Overcoming Immune Tolerance Against Multiple Myeloma With ...

2 downloads 7017 Views 637KB Size Report
University of Florida College of Medicine, Gainesville, Florida, USA; 3Department of Medicine, ... genetically engineering cancer patients' DCs may improve.
original article

© The American Society of Gene Therapy

Overcoming Immune Tolerance Against Multiple Myeloma With Lentiviral Calnexin-engineered Dendritic Cells Shuhong Han1, Bei Wang1, Matthew J Cotter1, Li-Jun Yang2, James Zucali3, Jan S Moreb3 and Lung-Ji Chang1 Department of Molecular Genetics and Microbiology, University of Florida College of Medicine, Gainesville, Florida, USA; 2Department of Pathology, University of Florida College of Medicine, Gainesville, Florida, USA; 3Department of Medicine, University of Florida College of Medicine, Gainesville, Florida, USA 1

The key to successful cancer immunotherapy is to induce an effective anticancer immunity that will overcome the acquired cancer-specific immune tolerance. In this study, we found that dendritic cells (DCs) from multiple myeloma (MM) patients suppressed rather than induced a cancer cell-specific immune response. We demonstrated that CD4+CD25high T cells from MM patients suppressed the ­ proliferation of activated peripheral blood lymphocytes. Further analysis illustrated that MM cell lysates or MM-specific idiotype immunoglobulins (MM Id-Ig) specifically induced the expansion of peripheral CD4+CD25highFoxP3high T regulatory (Treg) cells in vitro. Supraphysiological expression of calnexin (CNX) using lentiviral (LV) vectors in DCs of MM patients overcame the immune suppression and enhanced MM-specific CD4 and CD8 T-cell responses. However, overexpression of CNX did not affect the peripheral expansion of Treg cells stimulated by MM antigens. Thus, the immune suppression effect of Treg cells in cancer patients may be overcome by ­ improving antigen processing in DCs, which in turn may lower the activation threshold of the immune effector cells. This concept of ­modulating anticancer immunity by genetically engineering ­cancer patients’ DCs may improve immunotherapeutic ­regimens in ­cancer treatment. Received 14 March 2007; accepted 29 October 2007; published online 11 December 2007. doi:10.1038/sj.mt.6300369

introduction Multiple myeloma (MM), an incurable B-cell neoplasm, accounts for ~1% of all cancers but it is the second most common hematologic malignancy after lymphoma. The standard high-dose chemotherapy and autologous transplant used in MM treatment only offer limited control of the disease.1,2 A potentially effective strategy for curing MM is allogeneic stem cell transplant. However, allogeneic stem cell transplant is associated with high treatmentrelated mortality and limited to individuals with HLA-matched

donors. Furthermore, myeloma cells may escape from the graft versus myeloma effect in a large number of patients undergoing this treatment. Consequently, most patients eventually die from recurrent disease.3–5 Further studies are urgently needed to overcome these obstacles and lead to successful and effective immunotherapy for myeloma. Immunotherapy is considered an alternative means for treating MM and has been the focus of multiple studies.6–8 Dendritic cells (DCs) are a key player in establishing an effective anticancer immunity.9,10 The DC immunization approach for cancer therapy has gained limited success mainly because many patients have established strong immune tolerance toward their cancer cells. Much research is now focused on breaking the tolerance established in cancer patients.7,11 Recent evidence suggests that DCs in cancer patients are tolerized early in their development.11–13 Therefore, it may be difficult to alter the functions of DCs or their precursors by means of extracellular adjuvants such as cytokines and inflammatory mediators.14 However, DCs or their precursors may be genetically engineered to alter their T-cell stimulatory functions and re-direct immune reactivity. We have previously reported that DCs, modified by means of lentiviral (LV) vectors to express exogenous cytokines, costimulatory molecules or small interfering RNAs, can enhance T-cell responses.15 Others have shown that chaperones in the cytosol such as heat-shock proteins, calreticulin, gp96, tapsin, ER60 and calnexin (CNX) could enhance antigen-specific cellular immunity.16–18 Overexpression of CNX in DCs upregulates class I major histocompatibility complex (MHC-I) and promotes central memory T-cell response (L.-J. Chang and B. Wang, unpublished results). Thus, the activation of specific anticancer immunity using genetically modified DCs from cancer patients is a promising new immunotherapeutic modality but has not yet been clinically tested. A growing body of evidence now indicates that cancer cells promote T regulatory (Treg) cell trafficking, differentiation, and expansion, and the Treg cells may suppress the activities of ­cancer-specific T cells and natural killer cells.19–23 In MM patients, the number of CD4+, CD25high, FoxP3+ Treg cells is significantly increased and coincides with the progression of malignant transformation.24 On the other

