgG to PSA (left) or LGALS3 (right) versus OS are shown with patients grouped by median fold change in serum IgG level at week 10 versus pretreatment (≥median, <median); the dotted horizontal line indicates estimated median OS. The total number of patients in the analyses is given at the top right of each plot; the numbers of patients in the ≥median and <median groups are also shown within each of the figures. The forest plot below the Kaplan–Meier plots shows HR and 95% CI for the associations of change in serum IgG levels (log2 scale) with OS (adjusted for baseline PSA and LDH) from a multivariate Cox model (Supplementary Table S7A for details); blue boxes indicate HR and whiskers indicate 95% CI. HRs and P values for each of the IgGs are to the left of the forest plot. P values are shown to 3 significant digits. **, P ≤ 0.01; *, P ≤ 0.05; •, P ≤ 0.1.

Figure 3.

Association of posttreatment changes in serum levels of IgG (log2) against specified secondary antigens with OS in the sipuleucel-T arm in IMPACT. The Kaplan–Meier plots for serum level of IgG to PSA (left) or LGALS3 (right) versus OS are shown with patients grouped by median fold change in serum IgG level at week 10 versus pretreatment (≥median, <median); the dotted horizontal line indicates estimated median OS. The total number of patients in the analyses is given at the top right of each plot; the numbers of patients in the ≥median and <median groups are also shown within each of the figures. The forest plot below the Kaplan–Meier plots shows HR and 95% CI for the associations of change in serum IgG levels (log2 scale) with OS (adjusted for baseline PSA and LDH) from a multivariate Cox model (Supplementary Table S7A for details); blue boxes indicate HR and whiskers indicate 95% CI. HRs and P values for each of the IgGs are to the left of the forest plot. P values are shown to 3 significant digits. **, P ≤ 0.01; *, P ≤ 0.05; •, P ≤ 0.1.

Close modal

In the sipuleucel-T arm, IgG responses (≥2-fold increase at week 10) to PSA (responders, n = 36; nonresponders, n = 57; HR, 0.38; 95% CI, 0.19–0.80; P = 0.01) and LGALS3 (responders, n = 26; nonresponders, n = 67; HR, 0.25; 95% CI, 0.09–0.72; P = 0.01) were also significantly associated with improved OS (Supplementary Table S7B). Relative to patients in the control arm, patients in the sipuleucel-T arm who were anti-PSA IgG responders exhibited significantly improved OS (HR, 0.27; 95% CI, 0.12–0.58; P < 0.01; Fig. 4 and Supplementary Table S8), whereas OS in patients in the sipuleucel-T arm who were anti-PSA IgG nonresponders did not differ significantly from that in patients in the control arm (HR, 0.71; 95% CI, 0.41–1.23; P = 0.23). Similarly, patients in the sipuleucel-T arm who were anti-LGALS3 IgG responders exhibited significantly improved OS compared with those in the control arm (HR, 0.16; 95% CI, 0.06–0.49; P < 0.01; Fig. 4), whereas OS in patients in the sipuleucel-T arm who were anti-LGALS3 IgG nonresponders did not differ significantly from patients in the control arm (HR, 0.66; 95% CI, 0.38–1.12; P = 0.12). These results indicate that a relatively moderate increase in levels of IgGs to secondary antigens (≥2-fold) may identify patients who are more likely to have significant survival benefit after sipuleucel-T treatment than patients who do not show increase in these IgGs.

Figure 4.

The Kaplan–Meier plots of OS for anti-PSA and anti-LGALS3 IgG responders (R) and nonresponders (NR) in the sipuleucel-T arm and patients from the control arm (C) in IMPACT. These figures show the difference in OS between the control group and the IgG responder or nonresponder groups in the sipuleucel-T arm. Data for IgG responses at week 2 (left), week 10 (middle), and week 22 (right) are shown. Top, anti-PSA IgG; bottom, anti-LGALS3 IgG. IgG responders are defined as patients with ≥2-fold increase in serum IgG level posttreatment relative to pretreatment. The total number of patients in the analyses is given at the top right of each plot; the numbers of patients in the R, NR, and C groups are also shown within each of the figures. The P values and HRs (whiskers indicate 95% CI) in the forest plot inlays indicate the significance of the difference in OS in the R and NR groups relative to the C group using a multivariate Cox model (adjusted for baseline PSA and LDH levels) as described in the methods. Refer to Supplementary Table S8 (week 10) and S13 (weeks 2 and 22) for detailed information.

Figure 4.

The Kaplan–Meier plots of OS for anti-PSA and anti-LGALS3 IgG responders (R) and nonresponders (NR) in the sipuleucel-T arm and patients from the control arm (C) in IMPACT. These figures show the difference in OS between the control group and the IgG responder or nonresponder groups in the sipuleucel-T arm. Data for IgG responses at week 2 (left), week 10 (middle), and week 22 (right) are shown. Top, anti-PSA IgG; bottom, anti-LGALS3 IgG. IgG responders are defined as patients with ≥2-fold increase in serum IgG level posttreatment relative to pretreatment. The total number of patients in the analyses is given at the top right of each plot; the numbers of patients in the R, NR, and C groups are also shown within each of the figures. The P values and HRs (whiskers indicate 95% CI) in the forest plot inlays indicate the significance of the difference in OS in the R and NR groups relative to the C group using a multivariate Cox model (adjusted for baseline PSA and LDH levels) as described in the methods. Refer to Supplementary Table S8 (week 10) and S13 (weeks 2 and 22) for detailed information.

