Cancer Immunology, Immunotherapy

, Volume 56, Issue 9, pp 1485–1499

Vaccination with p53 peptide-pulsed dendritic cells is associated with disease stabilization in patients with p53 expressing advanced breast cancer; monitoring of serum YKL-40 and IL-6 as response biomarkers

Authors

    • Department of OncologyCopenhagen University Hospital
    • Center for Cancer Immune Therapy, Department of HematologyCopenhagen University Hospital
  • Anders E. Pedersen
    • Department of Medical Anatomy, The Panum InstituteUniversity of Copenhagen
  • Julia S. Johansen
    • Department of RheumatologyCopenhagen University Hospital
  • Hans E. Johnsen
    • Center for Cancer Immune Therapy, Department of HematologyCopenhagen University Hospital
  • Dorte Nielsen
    • Department of OncologyCopenhagen University Hospital
  • Claus Kamby
    • Department of OncologyCopenhagen University Hospital
  • Svend Ottesen
    • Department of OncologyRoskilde Hospital
  • Eva Balslev
    • Department of PathologyCopenhagen University Hospital
  • Eva Gaarsdal
    • Center for Cancer Immune Therapy, Department of HematologyCopenhagen University Hospital
  • Kirsten Nikolajsen
    • Center for Cancer Immune Therapy, Department of HematologyCopenhagen University Hospital
  • Mogens H. Claesson
    • Department of Medical Anatomy, The Panum InstituteUniversity of Copenhagen
Original Article

DOI: 10.1007/s00262-007-0293-4

Cite this article as:
Svane, I.M., Pedersen, A.E., Johansen, J.S. et al. Cancer Immunol Immunother (2007) 56: 1485. doi:10.1007/s00262-007-0293-4

Abstract

p53 mutations are found in up to 30% of breast cancers and peptides derived from over-expressed p53 protein are presented by class I HLA molecules and may act as tumor-associated epitopes in cancer vaccines. A dendritic cell (DC) based p53 targeting vaccine was analyzed in HLA-A2+ patients with progressive advanced breast cancer. DCs were loaded with 3 wild-type and 3 P2 anchor modified HLA-A2 binding p53 peptides. Patients received up to 10 sc vaccinations with 5 × 106 p53-peptide loaded DC with 1–2 weeks interval. Concomitantly, 6 MIU/m2 interleukine-2 was administered sc. Results from a phase II trial including 26 patients with verified progressive breast cancer are presented. Seven patients discontinued treatment after only 2–3 vaccination weeks due to rapid disease progression or death. Nineteen patients were available for first evaluation after 6 vaccinations; 8/19 evaluable patients attained stable disease (SD) or minor regression while 11/19 patients had progressive disease (PD), indicating an effect of p53-specific immune therapy. This was supported by: (1) a positive correlation between p53 expression of tumor and observed SD, (2) therapy induced p53 specific T cells in 4/7 patients with SD but only in 2/9 patients with PD, and (3) significant response associated changes in serum YKL-40 and IL-6 levels identifying these biomarkers as possible candidates for monitoring of response in connection with DC based cancer immunotherapy. In conclusion, a significant fraction of breast cancer patients obtained SD during p53-targeting DC therapy. Data encourage initiation of a randomized trial in p53 positive patients evaluating the impact on progression free survival.

Keywords

Dendritic cellsBreast cancerp53 peptidesImmunotherapyBiomarkers

Introduction

Dendritic cells (DCs) are utilized as natural adjuvant in cancer vaccines as they are extremely efficient antigen-presenting cells and potent stimulators of memory-B and -T cells as well as resting or naïve T cells [14]. Thus, several vaccines have been constructed comprising peptide-pulsed DCs based on prior identification of cytotoxic T-cell (CTL)-defined synthetic peptide epitopes and the presenting major histocompatibility complex (MHC) molecules. DCs for antigen loading can be generated in clinical scale from blood leukapheresis products followed by in vitro manipulation and growth factor stimulation [5, 6], and thereafter loaded with the synthetic MHC binding peptides known to stimulate CTLs.

p53 is an obvious target for cancer vaccination therapy [7]. Mutations that inactivate the p53 protein or members of the p53 pathways are the most common genetic alterations found in human cancers, and mutations in the p53 gene are found in approximately 50% of human tumors including about 30% of breast carcinomas associated with a poor prognosis following conventional therapy [810]. More than 85% of p53 mutations result in single amino-acid substitutions that lead to the synthesis of stable, inactive p53 protein, which accumulate in the nucleus and cytosol of the tumor cells. Tumor cells expressing high cytosolic levels of p53 are potential targets for recognition, and lysis by p53 specific CTLs as peptides derived from accumulated p53 protein are presented by class I MHC molecules on the cell surface, thereby, acting as tumor-associated epitopes. In accordance with this, CTLs with HLA class I restricted specificity for wild-type p53 peptides have been observed in peripheral blood of cancer patients suggesting immune surveillance of tumors expressing these peptides [1114]. Due to the diversity of p53 mutations, peptides representing wild-type sequences are preferable as basis for a broad-spectrum p53-targeting cancer vaccine.

