Analysis of naïve and memory CD4 and CD8 T cell populations in breast cancer patients receiving a HER2/neu peptide (E75) and GM-CSF vaccine

Abstract

We are conducting clinical trials of the E75 peptide as a vaccine in breast cancer (BrCa) patients. We assessed T cell subpopulations in BrCa patients before and after E75 vaccination and compared them to healthy controls. We obtained 17 samples of blood from ten healthy individuals and samples from 22 BrCa patients prior to vaccination. We also obtained pre- and post-vaccination samples of blood from seven BrCa patients who received the E75/GM-CSF vaccine. CD4, CD8, CD45RA, CD45RO, and CCR7 antibodies were used to analyze the CD4+ and CD8+ T cells by four-color flow cytometry. Compared to healthy individuals, BrCa patients have significantly more memory and less naïve T cells and more effector-memory CD8+ and less effector CD4+ T cells. Phenotypic differences in defined circulating CD4+ and CD8+ T cell subpopulations suggest remnants of an active immune response to tumor distinguished by a predominant memory T cell response and by untapped recruitment of naïve helper and cytotoxic T cells. E75 vaccination induced recruitment of both CD4+ and CD8+ naïve T cells while memory response remained stable. Additionally, vaccination induced global activation of all T cells, with specific enhancement of effector CD4+ T cells. E75 vaccination causes activation of both memory and naïve CD4+ and CD8+ T cells, while recruiting additional naïve CD4+ and CD8+ T cells to the overall immune response.

Introduction

Breast cancer (BrCa) ranks first among newly diagnosed non-skin cancers and second among cancer-related deaths in the US [1]. The HER2/neu proto-oncogene is amplified in approximately a third of human BrCas and serves as an independent predictor of aggressive tumor biology [2]. Patients with HER2/neu protein over-expressing breast carcinoma are at significantly increased risk of disease recurrence following conventional multi-modality cancer therapy [3, 4]. The finding of natural humoral and cellular immunity to HER2/neu protein in early and late stage HER2/neu over-expressing BrCa has helped to identify and establish this self-protein as a relevant target for the development of peptide-based vaccines in cancer immunotherapy [5–7]. Peptide-based vaccines have a number of important advantages over autologous and allogenic tumor cell as well as dendritic cell vaccines in that they are safe, inexpensive, easy to work with, chemically stable, readily modifiable, and lack potentially tolerizing cellular antigens. These immunogenic peptides derived from tumor-associated proteins lack oncogenic potential and can be combined with multiple other peptides and/or adjuvants.

We are conducting preventive clinical trials with E75, an immunogenic peptide from the HER2/neu protein recognized by cytotoxic T lymphocyte (CTL) from HLA-A2+ BrCa patients, to determine if long-lived, clinically relevant CD8+ T cell responses can be generated. However, the prediction of a clinical response to a tumor vaccine remains a formidable challenge largely due to a number of factors that include tumor heterogeneity, wide-ranging variations in immune responses, and degree of pre-existing immunity and T cell activation. As such, a critical aspect of evaluating our tumor vaccine initiative is the precise quantitative and qualitative monitoring of the antigen-specific T cell response. In this aspect we have recently validated the feasibility of the HLA-A2:Ig dimer reagent for directly quantifying, characterizing, and monitoring E75 peptide vaccine-specific peripheral blood CD8+ T lymphocytes, and are utilizing the dimer in all of our ongoing clinical cancer immunotherapy trials [8]. We have found that E75 immunization elicits tumor-specific immune responses characterized not only by antigen-specific cytolytic CD8+ T cell expansion and differentiation, but surprisingly also includes the activation and involvement of both CD4+ helper T cells and regulatory T cells. One possible explanation for this unanticipated co-activation of both CD4 and CD8 subpopulations by a MHC class I peptide is the inclusion of the biological adjuvant granulocyte macrophage-colony stimulating factor (GM-CSF) in our vaccine regimen. The global adjuvant activity of GM-CSF in stimulating both the cellular and humoral arms of the immune system has been well described in several other peptide-based and cellular vaccine studies [9–12]. Thus, single peptide-based vaccines including immunoadjuvants appear to have broad implications and may alter T cell subpopulations.

Most previous vaccine studies have focused almost exclusively on quantifying and characterizing the antigen-specific CD8 T cell responses in vaccinated individuals [13–17]. In order to move away from a restricted focus on the highly specific CD8 plane of the peptide-elicited immune response and obtain a rather global view of the immune landscape during the vaccination period, we have taken the approach of fully characterizing the naïve and memory CD4 and CD8 T cell phenotypes in the peripheral blood of these patients before and after administration of the vaccine. The immune compartment classification system of naïve, memory, and recently activated T cells utilized in our analysis is based on the well-defined and established surface antigens CD45RA, CD45RO, and CCR7, respectively, that have been reported previously by several investigators [18–22]. Based upon the findings from these extensive studies and their proposed lineage models, we have adhered to the accepted definition of CD45RA+ T cells as naïve and CD45RO+ T cells as memory and loss of CCR7 as an indication of antigen-specific activation. These T cell subpopulations (memory or naïve, CD4+ or CD8+) were then categorized as central- or effector-memory cells and naïve or end-effector T cells as described by Rene van Lier et al. and others based on the presence or loss of CCR7 expression [18–22].