Close modal

We next tested whether increasing numbers of IgG responses to secondary antigens was associated with progressively improved OS. Interestingly, as compared with patients with no IgG responses against the confirmed antigens (namely, PSA, LGALS3, E-Ras, LGALS8, K-Ras, and KLK2; n = 26), patients with ≥1 (n = 67; HR, 0.57; 95% CI, 0.29–1.10; P = 0.09), ≥2 (n = 52; HR, 0.51; 95% CI, 0.25–1.03; P = 0.06), ≥3 (n = 34; HR, 0.39; 95% CI, 0.17–0.92; P = 0.03), and ≥4 (n = 26; HR, 0.30; 95% CI, 0.11–0.82; P = 0.02) IgG responses showed progressively improved OS (Supplementary Fig. S2). The number of patients with IgG responses ≥5 was small (n ≤ 17), therefore, patients with ≥5 IgG responses were not analyzed. Overall, these data suggest that the total number of IgG responses (in other words, the extent of antigen spread) may be associated with improved OS.

Among patients from the sipuleucel-T arm in IMPACT, anti-PA2024 and anti-PAP IgG responders exhibited improved OS relative to control patients, whereas anti-PA2024 and anti-PAP IgG nonresponders did not (Supplementary Table S8). However, the increases in OS observed in the anti-PA2024 and anti-PAP IgG responders were not as significant as those in the anti-PSA or anti-LGALS3 IgG responders. A previous analysis of data from IMPACT, using different methodologies, showed that aggregated humoral (IgG and IgM combined) and cellular immune responses to PAP or PA2024 were associated with improved OS (26).

Serum IgG responses at earlier and later posttreatment time points

To determine whether the IgG responses to secondary antigens observed at the week 10 time point in IMPACT also occurred at other time points, we examined serum samples from the other available posttreatment time points in IMPACT (weeks 2 and 22) and ProACT (weeks 4 and 20; see Supplementary Table S9 for IMPACT and Supplementary Table S10 for ProACT results). In IMPACT, significant (P ≤ 0.01) increases in levels of IgGs against PSA, KLK2, K-Ras, E-Ras, LGALS8, and LGALS3 were observed in the sipuleucel-T group at week 2 (n = 142) and week 22 (n = 60). In ProACT, significant increases in levels of IgGs against PSA, KLK2, K-Ras, E-Ras, LGALS8, and LGALS3 were observed at week 4 (n = 33), but not at week 20 with sipuleucel-T (n = 19). The relatively small number of patients evaluable at week 20 in ProACT may have contributed to the lack of statistical significance.

Sipuleucel-T–induced IgG responses were largely consistent across the posttreatment time points. Patients from the sipuleucel-T arm in IMPACT with IgG responses to an antigen at week 10 frequently exhibited an IgG response to the same antigen at the earlier (week 2) and later (week 22) time points (P ≤ 0.01, hypergeometric test; Supplementary Table S11).

In the sipuleucel-T–treated patients from IMPACT, IgG responses to PSA at week 2 were significantly associated with improved OS (responders, n = 35; nonresponders, n = 107; HR, 0.42; 95% CI, 0.22–0.79; P < 0.01; Fig. 4; Supplementary Table S12B), and IgG responses to LGALS3 at week 2 showed a trend toward improved OS (responders, n = 41; nonresponders, n = 101; HR, 0.57; 95% CI, 0.31–1.02; P = 0.06). Patients in the sipuleucel-T arm who were anti-PSA or anti-LGALS3 IgG responders at week 2 (but not the corresponding nonresponders) exhibited significantly improved OS (P < 0.01) relative to the patients in the control arm (Fig. 4 and Supplementary Table S13). Patients in the sipuleucel-T arm with IgG responses to the primary antigen (PAP or PA2024) at week 2 also exhibited improved OS (P < 0.05; Supplementary Table S13) compared with patients in the control arm, whereas OS of patients without the IgG responses were not significantly different from patients in the control arm. IgG responses at week 22 were not significantly associated with OS; however, in the sipuleucel-T arm, patients with anti-PSA IgG responses showed a trend toward improved OS relative to nonresponders (responders, n = 18; nonresponders, n = 42; HR, 0.35; 95% CI, 0.12–1.05; P = 0.06; Supplementary Table S12B).

Thus, increases in levels of serum IgGs to multiple secondary antigens were observed as early as 2 weeks and up to 22 weeks (6 months) after completion of treatment with sipuleucel-T in patients from IMPACT. In the sipuleucel-T arm from IMPACT, IgG responses at multiple time points after treatment, spanning several months, were associated with improved OS.