We have established a p53-targeting vaccine using autologous DCs pulsed with six HLA-A2 binding p53 peptides for treatment of metastatic breast cancer. Three of the peptides represent wild-type sequences and were chosen to avoid the need for individualized vaccine targets [15, 16]. The other three p53 peptides included a single P2 anchor amino acid modification to increase HLA-A2 binding capacity and induction of p53-specific CTLs [17]. Preclinical studies have shown that these wild-type p53 derived HLA-A2 binding peptides are able to activate human T cells and generate effector T cells, which are cytotoxic to human HLA-A2+, p53+ tumor cells [18, 19]. In a recently published phase I trial, the safety of the p53-DC vaccine was analyzed in six patients with progressive breast cancer [20]. Induction of p53 specific T cells and indications of clinical effect were observed in some of the patients without any toxicity of significance.

Application of response biomarkers are important for identification of patients responding to treatment. The cytokine interleukine-6 (IL-6) and the protein YKL-40 could be potential candidates. They are normally produced by immune cells; T-and B-cells (IL-6) [21], macrophages, and granulocytes (YKL-40) [22]. However, they are also found to be constitutively expressed by a variety of tumor cells among these breast carcinoma cells [22, 23]. IL-6 can inhibit DC differentiation [24], and both IL-6 and YKL-40 are able to promote tumor growth by up-regulating anti-apoptotic proteins in the tumor cells and remodeling of the extracellular matrix during cell proliferation and invasiveness [25, 26]. Studies in several types of cancer patients including breast cancer have shown a negative impact on prognosis of elevated serum levels of YKL-40 and IL-6 [22, 27, 28]. However, knowledge of the value of serum IL-6 as predictive and response biomarker is very sparse and the applicability of monitoring serum YKL-40 during anticancer treatment for evaluation of response is previously not described.

In this phase II trial, autologous DCs loaded with a cocktail of 3 wild-type and 3 modified p53-peptides are analyzed in 26 HLA-A2+ patients with progressive metastatic breast cancer. Patients are evaluated for toxicity, clinical response, and induction of p53 specific T-cell immunity. Furthermore, measurement of selected serum biomarkers during treatment is carried out to evaluate their feasibility as indicators of clinical response.

Methods

Patients and eligibility criteria

Between June 2002 and June 2004, 26 patients with progressive metastatic breast cancer were enrolled in this phase II trial, in agreement with the inclusion criteria: (1) histologically proven metastatic or locally advanced carcinoma of the breast, (2) progressive disease and no standard systemic treatment indicated, (3) at least one measurable lesion or osteolytic bone metastasis, (4) expression of the HLA-A2 allele, (5) WHO performance status 0–2, and (6) life expectancy more than 3 months. Main exclusion criteria included: (1) evidence of brain metastasis, (2) use of immunosuppressive drugs such as glucocorticoids, (3) radiation therapy or chemotherapy within the prior four weeks, (4) significantly increased blood liver-enzyme level (>2.5× upper normal limit), (5) other malignancies, or (6) pregnancy. The study protocol was approved by the Institutional Ethical Committee, Copenhagen County and the Danish Medicines Agency. Written informed consent was obtained from all patients.

Vaccine preparation

Generation of DCs

All procedures were performed according to Good Laboratory Practice standards as approved by the Danish Medicines Agency. Patients underwent unmobilized leukapheresis using a continuous flow blood cell separator for isolation of large scale (>2 × 109) mononuclear cells. The remaining red blood cells were lysed with Orthomune lysing solution (provided by hospital pharmacy). Peripheral blood mononuclear cells (PBMC) were washed, resuspended in culture medium (CM) (X-VIVO15, 2% l-glutamin 200 mM, 1% autologous heat inactivated plasma) at 7 × 106 cells/ml, and separated by 1 h adherence to plastic Nunclon dishes (Nunc, Biotech Line, Slangerup, Denmark). Non-adherent cells were removed and adherent cells were subsequently cultured for 7 days in CM supplemented with 250 U/ml rh-IL-4 (CellGenix, Freiburg, Germany) and 1000 U/ml GM-CSF (Leucomax, Schering Plough, Farum, Denmark). Cells were harvested at day 7 using a cell-scraper and cells with typical DC morphology were counted by light microscopy. Aliquots of a minimum of 5 × 106 DCs were frozen in 85% autologous serum, 10% DMSO (BDH limited pool, UK, GMP), and 5% Glucosteril 40% (Fresenius, Albertslund, Denmark) using automated cryopreservation (Planer Freezing Unit, Planer, UK). Sterility controls of DCs were negative at all times. Aliquots of the cultured cells were subjected to phenotypic analysis at time of cryopreservation. The expressions of the cell surface antigens, HLA-A2, HLA-DR, CD1a, CD11c, CD33, CD40, CD54, and CD86 were analysed [29].

Peptide loading of in vitro generated DCs

On the day of vaccination, one vial of DCs was thawed and washed twice; DC viability as determined by tryphan blue staining was 85–95%. DCs were resuspended in 500 μl RPMI and pulsed for 2 h at 37°C with a mixture of the six HLA-A*0201 binding p53-derived peptides and the pan-MHC class II binding peptide, PADRE [30]; each peptide in a concentration of 40 μg/ml. After incubation, cells were washed twice resuspended in 500 μl RPMI and transferred to a 0.5 ml insulin syringe for injection.

Treatment

Eligible patients were to receive a total of 10 immunizations, with at least 5 × 106 peptides pulsed autologous DCs. The first four vaccinations were given weekly; thereafter biweekly. The vaccine was administered sc near the inguinal region on the same thigh each time. Concomitantly with each vaccination, 6 MIU/m2 interleukin-2 (Proleukine, Swedish Orphan, Denmark) was administered sc.