In this study, we have investigated the circulating levels of naïve and memory CD4+ and CD8+ T lymphocytes in healthy individuals and patients with non-metastatic BrCa. We have also performed an extended analysis of these cells utilizing monoclonal antibodies (mAb) directed to T cell surface differentiation markers, and four-color flow cytometry to include evaluation of recently activated subpopulations of T lymphocytes in the circulation of these two groups of individuals. The immunophenotype of the various subsets of the circulating T cell repertoire was then assessed before and after intradermal E75/GM-CSF vaccination of treated high-risk BrCa patients without apparent residual disease.

Materials and methods

Patients, normal donors, and E75 vaccine clinical trial

The Institutional Review Board at the Department of Clinical Investigation, Walter Reed Army Medical Center, Washington, DC, and the United State Army Medical Research and Materiel Command, Fort Detrick, MD, approved the clinical protocol. The study’s principal aim is to determine if the HER2/neu peptide, E75, is useful as a preventive vaccine in BrCa patients, particularly those who are at high risk of tumor recurrence. In addition, the purpose of the ongoing trials is to determine the optimal dose of the immunoadjuvant, GM-CSF, and the optimal schedule of inoculations for the E75/GM-CSF vaccine. We currently have active clinical trials for both node-positive (NP) and node-negative (NN) BrCa patients. All patients were tested to be immunocompetent and have had histologically confirmed NN BrCa. These patients have all also undergone primary surgical and medical therapies and were considered to be without evidence of disease at the time of enrollment into the trial. For the present study, we have limited our analysis to the NN patients because this clinical arm has a larger number of patients who have completed the vaccination series. Importantly, all of the patients reported in this study were close to a year or more post-completion of their standard treatment, thereby strongly reducing, if not entirely eliminating, the possibility of any remnant effects from their chemotherapy or radiation regimens.

The BrCa patients were enrolled prospectively and baseline blood samples were drawn in order to determine their HLA status. Subsequently, these patients were tested to determine if they were positive or negative for HLA-A2 expression. HLA-A2+ patients were enrolled in the treatment group, and HLA-A2- patients were offered enrollment in the control arm of the study. Treatment group patients received a series of intradermal inoculations consisting of 100, 500, or 1,000 μg of E75 peptide with 250 μg of GM-CSF (Berlex, Seattle, WA, USA) monthly for six inoculations. Blood samples were drawn from the vaccinated patients prior to each inoculation and 1 and 6 months following completion of the vaccination series. The NIH Common Terminology Criteria for Adverse Events (Version 3, March 31, 2003) definitions of adverse events were applied. Normal, healthy donor blood samples were obtained from BRT Laboratories (Baltimore, MD, USA).

E75 peptide and vaccination protocol

The E75 peptide (HER2/neu, 369-377 KIFGSLAFL) was produced commercially in good manufacturing practices grade (GMP) by Multiple Peptide Systems (MPS) Inc. (San Diego, CA, USA). The purity of the peptide was verified by high performance liquid chromatography and mass spectrometry, and the amino acid content determined by amino acid analysis. The peptide was purified to >95%. Sterility and general safety testing was carried out by the manufacturer. Lyophilized peptide was reconstituted in 0.5 ml of sterile saline at various concentrations (100 μg in 0.5 ml, 500 μg in 0.5 ml, and 1 mg in 0.5 ml). The peptide was mixed with GM-CSF (Berlex) at 250 μg in 0.5 ml. The 1 ml inoculation was split and given intradermally at two sites within 5 cm of each other. The group of NN patients reported in this study were vaccinated with 500 μg of E75 peptide and 125 μg of GM-CSF at 0, 1 and 5 months for a total of three doses. The monthly inoculations were all administered into the same extremity.

Peripheral blood mononuclear cell (PBMC) isolation

Blood drawn from every BrCa patient prior to vaccination was utilized to assess levels of circulating subsets of T cells and compare various T cell subpopulation levels in normal, healthy subjects. For vaccinated BrCa patients, additional blood was drawn prior to each inoculation and at 1 and 6 months after completing the vaccination regimen in order to determine the immunological responses of T cell populations to the vaccine. About 40 ml of peripheral blood was drawn into Vacutainer^® CPT^™ Tubes (Becton Dickinson, Franklin Lakes, NJ, USA) and centrifuged for the isolation of PBMC populations. The cells were washed in HBSS and re-suspended in complete culture medium consisting of Iscove’s Modified Dulbecco’s medium containing 10% human AB serum (Gemini Bio-Products, Woodland, CA, USA) supplemented with 1X penicillin/L-glutamine/streptomycin, 1X sodium pyruvate, 1X non-essential amino acids, and 50 μM of 2-mercaptoethanol (Life Technologies, Rockville, MD, USA). The PBMC were used as a source of lymphocytes for direct staining with fluorescent-labeled antibodies. Aliquots of freshly isolated PBMC were also cryo-preserved in 90% FCS and 10% DMSO in liquid nitrogen for future testing. The PBMC were also prepared from normal, healthy blood donor samples (BRT Laboratories) to serve as a basis for comparison of circulating T cell subpopulation levels between BrCa patients and healthy subjects. The pre- and post-vaccination blood samples used for these studies were obtained prior to the start of the vaccine series and 1 month after the final vaccination, respectively. Due to the time schedules of clinical visits of the participants in our trials, the collection of blood samples from patients occurred either during the early morning hours or in the late afternoon time period. Given this range of time for the patients’ visits we made sure that all samples were processed and analyzed within a 24-h period of collecting the blood samples. The same situation was also encountered with the processing of samples from the healthy donors such that all samples were collected similarly.