The results presented here show that humoral antigen spread occurred after treatment with sipuleucel-T. The IgG responses were not significantly associated with baseline prognostic characteristics or prior therapeutic interventions. Sipuleucel-T–induced IgG responses to PSA and LGALS3 were associated with improved OS in IMPACT.

Although antigen spread after treatment with cancer immunotherapies and its potential for association with clinical outcomes has been known for some time (13, 25), typically, such investigations have included relatively small cohorts or measured responses to a predefined set of antigens (11–15, 55, 56). To our knowledge, this is the most extensive demonstration yet of antigen spread and its relationship with clinical benefit in the context of a cancer immunotherapy, and the first such evaluation in a clinical study with a comparator control arm. The study provides insight into the temporal dynamic of the antigen spread induced by an immunotherapy using a broad platform (protein microarray) and demonstrates treatment-induced IgG responses to antigens against which immune responses had not been previously reported in prostate cancer (K-Ras and E-Ras). Finally, these data provide evidence in a relatively large population that IgG responses to secondary antigens can be associated with improved OS, and that such responses can be observed early and persist for several months after treatment.

A few other studies of cancer immunotherapies (15, 17, 30, 57) have reported unbiased or broad evaluations of humoral responses to self-antigens; some of these are of particular interest in light of the results presented here. Kwek and colleagues (15) reported a higher frequency (or breadth) of IgG responses against self-antigens in a small cohort of patients with malignant melanoma who responded clinically to ipilimumab and GM-CSF; the majority of these responses were patient-specific. In contrast, many patients shared responses to the same antigens (see Fig. 2B, right, for example) in our study. This may reflect differences in the modes of action of sipuleucel-T and ipilimumab/GM-CSF, or more generally, between antigen-directed [directed toward primary tumor antigen(s)] and nondirected (not directed toward specific tumor antigen[s]) immunotherapies. Compared with a nondirected immunotherapy (such as ipilimumab/GM-CSF), antigen spread after an antigen-directed immunotherapy (such as sipuleucel-T) may be limited to the antigens overexpressed or altered in tumor cells that express the primary antigen(s), and consequently, such responses may be shared among patients more often.

Using protein microarrays, Nguyen and colleagues (30) investigated antibody responses after treatment with GVAX, a vaccine for prostate cancer comprised two allogeneic prostate carcinoma cell lines, modified to secrete GM-CSF. Of the ≈30 candidate antigens against which antibody responses were observed after treatment (30), two (LGALS3 and LGALS8) were confirmed in our study; one (LGALS8) being a known prostate tumor antigen (51). Further studies may help determine whether immune responses to the galectins, which are highly expressed in prostate tumors (36), are common after treatment of prostate cancer with immunotherapies.

Another study used serologic analysis of expression cDNA libraries (SEREX) to investigate antigen spread in prostate cancer patients who received a viral vaccine encoding PSA (57). Of the vaccine-treated patients, 21% (7 of 30) developed antibody responses to a common set of ubiquitously expressed self-antigens, compared with none in the control groups. We did not observe responses to those antigens in our study. Patients with autoantibody responses after the vaccine treatment showed a trend toward decreased biochemical-free survival. The authors suggested that the nature of the secondary antigens may inform clinical outcome (57); immune responses to ubiquitously expressed self-antigens may reflect off-target activity and be associated with adverse effects, whereas responses to tumor antigens may reflect antitumor immune activity and therefore be associated with improved clinical outcomes (12–16). In this study, all of the confirmed IgG responses were associated with (or showed a trend toward) improved OS (Fig. 3).

The identification of posttreatment markers of an antitumor immune response is an important challenge in characterizing in vivo responses to cancer immunotherapies (7, 10). Measuring immune responses to the primary targets of antigen-directed immunotherapies (12, 26, 57) is a common practice. However, questions remain regarding the appropriate immunologic measures that provide the best information about clinical efficacy of immunotherapies (7, 11). Ideally, immunologic biomarkers should reflect the therapeutic mechanism of action and be quantifiable from easily accessed biospecimens (11). Humoral responses associated with antigen spread appear to have these properties and may hold potential utility as pharmacodynamic biomarkers for the assessment of clinical benefits of some cancer immunotherapies.

The associations of IgG responses with OS after sipuleucel-T treatment should be considered as hypotheses-generating observations. It will be important to establish the optimal methods for quantifying the IgG responses (e.g., with titrations of serum or antigen-conjugated Luminex beads) and to validate the associations with OS in additional clinical studies. The current results, along with previously published findings on sipuleucel-T (4, 26, 58), also point to opportunities to further characterize the observed IgG responses and to understand their origin and function. Some of these are discussed below.