Synthetic epitope peptides

DC were loaded with 6 different HLA-A*0201 binding p53 derived peptides (Table 1); three wild type peptides with high HLA-A*0201 binding affinities [18, 19, 3133] and three position 2 anchor modified HLA-A*0201 binding peptides [17]. An 11-amino acid pan-MHC class II binding peptide, PADRE (aKXVAAWTLKAAa), was added for induction of in vivo T-helper activation [30]. Purity of all peptides for clinical uses was >95%; these peptides were suspended at 1 mg/ml in RPMI, filtered through a 0.22 micron filter and tested for sterility.
Table 1

HLA-A2 binding p53 peptides used for DC loading prior to vaccination

Peptide number

Peptide name

Peptide sequence

Wild type p53 peptides

 1

R9V

p53 65–73 (RMPEAAPPV)

 2

L9V

p53 264–272 (LLGRNSFEV)

 3

G11V

p53 187–197 (GLAPPQHLIRV)

Modified p53 peptides

 4

S9V

p53 149–157 (SLPPPGTRV)

 5

K9V

p53 139–147 (KLCPVQLWV)

 6

Y9L

p53 103–111 (YLGSYGFRL)

ELISPOT control peptides

IMP58–66 GILGFVFTL from the influenza matrix protein, BMFL1280–288 GLCTLVAML from the immediate-early lytic protein of Epstein Barr virus and RTase309–317 ILKEPVHGV from the reverse transcriptase of HIV-1 virus. Wildtype HLA-A*0201 binding p53 derived peptides: p53149–157 STPPPGTRV (S9V), p53139–147 KTCPVQLWV (K9V) and p53103–111 YQGSYGFRL (Y9L). All peptides were purchased from Schafer N, Copenhagen, Denmark.

Clinical monitoring

Patients receiving one vaccination or more were evaluable for toxicity, and patients receiving four vaccinations or more were evaluable for tumor response. With the exception of patients with osteolytic bone metastasis, all patients underwent response evaluation according to the RECIST criteria [34, 35] at baseline, one week after the sixth (9 weeks after first treatment) and one week after the tenth (16 weeks after first treatment) DC vaccination. If treatment was continued, evaluation was performed every three months. Toxicity was graded according to the NCI Common Toxicity Criteria.

Blood samples

Heparinized peripheral blood samples were collected pre-immunization and post-immunization after the fourth, sixth, and tenth vaccinations. PBMC were separated by centrifugation on a Lymphoprep (NYCOMED, Norway) density gradient using standard procedures. Aliquots of PBMC were frozen in RPMI with 10% AB-serum and 10% DMSO.

In vitro stimulation of PBMC

Triplicates of PBMC (105/well) were cultured in X-VIVO15 medium (Biowhittaker, Wokingham, England) containing 2% inactivated human AB and 10 μg/ml relevant peptide in 96 well microtiter plates (Nunc, Denmark). RhIL-2 (Proleukine from Chiron, The Netherlands) 300 IU/mL was added on day 2. On day 10, cells were tested for peptide specific IFN-γ production in the ELISPOT assay.

ELISPOT assay

The IFN-γ ELISPOT assay was used to quantify p53 peptide specific CTLs during the vaccination trial. It was performed as previously described [20]. Each well of PBMC expanded for 10 days with the relevant peptide as described above were harvested separately and transferred directly to 96 well ELISPOT plates (Multiscreen, MAHAS4510 from Millipore, Molsheim, France) precoated with 7.5 μg/ml anti-human IFN-γ (M-700A from Endongen, USA). After 24 h culture at 37°C with relevant peptide, the ELISPOT plates were developed with 0.75 μg/ml biotinylated anti-human IFN-γ (M-701B from Endogen, USA) and HRP-streptavidin (DAKOCytomation, Denmark) and substrate (41CN from Sigma). The number of spot-forming cells was determined with computer-assisted image analysis software (KS ELISPOT, Zeiss, Munic, Germany).

DTH

Delayed-type hypersensitivity (DTH) reaction tests against: (1) RPMI (control), (2) mixed p53 peptides (0.5 μg/μl), and (3) mixed p53 peptides + PADRE peptide were applied intradermally at the same time points as the blood samples were drawn. DTH skin test reaction was measured as the diameter of induration and flushing at the injection site after 48 h.

ELISA for serum IL-6 and YKL-40

Serum samples were stored at −80°C until analysis for YKL-40 and IL-6. Serum concentrations of YKL-40 and IL-6 were determined using commercially available Elisa kits (YKL-40: #8020, Quidel, San Diego, CA [36]; IL-6: #HS600B, R&D Systems, Minneapolis, Minnesota). Sensitivity of the ELISA was 10 μg/l for YKL-40 and 0.11 ng/L for IL-6. The intra- and interassay coefficients of variation were less than 5.0 and 10.2% for YKL-40, and 10.5 and 17.7% for IL-6. All samples from each patient were analyzed on the same ELISA plate. The serum YKL-40 concentration in healthy subjects (= 245; median age 49, range 18–79) is 43 μg/l, and the plasma IL-6 concentration in healthy subjects (n = 320; median age 48, range 18–64) is 1.4 ng/l.

