Intradermal photosensitisation facilitates stimulation of MHC class-I restricted CD8 T-cell responses of co-administered antigen

Journal of Controlled Release 174 (2014) 143–150

Contents lists available at ScienceDirect

Journal of Controlled Release journal homepage: www.elsevier.com/locate/jconrel

Intradermal photosensitisation facilitates stimulation of MHC class-I restricted CD8 T-cell responses of co-administered antigen Monika Håkerud a,b,c, Ying Waeckerle-Men a, Pål Kristian Selbo b,c, Thomas M. Kündig a, Anders Høgset c, Pål Johansen a,⁎ a b c

Department of Dermatology, University Hospital Zurich, Gloriastrasse 31, 8091 Zurich, Switzerland Department of Radiation Biology, Institute for Cancer Research, Norwegian Radium Hospital, Oslo University Hospital, Oslo, Norway PCI Biotech AS, Strandveien 55, 1366 Lysaker, Norway

a r t i c l e

i n f o

Article history: Received 15 August 2013 Accepted 17 November 2013 Available online 23 November 2013 Keywords: Antigen delivery Cytosolic delivery Endosomes Photochemical internalization Cytotoxic T cells Melanoma

a b s t r a c t The protection or treatment of several immunological disorders is dependent on the antigen-specific and cytotoxic CD8 T cells. However, vaccines aimed at stimulating CD8 T-cell responses are typically ineffective because vaccine antigens are primarily processed by the MHC class-II and not the MHC class-I pathway of antigen presentation: the latter requires cytosolic delivery of antigen. In order to facilitate targeting of antigen to cytosol, the antigen was combined with the photosensitiser TPCS2a (disulfonated tetraphenyl chlorin) and administered intradermally to mice. The photosensitiser was activated by illumination of the injection site. This photochemical internalization (PCI) strongly increased the stimulation of CD8 T-cell responses as measured by antigen-specific proliferation and secretion of pro-inflammatory cytokines. Fluorescence microscopy showed that delivery to cytosol was TPCS2a dependent and occurred by light-induced disruption of TPCS2a- and antigen-containing endosomes. PCI-based vaccination prevented growth of malignant B16 cells as compared with vaccination without PCI. In conclusion, PCI represents a potent tool for delivery of antigens to cytosol for stimulation of cytotoxic CD8 T-cell responses. This study demonstrated a first proof-of-principle for PCI-mediated immunisation with potential application in cancer immunotherapy. © 2013 Elsevier B.V. All rights reserved.

1. Introduction Vaccines most probably represent the single medical intervention with the greatest impact on the global human health: childhood vaccines against bacterial toxins, measles, poliomyelitis and more prevent millions of fatalities each year. Vaccines primarily protect by means of neutralising antibodies [1] which are a result of antigen presentation via the MHC class-II pathway, the default pathway of exogenous antigens. When current vaccines fail to work, e.g. in tuberculosis, Hansens' disease (leprosy), most parasitic infections, some viral infections such as herpes and human immunodeficiency viruses, and tumours, it is mainly because protection or treatment is dependent on T cells, especially cytotoxic CD8 T cells. The signals for CD8 T-cell stimulation is mediated via peptide antigens loaded on MHC class-I molecules, and this generally occurs when the peptide is generated intracellularly; the source of such peptides is typically worn-out endogenous proteins or proteins from infecting virus. A major challenge in vaccine research

⁎ Corresponding author at: Department of Dermatology, University Hospital Zurich, Gloriastrasse 31, 8091 Zurich, Switzerland. Tel.: +41 44 255 8616; fax: +41 44 255 4418. E-mail address: [email protected] (P. Johansen). 0168-3659/$ – see front matter © 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.jconrel.2013.11.017