Although our data do not reveal the mechanism of origin of the IgG responses (note that tumor specimens were not available for analyses in IMPACT or ProACT), a few observations argue in favor of a possible treatment-induced antitumor effect. First, in a neoadjuvant setting, sipuleucel-T increases the frequency of CD4+ and CD8+ T cells within prostate tissue 2 to 3 weeks after treatment, particularly at the tumor margin (58). Second, sipuleucel-T–induced antigen spread was a delayed manifestation; a higher number of IgG responses were observed on ProtoArray at week 10 than at week 2. Third, IgG responses were mounted against antigens that are overexpressed in prostate tumors (35), including intracellular oncogenes (e.g., Ras members). Taken together, these observations suggest a mechanism by which the initial immune responses mobilized by sipuleucel-T are responsible for some degree of tumor cell lysis and release of secondary antigens; these secondary antigens may have triggered the IgG responses observed here.

The induction of IgG responses suggests that CD4+ helper T cells (Th2) have been engaged. Both cellular and humoral responses to the primary antigen, PAP, have been reported (4, 26), and characterization of cytokine profiles in sipuleucel-T product indicated engagement of both Th1 and Th2 mechanisms (26). It will be useful to determine whether sipuleucel-T–treated patients exhibit cellular immune responses to the secondary antigens. Combined cellular and humoral responses may be more effective in identifying the patients with antigen spread after-treatment. It is of note here that in another study, after treatment with PSA-TRICOM, a PSA-targeted viral vaccine for prostate cancer (59), primarily T-cell responses to PSA were generated, with only rare instances of anti-PSA antibody responses. In contrast, we found that anti-PSA IgGs were frequently induced after treatment with sipuleucel-T.

To date, the development of active cancer immunotherapies has focused on induction of CD8+ cytotoxic T-cell responses; however, recent studies have shown the importance of CD4+ T cells, B cells and humoral immunity in mediating tumor control (60–62). In particular, infiltration of plasma B cells and immunoglobulin gene expression have been identified as important predictors of good prognosis and response to chemotherapy (61, 62). The ability of IgG antibodies to mediate tumor cell killing suggests that tumor-specific antibodies may play a role in the overall ability of the immune system to combat tumors. Although the magnitudes of sipuleucel-T–induced IgG responses to secondary antigens were modest compared with that observed to the primary antigen, multiple IgG responses could collectively mediate an additive or synergistic antitumor effect. Several of the confirmed IgG responses were against antigens known to be involved with PSA and Ras-related molecular functions: PSA, KLK2, LGALS3, and K-Ras. KLK2 generates the enzymatically active form of PSA from pro-PSA (48). LGALS3 is cleaved by PSA, potentially modulating its function (38), and is a binding partner and activator in K-Ras signaling (63, 64). Whether there is a functional significance to the appearance of IgG responses to multiple antigens in this molecular pathway is unknown. It is also unclear why a subset of the confirmed IgG responses was more significantly associated with OS, although all showed a trend toward improved OS. The association of a given IgG response with OS may be determined by particularities of the expression or biologic role of the given antigen, or the functionality of the IgGs generated. Investigation of the isotypes of these IgGs and their ability to mediate antibody-dependent cell-mediated cytotoxicity (ADCC) and other functions (binding or neutralization), individually and in combination, may help shed light on their relevance to antitumor responses. In this study, we have confirmed only a relatively limited number of IgG responses; it will be of interest in future studies to assess a wider set of IgGs than those characterized here. Specifically, additional candidates from the protein microarrays as well as known prostate cancer-related gene variants, such as the TMPRSS2–ERG fusion product may help in gleaning a more complete understanding of IgG responses and their functional relevance.

Finally, treatment-induced immune responses to secondary antigens may provide targets for developing future therapies (15). Therefore, it will be of interest to ask in preclinical studies if immunization against the particular antigens to which IgG responses were observed here is protective against tumor growth.

D. GuhaThakurta, N.A. Sheikh, T. DeVries, and J.B. Trager have ownership interest (including patents) in Dendreon Corporation. P. Kantoff is a consultant/advisory board member for Dendreon Corporation. C.S. Higano reports receiving a commercial research grant from and is a consultant/advisory board member for Dendreon Corporation. C.G. Drake reports receiving commercial research grants from Aduro Biotech, Bristol-Myers Squibb, and Janssen; holds ownership interest (including patents) in Compugen and NexImmune; and is a consultant/advisory board member for Bristol-Myers Squibb, Compugen, Dendreon Corporation, ImmunExcite, NexImmune, Novartis, and Roche/Genentech. No potential conflicts of interest were disclosed by the other authors.

Conception and design: D. GuhaThakurta, N.A. Sheikh, M.W. Frohlich, J.B. Trager, C.G. Drake

Development of methodology: D. GuhaThakurta, L.-Q. Fan, M.W. Frohlich

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): D. GuhaThakurta, L.-Q. Fan, C.S. Higano, T.A. Gardner, K. Bailey, T. Vu

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): D. GuhaThakurta, N.A. Sheikh, H. Kandadi, T.C. Meagher, S.J. Hall, P.W. Kantoff, C.S. Higano, T.A. Gardner, T. DeVries, J.B. Whitmore, M.W. Frohlich, J.B. Trager, C.G. Drake