Immunohistochemical staining for p53

Sections of formalin-fixed, paraffin-embedded tissue from the primary breast carcinoma were cut and stained with antibody against p53 DO7, (DAKOcytomation), 1:50. In brief, the slides were immersed in citrate buffer solution for antigen retrieval and boiled in microwave for 10 min and washed in buffer solution (TBS). They were incubated with primary antibody for 1 h at room temperature and then washed in TBS. After 1 h of incubation in the secondary antibody, the sections were incubated with streptavidin-biotin-complex (DAKOcytomation). A multi-block including several different tissues was used for positive and negative controls. For estimation of positive reaction, only the strongly stained nuclei were counted as a percentage of all tumor nuclei. Less than 5% were estimated as negative [37].

Statistics

Survival was measured from the first vaccination until death or date of last follow-up. Survival distribution between the stable disease (SD) and progressive disease (PD) group was estimated using the Kaplan–Meier method. T-tests were performed to test for significant differences in mean values for DC yield and phenotype between the SD and PD group. Fisher’s exact test was used to test the association between clinical outcome and tumor p53 expression and vaccination induced p53 specific immune response.

Serum marker values were logarithmically transformed prior to analysis by Wilcoxon signed rank test applied to test for significant differences between pre- and post treatment values within the SD and PD groups. For inter-group comparisons of treatment associated changes in serum marker values, post values were divided with pre values in the respective groups and the computed ratios were subjected to Mann–Whitney test. All tests were two-sided and P-values less than 0.05 were considered significant.

Results

Patient characteristics

A total of 26 patients with progressive metastatic breast cancer were included. Characteristics of the treated patients are summarized in Table 2. All patients were HLA-A2 positive, and they had a mean age of 53 years. In 11 patients, metastatic disease was only present in one region, while the remainder of the patients had more widespread disease involving two to six regions. The majority (18/26) of the patients had previously received up to five different chemotherapy regimens, and 22/26 patients had received up to three different endocrine treatment regimes. All but one patient had at least one measurable tumor lesion; the remaining patient exclusively had osteolytic bone metastases. In 11/26 patients, tissue samples from the primary breast tumor stained positive (>5% positive cells) for p53 by immunohistochemistry.
Table 2

Patient and treatment characteristics

Patient Number

Age

Metastatic regions

Number of regions

Previous therapy (number of regimes)

p53 expr (%)

Total number of vacc

Clinical outcome after six vacc

Induced or increased p53 immune response

Survival after one vacc (months)

Chemo

Endocrine

Response at one evaluation

 14

58

liver

1

0

2

0

10

SD

+

14.5

 15

47

nodal

1

0

0

80

9

SD

+

9

 17

46

nodal

1

0

2

40

9

SD

NA

26.5

 19

64

nodal

1

0

2

30

>10

SD

21

 21

44

pleura, bone, nodal, ascites

4

0

1

30

10

PR minor

12.5

 29

58

lung

1

3

0

90

>10

SD

+

13

 35

74

lung, liver

2

2

2

nd

10

SD

+

5.5

 37

66

nodal, skin

2

5

1

nd

8

PR minor

>21

PD at one evaluation

 13

72

nodal, skin

2

1

1

0

7

PD

(+)

4

 16

54

pleura, liver, bone

3

2

2

0

5

PD

2

 20

59

skin, lung , bone, nodal

4

3

1

100

6

PD

NA

7.5

 22

33

lung, pleura, liver, bone, nodal

5

2

1

0

5

PD

NA

2

 23

46

nodal

1

1

0

90

7

PD

13

 26

51

lung

1

3

1

0

6

PD

10

 27

64

liver, bone

2

2

2

0

6

PD

2

 28

64

lung, pleura, liver, bone, nodal, ascites

6

2

1

50

5

PD

2

 31

52

bone

1

0

2

10

7

PD

3

 36

64

skin

1

1

2

0

6

PD

>21

 38

62

pleura, liver, bone, nodal

4

1

2

0

6

PD

(+)

9

Early termination

 10

51

lung, bone, nodal

3

0

2

0

3

PD

<1

 11

61

nodal

1

2

1

80

3

PD

ND

<1

 12

56

lung, liver, nodal

3

1

1

100

4

PD

+

8

 18

36

liver, bone

2

2

2

0

4

PD

+

4.5

 30

44

skin, lung, pleura, liver

4

2

3

0

4

PD

4

 32

58

skin

1

0

2

0

3

PD

4

 33

56

liver,bone, nodal

3

1

0

0

4

PD

1.5

+ Only immune response to a single peptide (see Table 4)

Dendritic cell preparations

The median number of PBMCs collected by leukapheresis was 4.2 × 109 (range 0.7–10.3 × 109) cells. Monocytes constituted median 14% (range 6.7–21.1) of the harvested PBMC and median 64% (range 17.4–100) of the monocytes developed into DC during culturing. According to the protocol, patients were to receive a minimum of 5 × 106 viable DCs pr vaccination; each patient was treated with a constant number of DCs pr vaccination but the DC content varied between patients with a median of 16 × 106 (range 8–43 × 106) DCs/vaccination as a consequence of DC preparation outcome.

None of these parameters were found to vary significantly between patients with PD or SD as clinical outcome; however, a tendency towards higher monocyte concentration (P = 0.15), DC yield (P = 0.15), and vaccine DC content (P = 0.16) were found in the SD group. Patients with early termination of treatment had significant lower monocyte (P = 0.02) and DC yield (P = 0.04).

The final DC cell population was phenotyped with a panel of antibodies and analyzed by four color flow cytometry. The phenotype was characteristic for intermediate mature DCs with some inter-patient variations [29], but no significant correlations between the DC parameters and clinical outcome or induced p53 specific immune responses were observed (data not presented).