Correspondence: Lung-Ji Chang, 1600 SW Archer Road, ARB, R1-252, University of Florida, Gainesville, Florida 32610-0266, USA. E-mail: [email protected] and Jan S Moreb, 1600 SW Archer Road, University of Florida, Gainesville, Florida 32610-0266, USA. E-mail: [email protected] Molecular Therapy vol. 16 no. 2, 269–279 feb. 2008

269

© The American Society of Gene Therapy

Overcoming Cancer Immune Tolerance

hand, the Treg cells in MM have been reported to be dysfunctional and unable to suppress anti-CD3 activated T-cell proliferation.25 The specific MM antigens that mediate the Treg cell increase have not been characterized. It is also still controversial whether the Treg cells are immunosuppressive, and whether the expansion of the Treg cells is a central or peripheral lymphoid event in MM patients.26,27 In this report, we analyzed CD4+CD25+Treg and DC functions in MM patients and provided evidence that CD4+CD25+Treg had immunosuppressive capacity and that both MM cell lysates and MM-specific idiotype immunoglobulin (MM Id-Ig) loaded DCs triggered a suppressive antimyeloma immune response exemplified by the expansion of peripheral Treg cells. These tolerizing DC functions, however, could be overcome by LV-mediated expression of the CNX gene in the patient’s DCs, which, in turn, lead to an anti-MM response by the effector T cells. We believe that this is the first study to report that engineered DCs from MM patients can overcome peripheral Treg cell-induced immune tolerance.

non-adherent peripheral blood mononuclear cells (PBMCs) as illustrated in Figure 1a. The MM cells were highly enriched (96–98%) as shown by using antibody (Ab)-conjugated magnetic beads specific to MM phenotype: CD38+CD138+CD56+CD45– (Figure 1b). To investigate whether DCs from MM patients were immunologically different, the phenotypes of mature DCs from five healthy donors (HDs) and five MM patients (MM) were compared. Figure 1c illustrates the results from CD11c+ DC phenotypes from HD and MM patients. Although minor variations existed, there was no significant difference in the surface phenotype between these two groups.

Expansion of CD4+CD25highFoxP3high Treg cells by MM cell lysate–pulsed DCs DCs of MM patients were pulsed with lysates derived from autologous MM cells (DC/MM), autologous normal PBMCs (DC/PBMC) or an allogeneic Epstein-Barr virus–transformed B-cell line (DC/BCL). Antigen internalization was confirmed by exposure of immature DCs to B-cell lymphoma (BCL) and double-staining for CD11c and immunoglobulin light chains. The κ or λ antigens of BCL were efficiently internalized by DCs (4–13.1%) as detected by flow cytometry (Figure 2a and Supplementary Figure S4). We transduced immature DCs

Results DC phenotype analysis in healthy donors and MM patients For immune analysis, we pulsed MM patients’ monocyte-derived DCs with MM cell lysates and cocultured them with autologous

a

c

BM cells (MM patient)