Writing, review, and/or revision of the manuscript: D. GuhaThakurta, N.A. Sheikh, H. Kandadi, T.C. Meagher, S.J. Hall, P.W. Kantoff, C.S. Higano, E.J. Small, T.A. Gardner, T. DeVries, J.B. Whitmore, M.W. Frohlich, C.G. Drake

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): N.A. Sheikh, H. Kandadi

Study supervision: D. GuhaThakurta, T.C. Meagher, C.S. Higano, M.W. Frohlich, J.B. Trager

Other (scientific interpretation of data (not necessarily computational)): C.G. Drake

Matt Harmon is thanked for providing the information from clinical trials. The authors thank Susan Veals and Johnathan Maher of Dendreon Corporation for suggestions on the article and medical writing support. Lawrence Fong and Frances Stewart are thanked for helpful discussions. Life Technologies Corporation is thanked for their expertise and consultation on the humoral assays and assay designs. This work would not be possible without the patients who enrolled in sipuleucel-T clinical trials and consented to contribute their biologic specimens for investigational purposes. Physicians and investigators of IMPACT and ProACT are thanked for their contribution in enrolling and treating patients.

Funding for the study was provided by Dendreon Corporation.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1.
Ribas
A
,
Hersey
P
,
Middleton
MR
,
Gogas
H
,
Flaherty
KT
,
Sondak
VK
, et al
New challenges in endpoints for drug development in advanced melanoma
.
Clin Cancer Res
2012
;
18
:
336
41
.
2.
Wolchok
JD
,
Hoos
A
,
O'Day
S
,
Weber
JS
,
Hamid
O
,
Lebbe
C
, et al
Guidelines for the evaluation of immune therapy activity in solid tumors: immune-related response criteria
.
Clin Cancer Res
2009
;
15
:
7412
20
.
3.
Madan
RA
,
Gulley
JL
,
Fojo
T
,
Dahut
WL
. 
Therapeutic cancer vaccines in prostate cancer: the paradox of improved survival without changes in time to progression
.
Oncologist
2010
;
15
:
969
75
.
4.
Kantoff
PW
,
Higano
CS
,
Shore
ND
,
Berger
ER
,
Small
EJ
,
Penson
DF
, et al
Sipuleucel-T immunotherapy for castration-resistant prostate cancer
.
N Engl J Med
2010
;
363
:
411
22
.
5.
Hodi
FS
,
O'Day
SJ
,
McDermott
DF
,
Weber
RW
,
Sosman
JA
,
Haanen
JB
, et al
Improved survival with ipilimumab in patients with metastatic melanoma
.
N Engl J Med
2010
;
363
:
711
23
.
6.
Hoos
A
,
Eggermont
AM
,
Janetzki
S
,
Hodi
FS
,
Ibrahim
R
,
Anderson
A
, et al
Improved endpoints for cancer immunotherapy trials
.
J Natl Cancer Inst
2010
;
102
:
1388
97
.
7.
Hoos
A
. 
Evolution of end points for cancer immunotherapy trials
.
Ann Oncol
2012
;
23
(
Suppl 8
):
viii47
52
.
8.
Stein
WD
,
Gulley
JL
,
Schlom
J
,
Madan
RA
,
Dahut
W
,
Figg
WD
, et al
Tumor regression and growth rates determined in five intramural NCI prostate cancer trials: the growth rate constant as an indicator of therapeutic efficacy
.
Clin Cancer Res
2011
;
17
:
907
17
.
9.
Ribas
A
,
Chmielowski
B
,
Glaspy
JA
. 
Do we need a different set of response assessment criteria for tumor immunotherapy?
Clin Cancer Res
2009
;
15
:
7116
8
.
10.
Fox
BA
,
Schendel
DJ
,
Butterfield
LH
,
Aamdal
S
,
Allison
JP
,
Ascierto
PA
, et al
Defining the critical hurdles in cancer immunotherapy
.
J Transl Med
2011
;
9
:
214
.
11.
Disis
ML
. 
Immunologic biomarkers as correlates of clinical response to cancer immunotherapy
.
Cancer Immunol Immunother
2011
;
60
:
433
42
.
12.
Butterfield
LH
,
Ribas
A
,
Dissette
VB
,
Amarnani
SN
,
Vu
HT
,
Oseguera
D
, et al
Determinant spreading associated with clinical response in dendritic cell-based immunotherapy for malignant melanoma
.
Clin Cancer Res
2003
;
9
:
998
1008
.
13.
Ribas
A
,
Timmerman
JM
,
Butterfield
LH
,
Economou
JS
. 