Safety and toxicity

The vaccine was well tolerated. None of the patients developed skin toxicity at the site of vaccine injection. No signs of autoimmunity were observed. The most common side effect was mild to moderate local reaction at the site of proleukine injection. Furthermore, as expected, patients experienced CTC grade 1–2 flu-like symptoms 12–24 h after proleukine injection. All patients received IL-2 as described except for a single patient (pt. no. 19), where IL-2 administration was stopped after the seventh vaccination due to noticeable joint and muscle pain.

Clinical response

Nineteen patients were available for first evaluation after six vaccinations; none of the patients achieved objective response according to the RECIST criteria; eight patients obtained SD or minor regression (n = 2) and 11 patients had PD. Seven patients did not reach the of evaluation as treatment was discontinued after only 3–4 immunizations (2–3 weeks) due to clinical signs of rapid disease progression or death (one patient). This group of patients was designated ‘early termination’ and categorized as PD. Thus, a total of 18 patients had PD during treatment.

One patient obtained minor regression (no. 37) as CT showed regression of an axillary lymph node (LN) from 4.5 to 3.0 cm. Patient no. 21 had end stage disease including weekly need for drainage of bilateral pleural effusions and ascites at inclusion. At time of evaluation, ascites production had ceased and the pleural effusions significantly decreased. Furthermore, minor regression of axillary LNs from 1.2/1.6 cm to 0.9/1.0 cm was noted. Due to appearance of a new lesion in the breast, she was excluded after ten vaccinations but survived for another year.

Only five patients completed all ten planned immunizations. Two patients (no. 19 and 29) received additional monthly immunizations in continuation of the study for 16 and 9 months, respectively, due to prolonged SD.

According to clinical outcome, patients were divided into two groups: SD (n = 8) and PD (n = 18). In Table 3, disease extent and previous treatment intensity are listed for the two groups; it appears that in the PD group, a higher fraction of patients had more wide spread disease at time of inclusion. The previous exposure to endocrine therapy did not differ considerably; however, even though the SD group also included patients previously treated with several different chemotherapy regimes, 5/8 patients in the SD group were chemo-naïve compared to only 3/18 in the PD group (Table 2). The mean duration of survival differed significantly (P = 0.04) with a median survival of 13.8 months in the SD group compared to 4 months for patients with PD at first evaluation (Fig. 1).
Table 3

Disease and previous treatment pattern in SD and PD patients

 

SD patients (n = 8)

PD patients (n = 18)

Number of metastatic regions

1

>1

1

>1

 

5

3

6

12

Number of previous therapy regimes

0–1

>1

0–1

>1

Chemotherapy

5

3

9

9

Endocrine

4

4

9

9

https://static-content.springer.com/image/art%3A10.1007%2Fs00262-007-0293-4/MediaObjects/262_2007_293_Fig1_HTML.gif
Fig. 1

Kaplein–Meier plot of survival after first vaccination for patients with SD at first evaluation contra patients with PD at first evaluation

Positive p53 expression of the primary tumor as measured by immunohistochemistry was not used as inclusion criteria. However, p53 expression was found more frequently in tumors from patients achieving SD during treatment (Table 2); thus, 5/6 patients tested in the SD group expressed p53 whereas only 6/18 in the PD group (P = 0.06), indicating a possibly correlation between p53 expression and clinical outcome.

p53 specific CTL response

IFN-γ ELISPOT assay was applied to analyze the induction of specific CTLs for each of the six included p53 peptides. Among 26 patients, 25 were available for evaluation of immune response; however, in three patients, the background level of IFN-γ production was too high for data interpretation; thus, 22 patients were evaluable for immune reactivity. Eight of these 22 patients had induced or increased numbers of CTLs against one or more p53 derived peptides after immunization (Table 4), while 14 patients had no measurable induction of or increase in p53 specific CTLs. Response against wild-type peptides (1–3, Table 1) and P2 anchor modified peptides (4–6, Table 1) seems to be induced equally well (Table 4).
Table 4

Vaccination related immune response to p53 peptide 1–6 as evaluated by Elispot

Patient number

p53 peptide specific T cell reactivity (peptide number according to Table 1)

Pre-existing

Induced

Increased

Reduced

Response at one evaluation

 14

1,2,3,4

5,6

2,4

 

 15

 

4,5,6

  

 29

 

1,2,3,4,5,6

  

 35

 

1,2,4

  

PD at 1.evaluation

 13

 

1

  

 27

3,4

  

3,4

 36

1,2,3,4,6

  

1,2,3,4,6

 38

 

1

  

Early termination

 12

3,4,6

1,5

6

 

 18

2,3,4,5,6

1

2,5,6

 

Elispot responses were defined as positive when the number of spots was >20 and more than 2 background (vehicle). Induced or increased responses was defined as responses >2 pre-vaccine levels, and reduced responses was defined as responses 2< prevaccine levels at a given time point. Data represents means of triplicates with background subtracted

As seen from Table 2, four out of seven patients with SD at first evaluation had induction or increase in p53 specific CTLs, compared to only 2 out of 11 patients with PD at first evaluation (P = 0.14). Furthermore, these two patients with PD only generated p53 specific reactivity against a single peptide while reactivity against multiple peptide epitopes were seen in patients obtaining SD (Table 4). It also appeared that pre-existing p53 specific CTL reactivity was prevailing in patients exhibiting PD, and reduction of pre-existing reactivity was only observed in patients with PD (Table 4).