HLA-A2 typing

The HLA-A2 status of the patients was confirmed by indirect staining with 10 μl of anti-HLA-A2 mAb, BB7.2, and MA2.1 (1:10 dilution of culture supernatant) (ATCC, Rockville, MD, USA) at 4°C for 30 min followed by 30 min of incubation with FITC-conjugated goat-anti-mouse antibodies (BioSource, Carlsbad, CA, USA) and analyzed on a FACSCalibur Analyzer (Becton Dickinson Biosciences, San Jose, CA, USA).

Immunofluorescent staining and flow cytometry analysis

The PBMC were stained with anti-CD45RA-FITC, anti-CD45RO-PE, anti-CD4-APC, anti-CD8-Cy5.5, and anti-CCR7-Biotin-avidin complex (Caltag Laboratories, Burlingame, CA, USA) to determine their immunophenotype. Four-color flow cytometry was performed using FACSCalibur (BD Biosciences). Data was collected on the total cell population and subsequent analysis was performed on lymphocytes (gated by forward and side scatter properties). The FACS datasets were analyzed using CellQuest software, Version 3.3 (BD). The E75 peptide was synthesized by MPS as per GMP specifications for use in our vaccine studies. The E37 peptide (FBP: 25–33) from folate-binding protein (FBP) was synthesized by the BIC Facility at the Uniformed Services University of the Health Sciences, Bethesda, MD, USA.

HLA-A2:Ig dimer assay and immuno-phenotype analysis

The HLA-A2:Ig dimer reagent was purchased from BD Biosciences. The HLA-A2 dimer is composed of a human HLA-A2 molecule fused to each of the antigen binding sites of a mouse IgG1 immunoglobulin molecule. The presence of peptide-specific CD8+ T cells in fresh PBMC was assessed using the dimer assay as previously described with modifications [8]. The HLA-A2:Ig dimer reagent was loaded with the peptide of interest by incubating 1 μg of dimer with an excess (5 μg) of the HER2/neu E75 vaccine peptide or the FBP E37-negative control peptide and 0.5 μg of β2 microglobulin (Sigma Chemical Co., St. Louis, MO, USA) at 37°C overnight. The incubated dimer preparations were then stored at 4°C until used. We have found that dimer preparations generated and stored in this manner can be stable for at least one week, but for this study all dimer preparations were used within a period of four days. The E37 peptide served as a negative control peptide, and dimers loaded with E37 were used for the determination of non-specific background staining by the dimer molecule. E75-loaded dimers were used to detect vaccine-specific T cells. Results are expressed as the percent of E75-specific CD8 cells (control E37 dimer results subtracted) of the total CD8+ population. Staining of freshly isolated PBMC with the HLA-A2:Ig dimer preparations was carried out in an identical manner and all incubations were done at 4°C. The PBMC were washed and re-suspended in PharMingen Stain Buffer (BD Biosciences) and were added at 5×10⁵ cells/100 μl/tube in 5 ml round-bottom polystyrene tubes (BD) and washed once with PSB. Human γ-globulin (Sigma) was added at 0.5 μg/tube and the samples were allowed to incubate for 5 min before adding the dimer preparations. The cells were then incubated with the E37- or E75-loaded dimers (at 1 μg dimer/tube) and CCR7-biotin for an additional 30 min and washed once in PSB. The cells were then re-suspended with 100 μl of rat-anti-mouse IgG1-PE (BD Biosciences) and incubated for 15 min. This was followed with the addition of CD8-FITC and StreptAvidin-PECy5 (both from BD Biosciences) to all the tubes and CD45RA-APC or CD45RO-APC (both from Caltag Laboratories) to one of two tubes for each of the dimers. The cells were incubated for a further 15 min and washed. Four-color fluorometric analysis was carried out on a FACSCalibur Analyzer (BD Biosciences). The lymphocyte population was gated on forward and side scatter and gated events were analyzed using the CellQuest program (BD Biosciences). The data is displayed as a dual parameter density plot correlating CD8-FITC and rat-anti-IgG1-PE fluorescence. Quadrants were set based upon staining obtained using the E37-negative control peptide-loaded dimers. Cellular events that were detected as staining positive for both CD8-FITC and anti-mouse-IgG1-PE (upper right quadrant of the histogram) were quantified as being CD8+ T cells expressing T cell receptors specific for the respective peptide-loaded HLA-A2 dimer molecule. The E75-dimer +ve CD8+ T cell population was then further analyzed for expression of CD45RA or CD45RO and CCR7.