Determinant spreading and tumor responses after peptide-based cancer immunotherapy
.
Trends Immunol
2003
;
24
:
58
61
.
14.
Yuan
J
,
Adamow
M
,
Ginsberg
BA
,
Rasalan
TS
,
Ritter
E
,
Gallardo
HF
, et al
Integrated NY-ESO-1 antibody and CD8+ T-cell responses correlate with clinical benefit in advanced melanoma patients treated with ipilimumab
.
Proc Natl Acad Sci U S A
2011
;
108
:
16723
8
.
15.
Kwek
SS
,
Dao
V
,
Roy
R
,
Hou
Y
,
Alajajian
D
,
Simko
JP
, et al
Diversity of antigen-specific responses induced in vivo with CTLA-4 blockade in prostate cancer patients
.
J Immunol
2012
;
189
:
3759
66
.
16.
Corbiere
V
,
Chapiro
J
,
Stroobant
V
,
Ma
W
,
Lurquin
C
,
Lethe
B
, et al
Antigen spreading contributes to MAGE vaccination-induced regression of melanoma metastases
.
Cancer Res
2011
;
71
:
1253
62
.
17.
Smith
HA
,
Maricque
BB
,
Eberhardt
J
,
Petersen
B
,
Gulley
JL
,
Schlom
J
, et al
IgG responses to tissue-associated antigens as biomarkers of immunological treatment efficacy
.
J Biomed Biotechnol
2011
;
2011
:
454861
.
18.
McRae
BL
,
Vanderlugt
CL
,
Dal Canto
MC
,
Miller
SD
. 
Functional evidence for epitope spreading in the relapsing pathology of experimental autoimmune encephalomyelitis
.
J Exp Med
1995
;
182
:
75
85
.
19.
Lehmann
PV
,
Forsthuber
T
,
Miller
A
,
Sercarz
EE
. 
Spreading of T-cell autoimmunity to cryptic determinants of an autoantigen
.
Nature
1992
;
358
:
155
7
.
20.
Tisch
R
,
Yang
XD
,
Singer
SM
,
Liblau
RS
,
Fugger
L
,
McDevitt
HO
. 
Immune response to glutamic acid decarboxylase correlates with insulitis in non-obese diabetic mice
.
Nature
1993
;
366
:
72
5
.
21.
Vanderlugt
CL
,
Miller
SD
. 
Epitope spreading in immune-mediated diseases: implications for immunotherapy
.
Nat Rev Immunol
2002
;
2
:
85
95
.
22.
Catterall
WA
. 
Structure and function of voltage-sensitive ion channels
.
Science
1988
;
242
:
50
61
.
23.
Powers
PA
,
Liu
S
,
Hogan
K
,
Gregg
RG
. 
Molecular characterization of the gene encoding the gamma subunit of the human skeletal muscle 1,4-dihydropyridine-sensitive Ca2+ channel (CACNLG), cDNA sequence, gene structure, and chromosomal location
.
J Biol Chem
1993
;
268
:
9275
9
.
24.
Markiewicz
MA
,
Fallarino
F
,
Ashikari
A
,
Gajewski
TF
. 
Epitope spreading upon P815 tumor rejection triggered by vaccination with the single class I MHC-restricted peptide P1A
.
Int Immunol
2001
;
13
:
625
32
.
25.
Gulley
JL
. 
Therapeutic vaccines: the ultimate personalized therapy?
Hum Vaccin Immunother
2013
;
9
:
219
21
.
26.
Sheikh
NA
,
Petrylak
D
,
Kantoff
PW
,
Dela Rosa
C
,
Stewart
FP
,
Kuan
LY
, et al
Sipuleucel-T immune parameters correlate with survival: an analysis of the randomized phase 3 clinical trials in men with castration-resistant prostate cancer
.
Cancer Immunol Immunother
2013
;
62
:
137
47
.
27.
Butterfield
LH
,
Palucka
AK
,
Britten
CM
,
Dhodapkar
MV
,
Hakansson
L
,
Janetzki
S
, et al
Recommendations from the iSBTc-SITC/FDA/NCI Workshop on Immunotherapy Biomarkers
.
Clin Cancer Res
2011
;
17
:
3064
76
.
28.
Schweitzer
B
,
Meng
L
,
Mattoon
D
,
Rai
AJ
. 
Immune response biomarker profiling application on ProtoArray protein microarrays
.
Methods Mol Biol
2010
;
641
:
243
52
.
29.
Gnjatic
S
,
Ritter
E
,
Buchler
MW
,
Giese
NA
,
Brors
B
,
Frei
C
, et al
Seromic profiling of ovarian and pancreatic cancer
.
Proc Natl Acad Sci U S A
2010
;
107
:
5088
93
.
30.
Nguyen
MC
,
Tu
GH
,
Koprivnikar
KE
,
Gonzalez-Edick
M
,
Jooss
KU
,
Harding
TC
. 
Antibody responses to galectin-8, TARP and TRAP1 in prostate cancer patients treated with a GM-CSF-secreting cellular immunotherapy
.
Cancer Immunol Immunother
2010
;
59
:
1313
23
.
31.
Pickering
JW
,
Martins
TB
,
Schroder
MC
,
Hill
HR
. 