In most cases, an increase in the number of p53 peptide specific CTLs during vaccination was measured (Fig. 2a). However, there are some fluctuations with low CTL levels at certain time points and a tendency towards a more marked decline at late time points after vaccination; for most peptides, this reduction did not reach pre-vaccination levels.
https://static-content.springer.com/image/art%3A10.1007%2Fs00262-007-0293-4/MediaObjects/262_2007_293_Fig2_HTML.gif
Fig. 2

a CTL reactivity against modified and unmodified p53 peptides in breast cancer patients with SD during vaccination. PBMC were analyzed for p53 specific CTLs in the ELISPOT assay. Data is presented as mean of triplicates with background subtracted. b CTL cross-reactivity against modified and unmodified p53 peptide variants in a breast cancer patient before and after vaccination. PBMCs from patient no. 14 were analyzed for p53 specific CTLs in the ELISPOT assay as described in [20]. Peptides 1  p53149–157modified, 2  p53139–147modified, 3  p53103–111modified, 4  p53149–157wild type, 5  p53139–147wild type, 6  p53103–111wild type, 7  vehicle background

The modified peptides in the vaccine are supposed to induce CTLs in vivo with preserved specificity for the corresponding native unmodified peptides that is processed by the tumour cells. We therefore tested whether CTL reactivity against unmodified versions of the modified vaccine p53 peptides were also induced during immunization. As an example, Fig. 2b illustrates an increased CTL reactivity against corresponding modified and unmodified p53 peptides induced during vaccination.

None of the patients developed positive DTH skin reaction against naked p53 peptides during immunization.

Serum biomarkers

Several serum disease markers were measured including lactate dehydrogenase (LDH), alkaline phosphatase (ALP), YKL-40, and IL-6. Figure 3a and b depicts patient specific changes and mean serum levels during treatment, illustrating the different patterns for patients with PD and SD. Two types of comparisons were performed testing the association with clinical outcome; one evaluating changes in serum levels for the different markers pre- and post-treatment within the SD and PD groups, respectively, and another comparing the magnitude of treatment associated changes between the SD and PD groups.
https://static-content.springer.com/image/art%3A10.1007%2Fs00262-007-0293-4/MediaObjects/262_2007_293_Fig3_HTML.gif
Fig. 3

a Individual changes in serum alkaline phosphatase (U/l), LDH (U/l), YKL-40 (μg/L), and Il-6 (ng/l) during treatment for patients with PD and SD. The red curve illustrates the course of serum IL-6 in patient 21, who suffered from extensive disease dissemination and had clinical benefit from the treatment. b Mean serum level ±SD of alkaline phosphatase (U/l), LDH (U/l), YKL-40 (μg/l), and IL-6 (ng/l) pre-vaccination and after fourth vaccination for patients with PD and SD. The treatment associated changes in these serum markers all differed significantly (P < 0.05) between SD and PD patients (detailed in the text)

LDH and ALP are well known serum disease markers. Comparable mean serum levels were found in the PD and SD groups prior to therapy (Fig. 3b); however, as it appears from Fig. 3a, a few patients in the PD group had deviating high pre-values. Both serum LDH and ALP increased significantly (P < 0.05) in the PD group after four vaccinations, in contrast to the SD group where LDH decreased (P = 0.06), while ALP remained stable. The treatment associated changes in LDH and ALP were found to differ significantly (P < 0.05) between SD and PD patients.

High serum IL-6 and YKL-40 levels have been verified to have independent prognostic value in several types of cancer. Both factors were measured in 18 of the patients during treatment to assess their feasibility as response markers in connection with DC based cancer immunotherapy. As for LDH and ALP, noticeable differences were found among SD and PD patients. In general, YKL-40 increased in PD patients (P = 0.06) during treatment but remained unchanged in SD patient (Fig. 3b). This difference in treatment associated changes in serum YKL-40 between SD and PD patients was significant (P = 0.03). As for LDH and ALP, some patients in the PD group had relatively high serum YKL-40 and IL-6 pre-values.

The pattern of serum IL-6 changes differed from the other serum markers as a significant decline was observed in both SD (P = 0.02) and PD patients (P = 0.04) after fourth vaccination. However, the relative decline in serum IL-6 was significantly higher in the SD group (P = 0.01). The accentuated curve in Fig. 3a represents patient no. 21, who suffered from extensive disease dissemination and had clinical benefit from the treatment.

Combined phase I and II data

This phase II study was carried out in direct continuation of the previously published phase I study [20] using exactly the same inclusion criteria and vaccination regime according to a common phase I/II protocol. Six patients were treated in the phase I part; thus, a total of 32 patients were treated in all and clinical outcome included 11 SD (of these, three patients had minor regression), 20 PD, and 1 mixed response (MR). Statistical analysis of the correlation between clinical outcome and p53 expression were applied to the combined data. In total, 30 patients were evaluable for p53 expression; 7/9 patients with SD were p53 positive compared to 7/21 patients with PD (P = 0.046). Thus, 7/14 (50%) p53 positive patients obtained SD while only 2/16 (12.5%) of p53 negative patients, reflecting a significant difference with an increased chance of treatment benefit for p53 positive patients.

Combined phase I and II data was also evaluated for correlation between clinical outcome and induced immune response; there was a trend towards a correlation between SD and induced p53 immune reactivity (P = 0.10) but no significance was reached.