and development is therefore to find ways to access the MHC class I pathway of antigen presentation with exogenous antigens. Phagocytosis or endocytosis normally leads to phago-lysosome fusion and MHC class II antigen presentation [2]. Several technologies have been proposed for stimulation of CD8 T-cells, e.g. recombinant but non-replicating viruses that are able to infect antigen-presenting cells by other ways than phagocytosis [3–5]. We recently demonstrated that endocytosed antigens can be redirected to the MHC class I pathway of antigen presentation by using the drug-delivery technology photochemical internalisation (PCI) [6]. Firstly, the protein antigen ovalbumin (OVA) and the photosensitiser disulfonated tetraphenyl chlorine (TPCS2a) [7], were co-incubated with bone-marrow-derived dendritic cells in vitro, and the cells illuminated with defined light doses and energies. Secondly, the dendritic cells were used to stimulate antigenspecific CD8 T-cell responses in vitro and in vivo. The strength of the stimulated CD8 T-cell responses was strongly dependent on the photosensitisation of dendritic cells. This current report is the first study to show direct immunisation with a vaccine and using photosensitiser and light as adjuvants. The intradermal application of photosensitiser and light strongly facilitated MHC class-I presentation and CD8 T-cell stimulation of the co-delivered protein antigen in mice. This PCI method could open new avenues in the vaccination or

144

M. Håkerud et al. / Journal of Controlled Release 174 (2014) 143–150

immunotherapy of infections and cancer where protection or treatment is dependent on the stimulation of cytotoxic CD8 T-cell responses.

A) D0 D1

2. Materials and methods

D1-3

OT-1, i.v. Bleeding, Immunisation ± TPCS2a, i.d. Light (435nm, 1-12min)

2.1. Mice For immunisation, female C57BL/6 mice were purchased from Harlan (Horst, The Netherlands) and used at 6–10 weeks of age. Rag2 deficient OT-I mice transgenic for the T-cell receptor that recognises the MHC class-I-restricted epitope OVA257–264 (SIINFEKL) from ovalbumin (OVA) were originally purchased from Taconic Europe (Ry, Denmark) and bred in own facilities at the University of Zurich. All mice were kept under specified pathogen-free conditions, and the procedures performed were approved by the Swiss Veterinary authorities (licence 69/2012).

D7

D14

Bleeding (FACS)

Euthanisia (spleen cultures, FACS & ELISA)

B)

2.2. Materials The antigen chicken ovalbumin (OVA; Grade V) was purchased from Sigma-Aldrich (Buchs, Switzerland) and dissolved in PBS. The octapeptide SIINFEKL was purchased from EMC microcollections (Tuebingen, Germany). The photosensitiser TPCS2a (tetraphenyl chlorin disulfonate or Amphinex®) was provided by PCI Biotech (Lysaker, Norway) at a concentration of 30 mg/ml in polysorbate 80, mannitol and 50 mM Tris pH 8.5. TPCS2a was protected from light and kept at 4 °C. OVA and TPCS2a were mixed together in PBS, and the vaccine was kept light protected and used within 60 min of preparation. The light source used for the activation of the photosensitiser was LumiSource™ (PCI Biotech), which contains four 18 W Osram L18/67 standard light tubes with a fluence rate of 13.5 mW/cm2 with a maximum wavelength peak of 435 nm. The light doses are given as illumination time, were 1 min corresponds to a light energy of 0.81 J/cm2.

2.3. Intradermal photosensitisation and immunisation of mice One day prior to immunisation, spleens and lymph nodes were isolated from female OT-I mice, and erythrocytes were removed by lysis (RBC Lysing Buffer Hybri-Max from Sigma-Aldrich) of the homogenised cell suspensions. The remaining cells were washed in PBS, filtered through 70 μm nylon strainers, and 2 × 106 OT-I cells were administered by intravenous injection into recipient female C57BL/6 mice; the adoptive transfer of SIINFEKL-specific CD8 T cells facilitates monitoring of the immune response by flow cytometry. On the day of OT-I transfer the fur were also shaven off the abdominal region in order to enable better illumination with light. One day later, mice were bled by tail bleeding, and the blood was collected in heparin-containing tubes for analysis of the baseline frequency of SIINFEKL-specific CD8 T cells. The mice were then given 100 μl of OVA or a mixture of OVA and TPCS2a intradermally using syringes with 29G needles. In order to target a larger area of the skin, the vaccines were split in two injections of each 50 μl to the left and right side of the abdominal midline. OVA was tested at 10–100 μg per dose. The TPCS2a doses were 7.5–250 μg. At various time points after immunisation (0–48 h), the mice were anaesthetised by intraperitoneal injection of a mixture of ketamine (25 mg/kg body weight) and xylazin (4 mg/kg) and placed on the light source for activation of the photosensitiser. The illumination time was 6 min (4.86 J/cm2; 0.81 J/cm2 per minute), if not otherwise stated. The whole procedure is illustrated in the scheme of Fig. 1A. The illumination of mice is imaged in Fig. 1B. Typically, mice were bled on day 7 by tail bleeding for analysis of antigen-specific CD8 T cells by flow cytometry. At the end of the experiment (typically day 14), the mice were euthanized and the splenocytes analysed ex vivo.