Comparison of a multiplex flow cytometric assay with enzyme-linked immunosorbent assay for auantitation of antibodies to tetanus, diphtheria, and Haemophilus influenzae Type b
.
Clin Diagn Lab Immunol
2002
;
9
:
872
6
.
32.
Smyth
GK
. 
Limma: linear models for microarray data
. In:
Gentleman VC
R.
,
Dudoit
S.
,
Irizarry
R.
,
Huber
W.
,
editors
. 
Bioinformatics and computational biology solutions using {R} and bioconductor
.
New York
:
Springer
; 
2005
.
p.
397
420
.
33.
Bejamini
Y
,
Hochberg
Y
. 
Controlling the false discovery rate: a practical and powerful approach to multiple testing
.
J R Soc, Ser B (Methodological)
1995
;
57
:
289
300
.
34.
Hudson
ME
,
Pozdnyakova
I
,
Haines
K
,
Mor
G
,
Snyder
M
. 
Identification of differentially expressed proteins in ovarian cancer using high-density protein microarrays
.
Proc Natl Acad Sci U S A
2007
;
104
:
17494
9
.
35.
Taylor
BS
,
Schultz
N
,
Hieronymus
H
,
Gopalan
A
,
Xiao
Y
,
Carver
BS
, et al
Integrative genomic profiling of human prostate cancer
.
Cancer Cell
2010
;
18
:
11
22
.
36.
Laderach
DJ
,
Gentilini
LD
,
Giribaldi
L
,
Delgado
VC
,
Nugnes
L
,
Croci
DO
, et al
A unique galectin signature in human prostate cancer progression suggests galectin-1 as a key target for treatment of advanced disease
.
Cancer Res
2013
;
73
:
86
96
.
37.
Califice
S
,
Castronovo
V
,
Bracke
M
,
van den Brule
F
. 
Dual activities of galectin-3 in human prostate cancer: tumor suppression of nuclear galectin-3 vs tumor promotion of cytoplasmic galectin-3
.
Oncogene
2004
;
23
:
7527
36
.
38.
Newlaczyl
AU
,
Yu
LG
. 
Galectin-3—a jack-of-all-trades in cancer
.
Cancer Lett
2011
;
313
:
123
8
.
39.
Fukasawa
K
,
Fujii
H
,
Saitoh
Y
,
Koizumi
K
,
Aozuka
Y
,
Sekine
K
, et al
Aminopeptidase N (APN/CD13) is selectively expressed in vascular endothelial cells and plays multiple roles in angiogenesis
.
Cancer Lett
2006
;
243
:
135
43
.
40.
Larkin
SE
,
Holmes
S
,
Cree
IA
,
Walker
T
,
Basketter
V
,
Bickers
B
, et al
Identification of markers of prostate cancer progression using candidate gene expression
.
Br J Cancer
2012
;
106
:
157
65
.
41.
Sorensen
KD
,
Abildgaard
MO
,
Haldrup
C
,
Ulhoi
BP
,
Kristensen
H
,
Strand
S
, et al
Prognostic significance of aberrantly silenced ANPEP expression in prostate cancer
.
Br J Cancer
2013
;
108
:
420
8
.
42.
Lambert
LA
,
Whyteside
AR
,
Turner
AJ
,
Usmani
BA
. 
Isoforms of endothelin-converting enzyme-1 (ECE-1) have opposing effects on prostate cancer cell invasion
.
Br J Cancer
2008
;
99
:
1114
20
.
43.
Herrmann
E
,
Bogemann
M
,
Bierer
S
,
Eltze
E
,
Hertle
L
,
Wulfing
C
. 
The endothelin axis in urologic tumors: mechanisms of tumor biology and therapeutic implications
.
Expert Rev Anticancer Ther
2006
;
6
:
73
81
.
44.
Nelson
JB
,
Hedican
SP
,
George
DJ
,
Reddi
AH
,
Piantadosi
S
,
Eisenberger
MA
, et al
Identification of endothelin-1 in the pathophysiology of metastatic adenocarcinoma of the prostate
.
Nat Med
1995
;
1
:
944
9
.
45.
Nelson
JB
,
Udan
MS
,
Guruli
G
,
Pflug
BR
. 
Endothelin-1 inhibits apoptosis in prostate cancer
.
Neoplasia
2005
;
7
:
631
7
.
46.
Darson
MF
,
Pacelli
A
,
Roche
P
,
Rittenhouse
HG
,
Wolfert
RL
,
Saeid
MS
, et al
Human glandular kallikrein 2 expression in prostate adenocarcinoma and lymph node metastases
.
Urology
1999
;
53
:
939
44
.
47.
Williams
SA
,
Xu
Y
,
De Marzo
AM
,
Isaacs
JT
,
Denmeade
SR
. 
Prostate-specific antigen (PSA) is activated by KLK2 in prostate cancer ex vivo models and in prostate-targeted PSA/KLK2 double transgenic mice
.
Prostate
2010
;
70
:
788
96
.
48.
Rittenhouse
HG
,
Finlay
JA
,
Mikolajczyk
SD
,
Partin
AW
. 
Human Kallikrein 2 (hK2) and prostate-specific antigen (PSA): two closely related, but distinct, kallikreins in the prostate