Discussion

This phase II study was carried out in continuation of a recently published phase I study [20] in which six patients with progressive advanced breast cancer were treated. Immunization with p53-peptide pulsed autologous DC was found to be safe and not associated with significant toxicity. Therapy associated expansion of p53 specific peripheral blood T-cells was demonstrated and, importantly, indications of a possible anti-tumor activity of the treatment were found. Now, additional 26 patients have been treated in the phase II part of the study. Unfortunately, a significant fraction of patients did not reach the evaluation time point due to disease deterioration within the first few weeks of treatment, which emphasizes the difficulties in conducting therapeutic vaccination trials in end stage cancer patients. Theoretically, when using tumor antigen immunization strategies such as ours, a time span of at least 6–8 weeks is necessary to establish a CTL response with a potential clinical capacity; furthermore, immunocompetent patients are a prerequisite.

No objective responses according to the RECIST criteria were achieved; however, nearly 1/3 of the patients, all having verified progressive disease at time of treatment initiation, attained disease stabilization during therapy. Two patients had minor regression of single lesions; and in two patients, one with nodal and the other with lung metastasis, prolonged stabilization was observed. Patients within the SD group survived significantly longer; however, this might be due to less wide spread disease at time of inclusion. Also, more patients in the SD group were chemo-naïve, which sustain the assumption that immunization therapy is more likely to succeed in less pretreated patients.

Hence, in the present study a significant fraction of the breast cancer patients experienced clinical benefit in association with the DC-p53 vaccination treatment, suggesting that vaccine-induced immunity may have therapeutic effect in some patients with defined tumor characteristics. To support this observation, it was important to document a p53 dependency of the observed clinical responses.

Wild-type p53 is only present in low levels in normal cells whereas mutant p53 has a prolonged half-life causing an accumulation of p53 protein. This differential level of p53 expression potentially allows the immune system to discriminate between normal and malignant cells and provides the basis for p53 targeting immunotherapy. Tumor p53 over-expression was however not used as inclusion criteria, but p53 expression of the primary tumor was measured by immunohistochemistry allowing us, retrospectively, to evaluate the influence of tumor p53 status on clinical outcome. Phases I and II data were combined to obtain a larger patient material, and we were thereby able to demonstrate that significantly more SD patients than PD patients had p53 over-expression in their primary tumor samples; a correlation which crucially sustains the p53 dependency of this p53 targeted immunotherapy. So, in spite of the fact that p53 over-expression is in general a predictor of poor prognosis and therapeutic response [38], these results indicate that p53 expression might be advantageous when therapy is targeting p53 specifically. Noticeably, if only p53+ patients had been included in this trial, it would have resulted in SD rates of up till 50%.

Tumor cells might, however, lose antigens and therefore display a reduced susceptibility to vaccine-induced immunity in the course of vaccinations. The use of multi-epitope vaccine strategies are a potential way to avoid this phenomenon; therefore, in combination with p53, we have introduced survivin [39, 40] and telomerase [41] as additional targets. In an ongoing trial, breast cancer patients are immunized with autologous DC loaded with up to 26 different HLA epitopes, including p53, survivin, and telomerase.

Despite the fact that p53 is an obvious candidate for immunotherapeutic strategies, p53 specific immunization has only been tested in very few clinical trials [4245] and, to our knowledge, this is the first time that clinical and immunologic efficacy of p53 peptide loaded DCs have been tested in cancer patients.

In one previous p53 vaccination study, the safety and efficacy of administration of a canarypox virus encoding the human wild-type p53 gene (ALVAC-p53) was evaluated in colorectal cancer patients [44, 45]. In another recent study, DC infected with an adenoviral construct containing the full length wild-type p53 was tested for treatment of patients with extensive small lung cancer [42]. Here, p53-specific T cell responses against ALVAC-p53 were observed in about half of all patients using IFN-γ ELISPOT, and against the L9V HLA-A*0201 binding p53 peptide in more than half of the HLA-A2+ patients. We were able to detect measurable induction of p53 peptide specific immunity by IFN-γ ELISPOT in one-third of the immunized patients; possibly, this lower frequency can be explained by differences in immunization concept. Utilizing DCs pulsed with predefined p53 peptides offers the possibility of more easily following the induced immunity and evaluate the immunological and clinical efficacy of these specified epitopes. Furthermore, the peptide immunization strategy circumvents the use of viral vectors, which holds the intrinsic problem of potential interference with adjuvant components due to antiviral immune reactivity. However, both antigen presentation spectrum and probably also presentation time are more limited in DC-peptide based vaccines.

The identification of class I binding peptides does not ensure that the epitopes are also processed by e.g. tumour cells. For the peptides used in this study, such processing and sensitivity of target cells to CTL killing were tested in preclinical studies [1719, 32, 33]. Similarly, induced immune responses against the three modified p53 peptides, with improved HLA class I binding, are directed against non self peptides per se, and it is therefore of importance to test cross-reactivity against wild-type peptides. Additional experiments confirmed that a parallel induction of CTL reactivity against modified and corresponding unmodified peptides could be attained. These results have important implications for the generation of anti-tumor reactivity against self epitopes as they indicate that the use of anchor-modified p53 peptides with increased affinity for HLA class I in cancer vaccines is capable of expanding CTL responses against the corresponding wild-type peptide epitopes.