Statistical analysis

Mean levels of circulating CD4+ and CD8+ CD45RA+ (naïve), CD45RO+ (memory), CCR7- (recently activated), CD45RA+CCR7- (effector), CD45RA+CCR7+ (naïve), CD45RO+CCR7- (effector-memory) and CD45RO+CCR7+ (central-memory) T cells between BrCa patients and normal, healthy subjects were compared by independent sample two-tailed Student’s t-test (assuming unequal variance). When comparing mean pre- and post-vaccination levels of circulating CD4+ and CD8+ CD45RA+ (naïve), CD45RO+ (memory), CCR7- (recently activated), CD45RA+CCR7- (effector), CD45RA+CCR7+ (naïve), CD45RO+CCR7- (effector-memory), and CD45RO+CCR7+ (central-memory) T cells in the vaccinated BrCa patients, the paired two-tailed Student’s t-test was used to calculate statistical significance. P < 0.05 was considered statistically significant in all comparisons.

Results

Analysis and comparison of CD4+ T cell subpopulations between healthy individuals and BrCa patients

The PBMC were prepared from 22 NN BrCa patients and from ten healthy volunteers without antecedent history of BrCa. The NN patients consisted of ten HLA-A2+ (subsequently vaccinated) and 12 HLA-A2- (not eligible for vaccination) individuals. Two of the healthy subjects had four and five blood samples, respectively, drawn at different times permitting us to assess (in a limited way) intra- and inter-donor variability due to confounding issues such as the flu or common cold, etc. This brought the total number of samples from healthy donors to 17 and the overall total PBMC samples to 39 for our analysis to compare the T cell populations in healthy individuals and BrCa patients. All 39 PBMC samples were stained with CD4, CD45RA, CD45RO, and CCR7 mAb in order to define the subpopulations of CD4+ T cells in these two groups of individuals. The results obtained (Table 1) show that BrCa patients had significantly less naïve CD4+CD45RA+ cells (31.1 ± 4.1 vs. 43.4 ± 4.1%, P < 0.04), and slightly more memory CD4+CD45RO+ T cells (72.9 ± 2.7 vs. 67.7 ± 2.8%, P < 0.19). The overall activation pattern of the CD4+ T cells was similar between the two groups (70.4 ± 4.8 vs. 67.3 ± 5.3%). A more thorough analysis was completed of the CD4+ T cells as defined by the four phenotypic and functional compartments: effector (CD45RA+CCR7-), naïve (CD45RA+CCR7+), effector-memory (CD45R0+CCR7-), and central-memory (CD45RO+CCR7+). As outlined in Table 1 and shown in Fig. 1a, there was a significant difference found in the effector compartment between BrCa patients with 18.9 ± 3.1% and healthy individuals with 28.9 ± 5.1% CD4+CD45RA+CCR7- T cells (P < 0.05). While the numbers of naïve and central-memory CD4+ T cells were quite similar between the BrCa patients and the healthy controls, there was a trend toward more effector-memory CD4+ T cells in the BrCa patients (54.9 ± 4.4 vs. 45.8+4.5%, P < 0.16).

Table 1.

Comparison of CD4+ T cell subpopulations between normal, healthy individuals, and BrCa patients

	CD45RAa	CD45ROb	CCR7-c	CD45RA		CD45RO
				CCR7-	CCR7+	CCR7-	CCR7+
Normal	43.4 ± 4.1d	67.7 ± 2.8	67.3 ± 5.3	28.9 ± 5.1	17.4 ± 3.7	45.8 ± 4.5	19.7 ± 4.6
BrCa	31.1 ± 4.1	72.9 ± 2.7	70.4 ± 4.8	18.9 ± 3.1	13.4 ± 2.7	54.9 ± 4.4	18.5 ± 3.3
P-value	0.04	0.19	0.7	0.05	0.38	0.16	0.83

BrCa breast cancer

^aCD45RA = naïve T cells

^bCD45RO = memory T cells

^cCCR7⁻ = recently activated T cells

^dValues show percent change ± standard error

Fig. 1.

Comparison of (a) CD4+ and (b) CD8+ T cell subpopulations between normal, healthy donors (n=17) and BrCa patients (n=22). In addition to CD4 and CD8, antibodies to the well-characterized cell surface markers CD45RA (naïve), CD45RO (memory), and CCR7 (recently activated) were utilized and subsequent T cell labeling analyzed by four-color flow cytometry. Phenotypic and functional subpopulations were categorized as effectors (CD45RA+CCR7-), naïve (CD45RA+CCR7+), effector-memory (CD45RO+CCR7-), and central-memory (CD45RO+CCR7+)

Analysis and comparison of CD8+ T cell subpopulations between healthy individuals and BrCa patients

A similar approach of immunofluorescent staining and flow cytometry analysis was performed using CD8 mAb instead of CD4 to look at the levels of CD8+ subpopulations in healthy individuals and BrCa patients. Interestingly, the results obtained (Table 2) revealed strikingly similar and even more significant findings whereby BrCa patients had significantly less naïve CD8+CD45RA+ cells (64.5 ± 3.8 vs. 73.3 ± 4.1%, P < 0.046) and significantly more memory CD8+CD45RO+ T cells (47.3 ± 3.0 vs. 35.5 ± 2.8%, P < 0.009) than the healthy controls. Additionally, further analysis into the phenotypic and functional compartments of these CD8+ T cells revealed a similar pattern to the CD4+ T cells, i.e., the same relationship between BrCa patients and healthy individuals in terms of the effector, naïve, and central-memory subpopulations, with the exception that the magnitude of the latter being appreciably lower in the CD8+ T cells compared to the CD4+ T cells (Table 2). However, the BrCa patients had a significantly higher percentage of effector-memory (CD8+CD45RO+CCR7-) T cells compared to the healthy controls (43.2 ± 3.7 vs. 35.2 ± 3.4%, P < 0.045) (Fig. 1b).