HD1

Negative selection MM cells

HD2

MM cell lysate

375

91

HD3

DC

MM1

PBMC minus DC

Functional and phenotype analysis

500

600

500

200

250

300

250

600

500

250

200

125

750

100 10

2

10

3

10

375

0 10 100 101 102 103 800 4

77

600

31

1000

500

250

250

0 10 100 101 102 103 800 600

76

18

750

500

4

92

0 100 101 102 103

104

0 104 100 101 102 103 1500 1125

0 104 100 101 102 103 1200

1

900

0 10 100 101 102 103 800 4

18

600

250

400

400

400

400

400

125

200

200

200

200

200

750

0 104 100 101 102 103 600

96

450

0 104 100 101 102 103 600

62

450

0 104 100 101 102 103 800

68

600

62

0 104 100 101 102 103 600

0 104 100 101 102 103 800

24

450

600

2

750

300

300

400

300

400

500

250

150

150

200

150

200

250

900

98

600

103

0 104 100 101 102 103 1200

82

900

0 104 100 101 102 103 1000

85

750

67

0 104 100 101 102 103 800

0 104 100 101 102 103 1000

45

600

750

0

900

400

600

500

400

500

600

300

200

300

250

200

250

300

MM3

900

0 100 101 102 103

0 100 101 102 103

104

1000

82

750

2000

84

1500

104

0 100 101 102 103

1700

73

1250

0 100 101 102 103

104

0 100 101 102 103

1800

1500

36

104

54

1125

1350

600

500

1000

900

750

900

300

250

500

450

375

450

104

750

0 100 101 102 103

1000

1

750

104

0 100 101 102 103

1000

0

750

104

0 100 101 102 103

1000

1

750

0

104 100 101 102 103

1

104

3

750

750

500

500

500

500

500

500

250

250

250

250

250

250

0 100 101 102 103

10

4

0 100 101 102 103

4

10

0 100 101 102 103

10

4

0 100 101 102 103

10

4

0

10

0

1

10

10

2

10

3

4

10

104

0 100 101 102 103 1200

90

600 300 0 104 100 101 102 103 1000

96

88

500

0 0

104 10 300

1

10

0 104 10 1000 750

0

101 102 103

250 0

0

100 101 102 103

104

100 101 102 103

HLA-DR

CD1a

CD86

CD40

CD83

CD14

HLA-I

CD80

HD

89 ±11.6

76 ±0.6

77 ±18.3

52 ±22.7

21 ±5.8

0.3 ±0.6

92 ±4.9

54 ±2.9

MM

92 ±8.7

76 ±12.1

75 ±8.7

55 ±16.6

41 ±15.3

0.6 ±1.1

91 ±4.2

66 ±10

96

104

1

500

250 4

104

75

500

0 100 101 102 103

101 102 103

62

225 150

0 100 101 102 103

750

104

77

750

250

1000

0

104

68

900

75 104

104

57

750

250

150

0 100 101 102 103

1000

1000

225

0 100 101 102 103 1000

500

0 104 100 101 102 103 300

0

10

4

95

0 0 1 2 3 104 10 10 10 10 1200

600

104

300

0 104 100 101 102 103 1000

500

0 104 100 101 102 800

600

0 10 100 101 102 103 800 4

0

600

104

52

900

300

375

600

0 104 100 101 102 103 1200

86

600

750

0 10 100 101 102 103 800 4

250

Isotype

CD45

750

250 1

500

99

0 104 100 101 102 103 1000

57

500

98 CD138

CD56

750 500

0

104

0 104 100 101 102 103 1000

76

375

400

64

CD38

0 100 101 102 103

104

76

52

750

400

1200

After selection

94

200

Before selection 98

900

400

0 100 101 102 103

69

1000

1200

0

750

CD80

HLA-I

CD14 1000

28

600

200

1000

68

CD83 800

49

400

0 100 101 102 103

b

600

75

MM2

14 days

800

84

150

0 100 101 102 103 1200

Coculture

600

125

0 100 101 102 103 1000

DC/MM

CD40

CD86 800

76

225

250

0 100 101 102 103 1000

Pulse (4 h)

20:1

CD1a 300

0 100 101 102 103 800

Freeze/thaw

PBMC (MM patient)

HLA-DR 500

104

Figure 1  Analysis of functional immunity and surface phenotype of dendritic cells (DCs). (a) Experimental diagram of immunity analysis. (b) Flow cytometry analysis of purified CD38+CD138+CD56+CD45− multiple myeloma (MM) cells. The purified MM cells were analyzed by surface staining for CD38, CD138, CD56, and CD45, before and after magnetic bead selection. The numbers denote percentages of the specific cell population. The results are representative of five MM patients. (c) Comparison of DC surface phenotype between healthy donors (HD) and MM patients (MM). The surface phenotype of mature DCs was analyzed with fluorochrome-conjugated antibodies against CD11c, CD1a, CD83, CD80, HLA-ABC, CD86, CD40 and HLA-DR as indicated. The percentages of positive cells against isotype Ab controls are summarized in the bottom. PBMC, peripheral blood mononuclear cell.

270

www.moleculartherapy.org vol. 16 no. 2 feb. 2008

© The American Society of Gene Therapy

Mock

104

0.4



103

BCL1

104 103

4.0

103

102

102

102

101

101

101

c

BCL2

104

10.8

CD25

a

Overcoming Cancer Immune Tolerance

0.5



BCL3

104

BCL4

104

6.0

103

13.1

103

102

101

101

101

101

100 101

101

101

100 102 103 104 100

MM2

104

12.5

104

103

103

103

102

102

102

102

102

101

101

101

101

101

Isotype 0.0

CD4

101 102 103 104

15 10 5 0

HD

MM

d CD8+

CD25

CD4+

DC/MM

12.2

103

DC-LV-LacZ/MM DC-LV-LacZ

MM1

104

103

CD11c

b

9.7

14.2

103

102

100 102 103 104 100

HD2

BCL5

104

102

100 102 103 104 100

104

20

102

100

5.7

CD4+ CD25high%

Mock

103

HD1

100 100 100 100 100 100 101 102 103 104 100 101 102 103 104100 101 102 103 104 100 101 102 103 104 100 101 102 103 104