A higher fraction of SD than PD patients had measurable induction of immune reactivity and also against a higher number of p53 peptides; furthermore, decrease of p53 reactivity was only observed in PD patients. Even though statistic significance was not reached, these observation are central as they point toward a possible connection between immune response and clinical outcome. Interestingly, in several patients with SD, we observed fluctuations in the number of p53 peptide specific CTLs during vaccination with low CTL levels at certain time points and decline at late time points. Since repeated IL-2 treatments have been shown to shift tumor specific T-cells from the blood to the tumor site [46], our data might suggest that p53 peptide specific T-cells leave the blood stream. To the contrary, induction of tolerance due to repeated injections of the antigen can not be ruled out [47], but does not seem likely as a reduction to background level is only seen in a few patients; further studies investigating the vaccine associated induction of p53 specific regulatory T cells are ongoing in order to clarify this.

Identification of response biomarkers are important to facilitate the identification of patient subgroups responding to treatment. In immunotherapy response, biomarkers is a particular important issue as objective responses frequently are lacking and clinical benefit often appears in the form of induced or prolonged stable disease with variable correlation to a measurable immune response.

The protein YKL-40 (also named CHI3L1) is a 40 kDa glycoprotein and a member of “mammalian chitinase-like proteins”. It is expressed by non-malignant cells such as macrophages and neutrophils and by several different types of malignant cells including breast cancer cells. The biological function of YKL-40 is partially unknown, but it is believed to be involved in inflammation and remodeling of the extracellular matrix through growth factor activity. Furthermore, several studies have indicated a role for YKL-40 in cancer cell proliferation and invasiveness [22]. A number of studies have suggested that one single measurement of serum YKL-40 in cancer patients at time of diagnosis, or at time of relapse, is a potential independent biomarker of short survival [22]. Three studies have suggested that serum YKL-40 may be a useful biomarker to monitor in cancer patients [4850].

The cytokine IL-6 is produced by both normal cells (e.g. macrophages, lymphocytes, and endothelia cells) and cancer cells. It plays a major role as immune modulator but also acts as a paracrine and autocrine growth factor for cancer cells and inhibits radio- and chemotherapy induced apoptosis of cancer cells [51]. The serum IL-6 concentration is increased in breast cancer patients and in patients with other kinds of solid tumors and is related to disease stage. In breast cancer patients, a high serum level of IL-6 is associated with shorter survival [27, 52]; only a few studies with IL-6 as response marker exist [53, 54].

In addition to LDH and ALP, we decided to measure serum IL-6 and YKL-40 during treatment to test their feasibility as markers of response to DC therapy. We found significant response related changes in the serum disease markers LDH; ALP as SD was associated with cessation of serum LDH and ALP rise. Serum YKL-40 and IL-6 changes were also significantly related to clinical outcome, as the YKL-40 level remained stable in SD patients, who also displayed the highest relative decline in IL-6 levels. These results imply that the role of serum YKL-40 and IL-6 as biomarkers of response in connection with DC based cancer immunotherapy should be further explored.

The phenotype and functionality of DCs employed in vaccination trials are believed to be decisive for the immunological and thereby the clinical outcome of the treatment [55]. To facilitate the ongoing optimization of clinical DC preparation procedures, it is therefore mandatory to collect relevant quality assessment data. The procedures for generation of DCs from leukapheresis separated PBMC used in the present work resemble those used by other authors [56, 57]. Large granular cells comprised around 60% of the cells in the final therapeutic DC preparations and the cells displayed a phenotype characteristic for intermediate-to-mature DCs, even though no maturation cytokine cocktail was employed. In a recently published study, we tested more thoroughly the function of DCs from breast cancer patients included in vaccination trials [29]. Here, we found an unimpaired capacity of the patient-derived DCs for cross-presentation of naïve (KLH) as well as recall (CMV and Tetanus) antigens.

Only a few human cancer vaccine studies have been made that directly compare immature and mature DCs. However, these studies clearly demonstrate that in vitro DC maturation is optimal for induction of a potent T cell response, also in immunotherapeutic settings [56, 58]. Phase I of our clinical trial was initiated in 2001 when no consensus regarding maturation of DCs for clinical applications existed; therefore, the protocol was planned without in vitro DC maturation. To maintain a homogenous treatment of all included patients, this strategy was not changed during phase II. However, immature DCs are also able to induce peptide specific CD8+ T cells [59] and the DCs used in our study induced significant levels of CTL activity despite the lack of in vitro maturation. Protocols that apply immature DC might not induce tolerance as feared, simply because the in vitro generation and manipulation of the DCs induce some degree of maturation. In addition, immature DCs pulsed with MHC class II binding epitopes, like PADRE employed in this protocol [30], may induce conditioning of the DC in vivo.

In conclusion, a significant fraction of the breast cancer patients obtained disease stabilization during p53-DC vaccination and correlation to tumor p53 expression, induction of p53 specific immunity, and observed changes in biomarkers supported the clinical results. All patients in this trial had metastatic breast cancer often with a high tumor burden, and the patients were frequently heavily pre-treated. As a consequence, transformation of a p53 specific activation of the immune system into significant tumor regression might not be obtainable in these patients. More relevant could be the ability of the treatment to induce prolonged survival, a question this study was not scaled to answer. However, the clinical results obtained here encourage further clinical studies at an earlier stage of the disease with progression free survival as endpoint.

Acknowledgment

This work was supported by grants from Dansk Kræftforsknings Fond, The Danish Cancer Society, Direktør Leo Nielsen og Hustru Karen Margrethe Nielsens Legat for Lægevidenskabelig Grundforskning, Michaelsen Fonden, and Aase og Ejnar Danielsens Fond.

Copyright information

© Springer-Verlag 2007