Table 2.

Comparison of CD8+ T cell subpopulations between normal, healthy individuals, and BrCa patients

	CD45RAa	CD45ROb	CCR7-c	CD45RA		CD45RO
				CCR7-	CCR7+	CCR7-	CCR7+
Normal	73.3 ± 4.1d	35.5 ± 2.8	75.2 ± 5	53.9 ± 5.9	17.4 ± 3.1	35.2 ± 3.4	3.2 ± 0.9
BrCa	64.5 ± 3.8	47.3 ± 3	81.6 ± 4.1	45.1 ± 4.2	13.7 ± 3.5	43.2 ± 3.7	4.5 ± 1.2
P-values	0.046	0.009	0.33	0.23	0.46	0.045	0.44

^aCD45RA = naïve T cells

^bCD45RO = memory T cells

^cCCR7⁻ = recently activated T cells

^dValues show percent change ± SE

Changes in the CD4+ and CD8+ T cell subpopulations in BrCa patients as a result of receiving the HER2/neu E75 peptide vaccine

The HER2/neu vaccine administered to the patients enrolled in our clinical trial consists of the E75 peptide mixed with GM-CSF as an adjuvant. The GM-CSF has been reported in previous studies to not only specifically activate the CD8+ arm of the immune system, but also potentially have a global stimulatory effect on other cells, including the CD4+ T cell population [9–12]. Therefore, it was clearly an important aim in this study to compare the immunophenotype of the CD4+ and CD8+ T cell subpopulations in the peripheral blood of the BrCa patients prior to and after receiving the HER2/neu E75 peptide vaccine. As shown in Tables 3 and 4, we were able to analyze and compare the levels of CD8+ and CD4+ subpopulations, respectively, in the peripheral blood samples of seven BrCa patients who had received the vaccine.

In our analysis of the CD8+ subpopulations of T cells from the pre- and post-vaccination peripheral blood samples, we found that the CD8+ naïve subpopulation (CD8+CD45RA+) was significantly increased in these patients as a result of vaccination with the E75/GM-CSF vaccine (67.4 ± 7.9 vs. 58.8 ± 6.7%, P < 0.011) (Table 3). The individual patient responses are given in Fig. 2a. The CD8+CD45RO+ T cells remained similar before and after vaccination (41.6 ± 4.7 vs. 41.9 ± 4.2%, P < 0.82). A more thorough analysis of the phenotypic and functional CD8+ T cell compartments failed to demonstrate any significant differences in the four CD8+ T cell subpopulations (Table 3). The individual patients’ responses to vaccination in terms of memory, activation, and subpopulation analysis are shown in Fig. 2a–d. These results indicate that there is recruitment of naïve CD8+ T cells and maintenance of the memory CD8+ T cells in response to the vaccination.

Table 3.

Changes in CD8+ T cell subpopulations in BrCa patients after receiving the HER2/neu E75 peptide vaccine

	CD45RAa	CD45ROb	CCR7-c	CD45RA		CD45RO
				CCR7-	CCR7+	CCR7-	CCR7+
Pre	58.8 ± 6.7d	41.6 ± 4.7	69.6 ± 8.7	36.5 ± 6.8	26.9 ± 7.1	35.2 ± 6.8	9.1 ± 3
Post	67.4 ± 7.9	41.9 ± 4.2	80.8 ± 8.7	50.8 ± 7.5	13.5 ± 6.7	35.2 ± 7.5	5.9 ± 2.7
P-values	0.011	0.82	0.4	0.21	0.2	0.99	0.5

^aCD45RA = naïve T cells

^bCD45RO = memory T cells

^cCCR7- = recently activated T cells

^dValues show percent change ± SE

Fig. 2.