100 100 100 100 101 102 103 104 100 101 102 103 104 100 101 102 103 104 104

104

*

*

*

*

CD4

DC alone

DC/PBMC

2.64

DC/MM

2.83

DC/BCL 8.09

2.95

Counts

DC/BCL DC/PBMC 0

2

4

6

8

10

0

1

IFN�+/CD4+T cells (%)

2

3

4

5

FOXP3

IFN�+/CD8+T cells (%)

DC/BCL DC-LV-LacZ/MM DC-LV-LacZ DC/MM

*

*

*

*

DC/MM

**

DC/PBMC

DC/BCL DC alone

DC/PBMC 0

1

2

3

4

5

6

7

IFN�+/CD4+T cells (%)

8

0

1

2

3

4

5

IFN�+/CD8+T cells (%)

6

0

2

4

6

8

10

FOXP3high/CD4+CD25highT cells (%)

Figure 2  Multiple myeloma (MM) cell lysate–pulsed dendritic cells (DCs) impair T cell response and increase the frequency of CD4+CD25highFoxP3high Treg cells. (a) Internalization of cancer cell lysate immunoglobulins by DCs. Immature DCs were pulsed with cancer cell lysates from five different donors (BCL-1 to 5) for 3 hours. Antigen internalization in DCs was evaluated by double antibody (Ab) staining for CD11c and cytoplasmic BCL-specific immunoglobulins (κ and λ light chains). (b) Intracellular cytokine staining (ICCS) of DC-activated T cells. After restimulation with phorbol myristate acetate and ionomycin, CD4+ or CD8+T cells secreting tumor necrosis factor-α (TNF-α) and interferon-γ (IFN-γ) was detected by ICCS. (c) Relative levels of peripheral blood CD4+CD25high T cells in MM patients versus healthy donors (HD). Peripheral blood mononuclear cells (PBMCs) were incubated with anti-CD4 and anti-CD25 Abs and analyzed by flow cytometry. Lymphocytes were gated according to forward scatter and side scatter, and CD4+ T cells were gated for further analysis. Representative flow cytometry analyses of two MM patients and two HD are shown and total sample distributions presented underneath. (d) Expansion of Treg cells upon exposure to MM cell lysates. After 13 days in coculture, the frequencies of CD4+CD25highFoxP3high Treg cells were analyzed with Abs against CD4, CD25, and FoxP3. The CD4+CD25high T cells were gated for FoxP3 analysis. Representative triplicate assays from four MM patients’ specimens are shown (*P < 0.05 and **P < 0.001). BCL, B-cell lymphoma.

with LV-LacZ expressing a highly ­ immunogenic bacterial β-galactosidase protein for 16–24 hours and induced them into maturation (DC-LV-LacZ). Similarly treated DCs were exposed to MM cell lysates (DC-LV-LacZ/MM). The DCs were cocultured with non-adherent autologous PBMCs at a ratio of 1:20 for 14 days and TNF-α− and IFN-γ−producing CD4 and CD8 T cells were detected by intracellular cytokine staining (ICCS) after stimulation with phorbol myristate acetate and ionomycin. We noted that the MM cell lysates (DC/MM) were not as immunogenic as the BCL lysates (DC/BCL), and furthermore, T cell activation by DC/LV-LacZ cells was significantly suppressed after exposure to MM cell lysates (Figure 2b, DC/LV-LacZ ­versus DC-LV-LacZ/MM). Molecular Therapy vol. 16 no. 2 feb. 2008

It is known that during cancer progression, patients may develop strong Treg activities. We examined ­ Treg-related CD4+CD25high lymphocytes in the peripheral blood of MM patients and HD. MM patients showed elevated levels of CD4+CD25high lymphocytes (mean 11.51 ± 1.716%) in the peripheral blood than did HD (mean 6.57 ± 1.082%; P = 0.03, Figure 2c and Supplementary Figure S2). To further examine Treg cells, the expression of FoxP3, an important transcription factor associated with Treg cells, was analyzed. Figure 2d illustrates a significant increase in gated CD4+ T cells with CD25high FoxP3high phenotype when the cells encountered MM cell lysates (DC/MM, mean fluorescence index 8.09 versus 2.64, 2.83, and 2.95 of DC alone, DC/PBMC and DC/BCL, respectively). Thus, 271

© The American Society of Gene Therapy

Overcoming Cancer Immune Tolerance

these results suggest that the MM cells selectively induced CD4+CD25highFoxP3high Treg cell expansion.