.
Crit Rev Clin Lab Sci
1998
;
35
:
275
368
.
49.
Kubota
E
,
Kataoka
H
,
Aoyama
M
,
Mizoshita
T
,
Mori
Y
,
Shimura
T
, et al
Role of ES cell-expressed Ras (ERas) in tumorigenicity of gastric cancer
.
Am J Pathol
2010
;
177
:
955
63
.
50.
Arencibia
JM
,
Martin
S
,
Perez-Rodriguez
FJ
,
Bonnin
A
. 
Gene expression profiling reveals overexpression of TSPAN13 in prostate cancer
.
Int J Oncol
2009
;
34
:
457
63
.
51.
Su
ZZ
,
Lin
J
,
Shen
R
,
Fisher
PE
,
Goldstein
NI
,
Fisher
PB
. 
Surface-epitope masking and expression cloning identifies the human prostate carcinoma tumor antigen gene PCTA-1 a member of the galectin gene family
.
Proc Natl Acad Sci U S A
1996
;
93
:
7252
7
.
52.
McQuillan
GM
,
Kruszon-Moran
D
,
Deforest
A
,
Chu
SY
,
Wharton
M
. 
Serologic immunity to diphtheria and tetanus in the United States
.
Ann Intern Med
2002
;
136
:
660
6
.
53.
Halabi
S
,
Small
EJ
,
Kantoff
PW
,
Kattan
MW
,
Kaplan
EB
,
Dawson
NA
, et al
Prognostic model for predicting survival in men with hormone-refractory metastatic prostate cancer
.
J Clin Oncol
2003
;
21
:
1232
7
.
54.
Halabi
S
,
Lin
CY
,
Kelly
WK
,
Fizazi
KS
,
Moul
JW
,
Kaplan
EB
, et al
Updated prognostic model for predicting overall survival in first-line chemotherapy for patients with metastatic castration-resistant prostate cancer
.
J Clin Oncol
2014
;
32
:
671
7
.
55.
Disis
ML
,
Goodell
V
,
Schiffman
K
,
Knutson
KL
. 
Humoral epitope-spreading following immunization with a HER-2/neu peptide based vaccine in cancer patients
.
J Clin Immunol
2004
;
24
:
571
8
.
56.
Gulley
JL
,
Arlen
PM
,
Bastian
A
,
Morin
S
,
Marte
J
,
Beetham
P
, et al
Combining a recombinant cancer vaccine with standard definitive radiotherapy in patients with localized prostate cancer
.
Clin Cancer Res
2005
;
11
:
3353
62
.
57.
Nesslinger
NJ
,
Ng
A
,
Tsang
KY
,
Ferrara
T
,
Schlom
J
,
Gulley
JL
, et al
A viral vaccine encoding prostate-specific antigen induces antigen spreading to a common set of self-proteins in prostate cancer patients
.
Clin Cancer Res
2010
;
16
:
4046
56
.
58.
Fong
L
,
Carroll
P
,
Weinberg
V
,
Chan
S
,
Lewis
J
,
Corman
J
, et al
Activated lymphocyte recruitment into the tumor microenvironment following preoperative sipuleucel-T for localized prostate cancer
.
J Natl Cancer Inst
2014
;
106
:
268
280
.
59.
Gulley
JL
,
Madan
RA
,
Tsang
KY
,
Jochems
C
,
Marte
J
,
Farsaci
B
, et al
Immune impact induced by PROSTVAC (PSA-TRICOM), a therapeutic vaccine for prostate cancer
.
Cancer Immunol Res
2014
;
2
:
133
41
.
60.
Hunder
NN
,
Wallen
H
,
Cao
J
,
Hendricks
DW
,
Reilly
JZ
,
Rodmyre
R
, et al
Treatment of metastatic melanoma with autologous CD4+ T cells against NY-ESO-1
.
N Engl J Med
2008
;
358
:
2698
703
.
61.
Schmidt
M
,
Hellwig
B
,
Hammad
S
,
Othman
A
,
Lohr
M
,
Chen
Z
, et al
A comprehensive analysis of human gene expression profiles identifies stromal immunoglobulin kappa C as a compatible prognostic marker in human solid tumors
.
Clin Cancer Res
2012
;
18
:
2695
703
.
62.
Lohr
M
,
Edlund
K
,
Botling
J
,
Hammad
S
,
Hellwig
B
,
Othman
A
, et al
The prognostic relevance of tumour-infiltrating plasma cells and immunoglobulin kappa C indicates an important role of the humoral immune response in non-small cell lung cancer
.
Cancer Lett
2013
;
333
:
222
8
.
63.
Elad-Sfadia
G
,
Haklai
R
,
Balan
E
,
Kloog
Y
. 
Galectin-3 augments K-Ras activation and triggers a Ras signal that attenuates ERK but not phosphoinositide 3-kinase activity
.
J Biol Chem
2004
;
279
:
34922
30
.
64.
Shalom-Feuerstein
R
,
Cooks
T
,
Raz
A
,
Kloog
Y
. 
Galectin-3 regulates a molecular switch from N-Ras to K-Ras usage in human breast carcinoma cells
.
Cancer Res
2005
;
65
:
7292
300
.

Supplementary data