Individual patients’ pre- and post-vaccination levels of CD8+ T cell populations in vaccinated patients. a Naïve CD8+ T cells (CD8+CD45RA+). b Recently activated CD8+ T cells (CD8+CCR7-). c Effector CD8+ T cells (CD8+CD45RA+CCR7-). d Effector-memory CD8+ T cells (CD8+CD45RO+CCR7-). The mean pre- versus post-values were significant for the naïve CD8 T cell population (67.4 ± 7.9 vs. 58.8 ± 6.7%, P < 0.011) shown in (a)

Somewhat surprisingly, some of the most significant differences post-vaccination were found in the CD4+ T cell compartment (Table 4). In this compartment, the level of CD4+CD45RA+ naïve T cells was significantly increased in post-vaccination samples of peripheral blood (43.1 ± 11.1 vs. 26.7 ± 8.3%, P < 0.036). The individual patient responses are shown in Fig. 3a. Similarly, the overall levels of recently activated CD4+ T cells (CD4+CCR7-) was also significantly increased in the post-vaccination samples (93.4 ± 1.1 vs. 53.8 ± 7.2%, P < 0.004) (Fig. 3b). In an extended compartment analysis, Table 4 reveals an almost 3-fold specific increase (41.9 ± 11.2 vs. 15.1 ± 5.7%, P < 0.027) in the CD4+ effector T cell (CD4+CD45RA+CCR7-) subpopulation (Fig. 3c). In contrast, the levels of memory (CD4+CD45RO+) and effector-memory CD4 T cells (CD4+CD45RO+CCR7-) remained almost constant (66.9 ± 5.3 vs. 61.8 ± 8.2%) or only showed a non-significant trend (37.5 ± 6.1%, 58.0 ± 9.1%, P < 0.12), respectively, in pre- versus post-vaccinated blood samples (Fig. 3d). Both naïve and central-memory CD4+ T cells were significantly less in the post-vaccination samples reflecting the global activation due to the vaccination process (Table 4).

Table 4.

Changes in CD4+ T cell subpopulations in BrCa patients after receiving the HER2/neu E75 peptide vaccine

	CD45RAa	CD45ROb	CCR7-c	CD45RA		CD45RO
				CCR7-	CCR7+	CCR7-	CCR7+
Pre	26.7 ± 8.3d	66.9 ± 5.3	53.8 ± 7.2	15.1 ± 5.7	17.2 ± 4	37.5 ± 6.1	25.4 ± 5.3
Post	43.1 ± 11.1	61.8 ± 8.2	93.4 ± 1.1	41.9 ± 11.2	3.3 ± 1.1	58.0 ± 9.1	3.8 ± 0.7
P-value	0.036	0.9	0.004	0.027	0.011	0.12	0.012

^aCD45RA = naïve T cells

^bCD45RO = memory T cells

^cCCR7- = recently activated T cells

^dValues show percent change ±SE

Fig. 3.

Individual patients’ pre- and post-vaccination levels of CD4+ T cell populations in vaccinated patients. a Naïve CD4+ T cells (CD4+CD45RA+). b Recently activated CD4+ T cells (CD4+CCR7-). c Effector CD4+ T cells (CD4+CD45RA+CCR7-). d Effector-memory CD4+ T cells (CD4+CD45RO+CCR7-). The mean pre- versus post-values were significant for the a naïve CD4+ T cells (43.1 ± 11.1 vs. 26.7 ± 8.3%, P < 0.036), b recently activated CD4+ T cells (93.4 ± 1.1 vs. 53.8 ± 7.2%, P < 0.004), and c effector CD4+ T cells (41.9 ± 11.2 vs. 15.1 ± 5.7%, P < 0.027)

CD45RA, CD45RO, and CCR7 expression on E75-HLA:Ig dimer-positive, vaccine-specific CD8 T cells pre- and post-vaccination

We have previously shown, by using the HLA-A2:Ig dimer assay, that the E75 peptide vaccine used in our clinical trials has the ability to induce the expansion of E75 vaccine-specific CD8 T cells in the vaccinated patients [8]. Furthermore, we have shown by cytotoxicity [23] and ELISPOT [24] assays that these vaccine-induced E75-specific CTL are functional. We have recently been successful at modifying and extending the application of the HLA:Ig dimer assay to define and dissect the specific immunophenotype/s of the E75 vaccine-specific CD8 T cells. Therefore, in an extension of the current study we have measured the CD45RA, CD45RO, and CCR7 expression on the E75 vaccine-specific (E75-HLA:Ig dimer +ve cells) in peripheral blood samples from a group of five BrCa patients prior to and after receiving their first vaccine dose. The pre- and post-vaccine levels of E75-dimer +ve CD8 cells and the expression of the above molecules on these antigen-specific cells for these five patients are shown in Table 5. These results demonstrate a wide range of cell numbers emerging from both the naïve and/or memory pools depending on the individual patient. The technical challenges of further characterizing the small population of E75-dimer +ve CD8 cells make it difficult to fully interpret these findings; however, the possibility remains that the pattern within these responses may be related to the patient’s pre-existing HER2/neu immunity.

Table 5.

Expression of CD45RA, CD45RO, and CCR7 on E75-HLA:Ig dimer-positive vaccine-specific CD8 T cells pre- and post-vaccination

E75-dimer +ve CD8 T cells (% of total CD8 cells)			% of E75-dimer +ve CD8 T cells
			CD45RA		CD45RA		CD45RO		CD45RO
			CCR7+	CCR7+	CCR7-	CCR7-	CCR7+	CCR7+	CCR7-	CCR7-
Patient	Pre	Post first vaccine	Pre	Post	Pre	Post	Pre	Post	Pre	Post
BrCa #1	0.72	0.59	40.0	50.0	11.3	9.5	69.9	29.9	10.5	16.9
BrCa #2	0.22	0.49	52.8	25.9	19.4	46.6	65.9	24.2	29.3	39.0
BrCa #3	0.26	0.42	22.1	61.0	40.0	0.9	11.4	45.2	22.9	1.9
BrCa #4	0.29	0.41	23.5	47.2	45.1	36.8	5.6	31.8	48.2	18.2
BrCa #5	0.57	0.71	24.43	0	48.85	63.77	16	34.59	26	62.41