Isolation and expression of MM Id-Ig gene The activation of Treg cells by tumor cell lysate–pulsed DCs, may be induced by multiple cancer-related antigens. To see if this effect can be induced by MM Id-Ig antigens, we cloned MM Id-Ig genes for further investigation. The MM cells of a κ-chain specific patient (MM3) were fluorescence-activated cell sorting (FACS)-sorted (CD38+CD138+CD56+CD45−, right panel, Figure 3a), and the RNA was harvested for complementary DNA (cDNA) synthesis. We used specific 5′ V-region and common 3′ C-region primers to amplify the MM κ gene (Figure 3a). After polymerase chain reaction amplification, one of the six primer pairs (Figure 3b, circled) revealed a discordant pattern between MM+ (purified MM cells) and MM− cDNAs (MM-minus BM cells). cDNA analysis revealed a consensus CDR3 sequence as underlined in Figure 3c, resulting from clonal plasma cells. To confirm this, an oligo-primer, specific

a

for the CDR3 sequence of the MM3 patient, was used to amplify cDNAs of different MM cells (Figure 4a). In separate studies, we have noted that MM patients’ BM stromal cells continue to express a high level of MM-­specific surface markers (unpublished results). Therefore, we used both BM cells and BM stromal cells for this analysis. A positive band was amplified from the corresponding MM3 patient’s BM cDNA (MM3 BM cells, L1, Figure 4b) as well as the corresponding BM stromal cell cDNA (MM3 stromal cells, L2), but not from a ­different MM patient’s (MM4) BM stromal cell cDNA (MM4 ­stromal cells, L3). To express myeloma-specific Id-Ig, we cloned the MM κ cDNA into an LV vector (pTYF-EF) under the control of a strong elongation factor 1α promoter. The cDNA was fused with an N-­terminal Flag tag (pTYF-EF-κ-flag), and the expression confirmed by Western analysis using an anti-Flag Ab (Figure 4c, with an internal expression control of α-tubulin). Efficient transduction of DCs with LV vectors is illustrated in Figure 4d; up to 40% of DCs were transduced with a reporter LV-eGFP at a multiplicity of infection of 10.

SSC

BM cells (MM) patient FACS sort

FSC BM cells (minus MM cells) CD138

MM cells (CD38+CD138+CD56+CD45−)

mRNA

CD38

RT-PCR 5′

FR1 CDR1 FR2 CDR2 FR3

CDR3

FR4

C

3′

CD56

1 2 3 4 5 6

c

CD45

b Purified MM cells (CD38+CD138+CD56+CD45–) 1

2

3

4

5

6

BM cells (minus MM cells) 1

2

3

4

5

6

K GAPDH

Figure 3 The cloning strategy and sequencing of multiple myeloma (MM)-specific Id-Ig complementary DNA (cDNA). (a) The strategy of cloning MM Id-Ig cDNA. In this representative MM patient, bone marrow (BM) cells were sorted by flow cytometry based on specific MM surface markers. The RNA was isolated and reverse transcribed into cDNA, and amplified with six pairs of primers specific for the known κ light chain families associated with the specific MM patients. (b) PCR identification of MM-specific κ chain cDNA. (c) The MM κ chain cDNA sequence (700 base pair with estimated MW of 23 kd). The CDR3 region of the MM κ gene is underlined. FACS, fluorescence-activated cell sorting; FSC, forward scatter; Id, idiotype; Ig, immunoglobulin; mRNA, messenger RNA; RT-PCR, reverse transcriptase polymerase chain reaction; SSC, side scatter.

272

www.moleculartherapy.org vol. 16 no. 2 feb. 2008

© The American Society of Gene Therapy

Overcoming Cancer Immune Tolerance

BM cells (MM3) Stromal cells from BM cells (MM3) Stromal cells from BM cells (MM4)

a

c

pTYF-EF-�-flag CMV-TAR EF1�

�-flag

SIN-LTR

cPPT mRNA

293T -�-Flag

293T

27 kd

Anti-flag

54 kd

Anti-tubulin

RT-PCR 5� FR1 CDR1 FR2 CDR2 FR3 CDR3 FR4

L1

L2

3�

d

L3 �

MOI:

0

Transduction efficiency of DC with LV vector 10 20 40 80 0.5 41.7 44.2 53.7 47.0

GFP

b

C

GADPH CD11c

Figure 4 Expression of multiple myeloma (MM) Id-Ig gene with lentiviral (LV) vectors. (a) Polymerase chain reaction (PCR) confirmation of MM Id-Ig gene. The Id-Ig gene isolated from an MM patient (MM3) was confirmed by PCR using a sequence-specific CDR3 primer. A different MM patient’s bone marrow (BM) cells were included as control (MM4). (b) Specific PCR amplification of the MM Id-Ig gene. L1, DNA from MM3 patient’s BM cells; L2, DNA from MM3 patient’s BM-derived stromal cells; L3, DNA from the control MM patient’s (MM4) BM-derived stromal cells. (c) LV expression of the MM κ-Flag fusion protein. The self-inactivating (SIN) LV vector construct (pTYF-EF-k-flag) was used to transduce 293T cells. The expression of the κ-Flag fusion protein (700 base pair complimentary DNA with estimated MW of 27 kd) was confirmed by Western analysis using an anti-Flag antibody (Ab). Anti-α−tubulin Ab was included as control. (d) LV-eGFP transduction of dendritic cells (DCs) and analysis of green fluorescent protein (GFP) expression. On Day 5 immature DCs were transduced with LV-eGFP at different multiplicity of infections (MOIs of 10, 20, 40, and 80) and 48 hours later, the expression of GFP was analyzed by flow cytometry. cPPT, central polypurine tract; EF1α, elongation factor 1α; eGFP, enhanced green fluorescent protein; GADPH, glyceraldehyde-3-phosphate dehydrogenase; Id, idiotype; Ig, immunoglobulin.

MM Id-Ig displays low immunogenicity but induces a specific Treg-cell response To assess the immunogenicity of the MM Id-Ig, we transduced ­autologous immature DCs with the MM Id-Ig LV-κ vector (DCLV-κ) or a control LV-LacZ vector (DC-LV-LacZ). As positive control, immature DCs were pulsed with memory antigen tetanus toxoid (TT) for 4 hours before maturation (DC/TT). The DCs were cocultured with autologous T cells for 14 days and immune effector function was examined. Background response to the nontransduced DCs was subtracted. For HD there was no difference between DC-LV-κ and DC-LV-lacZ in the CD4 and CD8 T-cell response (Figure 5a). In contrast, DC-LV-κ induced less CD4 and CD8 T-cell response than did DC-LV-LacZ as illustrated by intracellular analysis of TNF-α and IFN-γ (Figure 5a). Further analysis of CD4+CD25+ and FoxP3 Treg cells showed that DCLV-κ, but not DC-LV-LacZ or DC/TT, markedly upregulated FoxP3 expression in CD4+CD25high T cells in the corresponding MM patients (mean fluorescence index of 16.85 for DC-LV-κ, versus 8.23 and 8.21 for DC-LV-LacZ and DC/TT, respectively, P < 0.001). By contrast, no significant difference was found for HD (Figure 5b and c). It has been reported that CD4+CD25+ Treg cells of MM patients are dysfunctional.27 We evaluated CD4+CD25+ versus CD4+CD25− T cells of MM patients and HD in a PBMC proliferation assay. Autologous CD4+CD25+ or CD4+CD25− T cells were cocultured with 5-(and 6)-carboxyfluorescein diacetate succinimidyl ester (CFSE)-labeled PBMCs at different ratios and activated with phytohemagglutinin. The intensity of CFSE in the culture decreases with increased cell ­ proliferation. The Molecular Therapy vol. 16 no. 2 feb. 2008

results showed that CD4+CD25+ T cells from MM patients suppressed PBMC proliferation in a dose–dependent manner (Figure 5d and e), similar to that of HD (not shown). The control CD4+CD25− T cells, by contrast, did not show such an effect.

LV-CNX-transduced DCs enhance the MM-specific CD4 and CD8 T-cell response Both MM cell lysates and the specific Id-Ig antigens failed to induce an immune effector response but instead, promoted a strong Treg-cell response. To overcome this MM-specific immune suppression, we attempted to modify the antigen presentation functions of MM patients’ DCs. CNX is a chaperone in the endoplasmic reticulum critical to the processing of glycoproteins and has been shown to promote antigen presentation in immune cells.28,29 We explored the effect of CNX on MM patients’ DCs and T cells by LV-­mediated overexpression of CNX in patients’ DCs. The human CNX cDNA was cloned into LV vector (LVCNX) behind a strong elongation factor 1α promoter. Western analysis detected high expression of CNX in CEM-NKR cells (a CNX-defective cell line) after LV-CNX transduction (Figure 6a). Upregulation of CNX expression was also confirmed in DCs when transduced with LV-CNX (not shown). To assess the effect of CNX, immature DCs from MM patients were transduced with LV-CNX (DC-LV-CNX/MM), and control DCs were transduced with LV-LacZ (DC-LV-LacZ/MM), and then pulsed with MM lysates. After maturation, the DCs were cocultured with autologous PBMCs at a ratio of 1:20 for 14 days. The IFN-γ and TNF-α response of the T cells was analyzed upon re-stimulation 273