Fig. 1. Experimental setup. (A) Scheme showing the experimental set of PCI-mediated immunisation using mice adoptively transferred with OVA-specific CD8 T-cell transgenic OT-1 cells prior to immunisation. (B) At various time points after intradermal injection of antigen (OVA) and photosensitiser (TPCS2a) in the abdominal region, the mice were anaesthetised and placed with the abdominal side down directly on the glass plate of the LumiSource light table. The light table allows the administration of a precise light dose to the treated region. An illumination time of 1 min corresponds to 0.81 J/cm2.

2.4. Analysis of immune responses by flow cytometry and ELISA The frequency of OVA-specific CD8 T-cells in blood was monitored by staining the cells with anti-CD8 antibody and H-2Kb/SIINFEKL Pro5 pentamer (Proimmune, Oxford, UK) for analysis by flow cytometry. The activation status of the cells was further analysed by testing the expression of CD44 and CD69. Intracellular staining for IFN-γ was done after overnight SIINFEKL stimulation of splenocytes at 37 °C in 24-well plates. Brefeldin A was added during the last 4 h. The cells were then washed and fixed with 4% formaldehyde in PBS on ice for 10 min. Anti-CD16/32 was added to block unspecific antibody binding to Fc receptors. The cells were permeabilised with 0.1% NP40 in PBS for 3 min and washed before being stained with anti-IFN-γ, anti-CD8 and antiCD44 antibodies (eBioscience or BD Pharmingen) for 35 min. The cells were acquired using FACSCanto (BD Biosciences, San Jose, USA) and analysed using FlowJo 8.5.2 software (Tree Star, Inc., Ashland, OR). Alternatively, 2 × 105 splenocytes were re-stimulated in 96-well plates with OVA protein or the SIINFEKL. After 24 and 72 h, supernatants were collected and analysed IL-2 or IFN-γ, respectively, by ELISA (eBioscience). 2.5. Live cell fluorescence microscopy Fifty thousand J774.1 cells (ATCC no. TIB-67 mouse monocyte macrophage cell line) were seeded out on no. 1.5 glass coverslips (Glasswarenfabrik Karl Hecht KG, Sondheim, Germany) in 4-well plates overnight. The cells were incubated with 0.05 or 1.0 μg/ml TPCS2a for 18 h and washed three times in drug-free culture medium prior to incubation with 25 μg/ml OVA-Alexa488 for 4 h. Cells were subsequently washed in ice-cold PBS with Ca2+ and Mg2+, illuminated at for 5 min (approx. 4 J/cm2) and rested 90 min prior to microscopy. Images of cellular localization and PCI-induced cytosolic release of OVA were obtained by epi-fluorescence microscopy using a Plan-Apochromat 63×/1.40 Oil differential interference contrast (DIC) objective or 40×/0.95 PlanApochromat phase contrast (Korr Ph3 M27) objective with a Zeiss Axioimager Z.1 microscope (Carl Zeiss, Oberkochen, Germany). Fluorescence of Alexa488-labelled OVA was obtained by using a 470/40 nm