Discussion

In this study, we have utilized immunostaining and four-color flow cytometry to characterize the immunological phenotype of circulating CD4+ and CD8+ T lymphocytes in the peripheral blood of BrCa patients and healthy individuals. Additionally, in a subpopulation of the BrCa patients we have also analyzed these cells before and after the administration of a HER2/neu peptide vaccine. In so doing, important distinctions have been established for specific T cell subpopulations between normal subjects and immunocompetent BrCa patients. The study’s findings suggest an active immune response to tumor distinguished by a predominant memory T cell response and also by untapped recruitment of naïve helper and cytotoxic T cells. Furthermore, alterations in the immune landscape were noted as phenotypic changes in defined circulating CD4+ and CD8+ T cell subpopulations following vaccination with E75 peptide/GM-CSF in HLA-A2, high-risk patients treated for HER2/neu over-expressing BrCa.

Specifically, peptide vaccination was associated with expansion of circulating naïve CD4+ and CD8+ T cells, recently activated CD4+CCR7- T cells in general, and recently activated CD4+CD45RA+CCR7- naïve T cells in specific. Peripheral CD4+ and CD8+ memory T cell populations including recently activated memory T cells remained stable following vaccination. Although, the current peptide vaccination strategy is associated with significant antigen-specific T cell expansion, differentiation of naïve to terminal effector CD8+ T cells is incomplete. The finding of a significant proportional decrease in naïve CD4+ and naïve CD8+ T cells in the peripheral blood of non-vaccinated, immunocompetent BrCa patients may be related to tumor immune tolerance in the context of blunted immune surveillance and/or immunosuppression effects from occult tumor. Interestingly, the BrCa patients demonstrated significant increases in circulating memory CD8+ T cells, suggesting the possibility of an active, though predominantly memory T cell-driven, anti-tumor response characterized by untapped recruitment of naïve T cells. Thus, HLA-A2 patients with treated NN HER2/neu-expressing BrCa demonstrate reductions in peripheral and effector CD4+ T cells consistent with increased circulating memory lymphocyte response to prior tumor antigen exposure and/or cancer-related immunosuppression.

Based upon the studies of Rene van Lier et al. and others, mAb directed to CD45RA, CD45RO, and CCR7 molecules can, up to a certain extent, stratify naïve and memory T cells into different effector subsets [18–22]. The results reported by these different groups of investigators who analyzed immune responses in a large number of individuals against endogenous viruses such as CMV, EBV, and influenza have resulted in the emergence of lineage models that are still in an “evolving” stage of clarity. The ambiguous features associated with the proposed differentiation pathways from their studies is largely due to a combination of complex factors such as time points of initial and subsequent encounters of the T cells with the antigen(s) as well as the plasticity and/or reversibility of expression of some of the above molecules [18–22]. Long-lived lymphoid tissue-centric central-memory T cells (CD45RA-CCR7+) lack immediate effector function but are able to respond to recall antigens; conversely, effector-memory T cells (CD45RA-CCR7-) target peripheral potentially detrimental antigens and are capable of immediate and potent cytolytic response [18]. Effector T lymphocyte populations are also subdivided into two phenotypic subsets on the basis of CD45RA and CCR7 expression: unprimed-naïve (CD45RA+CCR7+) and antigen-primed end-effector (CD45RA+CCR7-) T cells. CD4+ T cells contribute to CD8+ T lymphocyte activation and differentiation to cytolytic effector T cells and are a critical element in sustaining CD8+ T cell responses to MHC class I-restricted peptides. In this study, both recently activated CD4+CCR7- T cells (i.e., all CD4 T cells) as well as recently activated naïve CD4+CD45RA+CCR7- T cell subpopulations (i.e., only CD4+CD45RA+ T cells) demonstrated significant expansion over pre-immunization baseline levels supporting further the effectiveness in generating immune responses with the vaccination schedule being implemented.

An immunodominant peptide epitope of the HER2/neu proto-oncogene, E75 (HLA-A2 binding epitope, p369-377 KIFGSLAFL) is an appealing candidate antigen for protective peptide vaccine trials given its low level of expression in normal tissue and frequent over-expression in various common epithelial cancers such as breast, ovary, and prostate carcinoma [25–32]. In our ongoing adjuvant peptide vaccine trials for patients with high-risk BrCa but without evidence of detectable disease, we have found that E75 in conjunction with the GM-CSF immunoadjuvant safely and effectively stimulates peripheral CTL [23]. Similar clinical safety profiles have been demonstrated with other HER2/neu helper peptide vaccine trials in patients with breast and ovarian cancer [26]. Our tumor-specific E75 peptide vaccination has as its primary aim the establishment of long-lived immune memory and prevention of cancer recurrence. Reduced levels of peripheral naïve CD4+ and CD8+ T cells demonstrated in BrCa patients in this study may represent untapped potential for anti-tumor immunity. E75/GM-CSF vaccination was associated with significant increases in circulating naïve CD4+ and CD8+ T cells, suggesting that this vaccination strategy might be recruiting and activating a dormant limb of the T cell repertoire. The finding of stable post-vaccination CD4+ and CD8+ memory T cell populations in the context of expansion of the naïve T cell compartment points to the potential to expand the memory T cell pool with the current preventive peptide vaccine strategy.