© The American Society of Gene Therapy

Overcoming Cancer Immune Tolerance

b a

CD25

DC/TT DC-LV-� DC-LV-LacZ

HD

c CD4

MM

**

0.0

2.5 5.0 7.5 % TNF�+ CD4 T cells

DC/TT DC-LV-� DC-LV-LacZ

HD

* 10.0

0

1 2 % TNF�+ CD8 T cells

DC-LV-LacZ

3

DC-LV-�

4.36

DC/TT

4.82

4.52 MM

**

HD HD 0 8.23 MM

* 0

5

0.0

0.3 0.6 0.9 % IFN�+ CD8 T cells

12

16

20

FOXP3 Ratio of PBMC:CD4+CD25+ (or CD4+CD25–) T cells

CD4+CD25+

104

PBMC 3 20

820

99

103

615

10

2

410

101

205

0

10

102 101

58

19

100 100 101 102 103 104

100 101 102 103 104

CD4+CD25–

4

10

3 2

CD25

101

1122

10

0

10

1

10

2

10

3

10

4

1:2 25.0

0 100 101 102 103 104

1:1

1260 945

960

28.6

720

1:0.5 38.4

816 612

748

630

480

408

374

315

240

204

0 100 101 102 103 104

1137

43.1 M1

301

99

0

CD4

1496

M1

1205

603

10

1:4 15.0

0 100 101 102 103 104

904

10

10

Counts

103

8

% FOXP3high/CD4+CD25high T cells

1.2

d

104

4

8.21

MM

*

1 2 3 4 % IFN�+ CD4 T cells

16.85

853

0 100 101 102 103 104

812

699

46.7

51.4

612

569

541

439

408

284

271

220

204

0 100 101 102 103 104

0 100 101 102 103 104

0 100 101 102 103 104

e

1:4 1:2 1:1 1:0.5

CD4+CD25–

0 100 101 102 103 104 816

878

1082

43.9

0 100 101 102 103 104

1:0 42.8

42.8

CD4+CD25+

0

0 100 101 102 103 104

10 20 30 40 50 60 70 80 % Suppression

CFSE

Figure 5 Dendritic cells (DCs) transduced with LV-MM Id-Ig suppress CD4 and CD8 T-cell response and upregulate Treg response. (a) Analysis of Ag-specific effector cell response. The immature DCs from multiple myeloma (MM) patients (MMs) and healthy donors (HDs) were transduced with LV-LacZ (DC-LV-LacZ), LV-κ (DC-LV-κ) or pulsed with tetanus toxoid (TT) (DC/TT), and cocultured with autologous non-adherent peripheral blood mononuclear cells (PBMCs) for 14 days. For specific response, the T cells were re-stimulated with the same antigen-treated DCs. CD4 and CD8 T-cell secreting tumor necrosis factor-α (TNF-α) and interferon-γ (IFN-γ) were analyzed by intracellular cytokine staining. (b, c) Flow cytometry analysis of Treg cells. On day 13, the cocultured cells were stained with fluorochrome-labeled CD4, CD25, and Foxp3 antibodies. Representative of three repeated experiments is shown (*P < 0.05 and **P < 0.001). (d) Dose–dependent suppression of PBMC proliferation by CD4+CD25+ T cells. 5-(and 6)-carboxyfluorescein diacetate succinimidyl ester (CFSE)-labeled PBMCs were activated with phytohemagglutinin and incubated with CD4+CD25+ or CD4+CD25− T cells at increasing ratios as indicated. After 3 days, the proliferation of PBMCs was measured based on CFSE intensity. Representative CFSE profile from one MM patient is shown (n = 3, plus three healthy donors); M1 represents the percentage of dividing cells. (e) Bar graph analysis of suppression of PBMC proliferation (one of three assays).

with phorbol myristate acetate and ionomycin. The MM lysate– pulsed DCs again illustrated immune suppression of both CD4 and CD8 T cells (DC-LV-LacZ/MM versus DC-LV-LacZ, Figure 6b). Importantly, LV-CNX effectively reversed this trend (Figure 6b, DC-LV-CNX/MM versus DC/MM). However, analysis of the LVCNX effect on Treg cells revealed no reduction in the number of CD4+CD25highFoxP3high T cells in coculture with DC-LV-CNX/MM (data not shown).

LV-CNX enhances the MM Id-Ig-specific DC immunity The above study demonstrates that DCs from MM patients, when transduced with LV-CNX can effectively upregulate MMspecific CD4 and CD8 T-cell responses. We have observed an enhanced immune activation effect of CNX in a separated study (Supplementary Figure S3). T cells were rapidly expanded within 3 days when LV-CNX-DCs were included in the coculture (Figure 7a, clusters of expanded cells). This was verified by CFSE proliferation analysis. We pre-stained the T cells with CFSE before DC coculture. After 3 days, the CFSE intensity of the LV-CNX DC coculture group markedly decreased as compared with that of the other three groups, indicating an increased T-cell proliferation induced by LV-CNX (Figure 7a, right panel). 274

To investigate whether CNX could overcome the tolerance effect of MM-specific antigens, the MM Id-Ig gene (LV-κ) was transduced into autologous immature DCs; also included were LV-CNX (DC-LV-κ + LV-CNX), and LV-LacZ alone as control (DC-LV-LacZ). The mature DCs were cocultured with autologous non-adherent PBMCs, and antigen-specific T-cell response was examined by ICCS for IFN-γ and TNF-α. The result showed an enhanced response of CD4 and CD8 effector T cells when DCs were co-transduced with LV-CNX (Figure 7b). The number of CD4+CD25highFoxP3high Treg cells between the DC-LV-κ + LVCNX and the DC-LV-κ alone, however, was not significantly changed (not shown). To directly correlate the immune effector function with anticancer immunity, we examined the cytotoxic activity of the MMspecific T cells, using a non-radioactive target cell killing assay. The T cells were re-stimulated with the corresponding antigentreated DCs for 5 days and harvested as effector cells. Autologous stromal cells transduced with LV-κ (Stromal cells-LV-κ) or LVLacZ (Stromal cells-LV-LacZ), were used as target cells. T cells derived from the DC-LV-κ + LV-CNX coculture killed target stromal cells-LV-κ with increased activity and specificity, as compared to T cells from DC-LV-κ at effector/target ratio of 50:1 or 25:1 (P < 0.03, Figure 7c). www.moleculartherapy.org vol. 16 no. 2 feb. 2008

© The American Society of Gene Therapy

Overcoming Cancer Immune Tolerance

b

CD4+ DC-LV-CNX/MM

CD8+

*

DC-LV-CNX

*

**

**

DC/MM

C EM -N KR -L Vh

C EM -N KR

C N X

a

90 kd

hCNX

54 kd

-Tubulin

DC-LV-LacZ/MM

* *

DC-LV-LacZ DC/MM 0

2 4 6 8 IFN+/CD4+T cells (%)

* ** 10

0

1 2 3 4 IFN+/CD8+T cells (%)

5

DC-LV-CNX/MM

* *

DC-LV-CNX DC/MM

*

*

DC-LV-LacZ/MM

* **

* *

DC-LV-LacZ DC/MM 0

1 2 3 4 5 TNF+/CD4+T cells (%)

6

0

1 2 3 4 5 TNF+/CD8+T cells (%)

6

Figure 6 Exogenous calnexin (CNX) expression in dendritic cells (DCs) upregulates multiple myeloma (MM)-specific T-cell response. ­(a) Expression of LV-CNX. LV-CNX was used to transduce CEM-NKR cells, a CNX-deficient cell line. After 96 hours, the expression of CNX was verified by Western analysis. CNX and control α-tubulin were detected using specific monoclonal antibodies. (b) LV-CNX-transduced DCs enhance MM-specific T-cell immunity. MM patients’ immature DCs were transduced with LV-LacZ or LV-CNX (DC-LV-LacZ and DC-LV-CNX), and pulsed with MM cell lysates (DC-LV-LacZ/MM and DC-LV-CNX/MM). After coculture with autologous non-adherent peripheral blood mononuclear cells, the T cells were stimulated with phorbol myristate acetate and inomycin and analyzed by intracellular cytokine staining. Representative result of triplicates of four MM patients’ specimens is shown (*P < 0.05 and **P < 0.001). IFN-γ, interferon-α; TNF-γ, tumor necrosis factor-α.

To see if this cytotoxic activity could target primary MM cells, autologous MM cells were isolated from BM (MM cells), and autologous BCL cells were used as control cells. T cells from the coculture of DC-LV-κ + LV-CNX killed the MM target cells with increased activity (>40% of specific cell lysis at an effector/target ratio of 30:1) as compared to the T cells from the DC-LV-κ coculture (