In this study, BrCa patients apparently free of disease following multimodality therapy began peptide-specific vaccination with a significant deficit of end-effector CD8+CD45RA+CCR7- T cells. There was a tendency of a proportional, relative increase of antigen-specific CD8+ effector T cells with vaccination compared to pre-immunization baseline levels. However, this amounted to restoration of the effector CD8+ T cells to circulating levels quantified in healthy subjects. The current vaccination strategy is associated with antigen-specific T cell expansion, but differentiation of naïve to end-effector CD8+ T cells appears incomplete. This partial differentiation to tumor-specific cytotoxic CD8+ T lymphocytes has also been recognized in HLA-A2 melanoma patients vaccinated with Melan-A-peptide [33]. This finding requires further study, and if confirmed may warrant modification of the current peptide vaccine approach. The addition of HER2/neu-specific MHC class II-restricted epitopes to the E75/GM-CSF vaccine or vaccination with HER2/neu-specific T helper epitopes containing MHC class I motifs are strategies being explored for enabling peptide-specific T cells to recognize naturally processed endogenous antigen and achieve long-lived anti-tumor immunity [26, 34]. Although, distinctly remote, there is a possibility based on previous investigations by others looking at lymphocyte recovery in patients with advanced BrCa receiving high-dose chemotherapy that the standard clinical treatment administered to the patients prior to entering our study could somehow have an impact on the levels of the peripheral blood T cells analyzed in this study [35, 36]. We would consider this to be extremely unlikely based on the fact that all of the patients reported here were approaching 1 year or more post-completion of their treatment and did not receive high dose chemotherapy.

Phenotypic and functional characterization and correlation is an important aspect of immune surveillance in clinical vaccine trials; however, consistently correlating effector T cell response with clinical efficacy is challenging. One phase I trial of CEA peptide-loaded dendritic cells found a significant correlation between magnitude of expansion of CD8+CD45RA+CD27-CCR7- T cells and clinical response in patients with advanced colorectal and lung cancer [37]. Another study of patients with metastatic melanoma vaccinated with autologous tumor-derived heat-shock-protein-peptide complexes showed similar parallels between tumor-specific cytolytic T cell and clinical responses [38]. A more recent study extended these observations to determine if pre-immunization T cell activation could predict the response to a peptide vaccine [33]. Speiser et al. found that the percentage of tumor-specific CD8+CD28- T cells pre-vaccination was significantly increased in HLA-A2 melanoma patients responding to vaccination (response defined by >2-fold expansion of Melan-A-specific T cells) [33]. Responders to immunization had significantly higher circulating levels of pre-immunization antigen-specific CD28 cytotoxic CD8+ T cells than non-responders, consistent with pre-existent “tumor-driven” immunity. Ongoing research efforts are aimed at clarifying the relationship between degree of immune pre-activation and clinical efficacy in our preventive peptide vaccine trials.

The present study represents an important first step toward precise phenotypic and functional characterization of the naïve and memory T cell repertoire in the context of a preventive HER2/neu peptide vaccine strategy. Such a detailed analysis of the immune landscape in cancer patients will be integral to the continued improvement and development of successful vaccines and preferred adjuvants for immunotherapy. The findings reported here provide a preliminary platform to explore these complex interactions and consider further study refinements of the present trial. For example, a control arm, consisting of GM-CSF administration only, will be incorporated to the present clinical trial so as to further delineate the phenotypic T cell response to our peptide vaccination and address the effects that may be due specifically to GM-CSF.

In the current study, we have also made an initial attempt to clearly define and dissect the specific phenotypes of the E75 vaccine-specific CD8 T cells. At this time we have only been able to study the CD45RA, CD45RO, and CCR7 expression on the E75 vaccine-specific (E75-HLA:Ig dimer +ve cells) for an initial group of patients prior to and after receiving their first vaccine dose. The results obtained indicate the presence of a wide range of subpopulations of cells residing and/or emerging from both the naïve and/or memory pools depending on the individual patient. Interestingly however, the range and degree of variability appears to mirror similar findings reported by Ravkov et al. in their analysis of T cell subsets in CD8 T cells specific for viral epitopes (CMV, EBV, and Flu) using HLA tetramer-staining assays [22]. In future experiments, we also hope to assess the cytokine secretion activity in the T cell populations from these patients using an ex vivo ELISPOT assay which would then allow us to correlate important functional aspects associated with the various subsets of the E75 vaccine-specific CD8 T cells.

In conclusion, we hope that the eventual phenotypic and functional characterization of E75-specific subsets of CD8+ cytotoxic T cells will help further define tumor-specific immune responses, with the aim of enhancing predictability of cellular immune response to the peptide that will correlate with clinically relevant endpoints.