M. Håkerud et al. / Journal of Controlled Release 174 (2014) 143–150

3. Results 3.1. PCI facilitates antigen-specific stimulation of IFN-γ-secreting CD8 T cells To facilitate analysis of MHC class-I antigen presentation, we used the class-I-binding octapeptide SIINFEKL from OVA (aa257-264) in combination with SIINFEKL-specific CD8 T cells from T-cell receptor transgenic OT-I mice. OT-I lymphocytes, 2 × 106 cells, were adoptively transferred to syngeneic and sex-matched wild type C57BL/6 mice. One day after the transfer, approximately 1.4% of all CD8-positive T cells in peripheral blood was SIINFEKL-specific (Fig. 2A); the frequency of SIINFEKL-specific CD8 T cells in C57BL/6 mice, which did not receive an adoptive transfer of OT-I cells was less than 0.05% (data not shown). The mice were then typically immunised with 10–100 μg OVA protein or with a mixture of OVA and 7.5–250 μg of the photosensitiser TPCS2a by intradermal administration in the abdominal region (not all conditions given in the representative data illustrated as figures). At different time points thereafter, the mice were anaesthetised and placed belly-down with the treated abdominal side facing the light source (Fig. 1B), and the site of vaccination was illuminated for 6 min (4.86 J/cm2). By day 6 after vaccination, the frequency of SIINFEKLspecific CD8 T cells in the peripheral blood of mice vaccinated with 100 μg OVA had increased to approximately 3.5% (Fig. 2B). A similar frequency was measured in mice that also received 15 μg TPCS2a and were illuminated 2 h after vaccination (Fig. 2B). However, when mice were illuminated 18 h post-vaccination, a significant increase in the number of SIINFEKL-specific CD8 T cells was measured in blood (Fig. 2B; P = 0.0286 by Mann Whitney). Typically, a retraction of the number of SIINFEKL-specific CD8 T cells in blood was observed 10–15 days after vaccination. By day 23 post-vaccination, the numbers of antigenspecific CD8 T cells had retracted to baseline levels in mice immunised with OVA alone or OVA plus TPCS2a and illuminated 2 h after administration (Fig. 2C). Also mice immunised with OVA and TPCS2a and illuminated at 18 h after immunisation showed reduced frequencies after 23 days, but still significantly higher than baseline (P = 0.0294 by two-way Mann Whitney). While the SIINFEKL-specific cells in blood of untreated animals had a non-activated phenotype with lack of activation markers such as CD44 (Fig. 2D), CD25 and CD69 (not shown), both immunisation with OVA and OVA–PCI caused strong upregulation of these markers by day 6. The up-regulation induced by PCI was specific for the antigen-specific T cells, since no up-regulation

MHC-SIINFEKL pentamer (% of CD8)

B) 10

10

D0

C) 10

D6

8

8

8

6

6

6

4

4

4

2

2

2

0

0

n=20

0

OVA

-

-

+

PCI

-

-

2h Treatment

D)

D23

+

Untreated (D6)

+

-

+

+

+

18h

-

-

2h

18h

OVA (D6)

OVA + PCI (D6)

MHC-SIINFEKL pentamer

Animals were immunised intradermally as described above with 10 μg OVA with our without 200 μg TPSC2a. TPSC2a-treated mice were illuminated 18 h after OVA and TPSC2a treatment, and the illumination time was 6 min (4.86 J/cm2). One day prior to vaccination, the mice received 10,000 OT-I cells intravenously. On day 4 after vaccination, the mice received 5 × 105 SIINFEKL-expressing malignant B16 cells by intradermal injection into one of the flanks. The B16 cell line (ATCC® CRL-6322™) stems from C57BL/6 mice, and the SIINFEKL-expressing line was kindly provided by Emmanuel Contassot (University of Zurich). The growth in the mice was monitored by measuring the size of the neoplasm with a calliper 14 days after injection, the endpoint of the investigation. The volume was calculated by using the modified ellipsoid formula: (length × width2)/2.

CD44 expression

E) 2.0

IFNγ +CD44+CD8+ (%)

2.6. Effect of vaccination on growth of malignant cells

A)

1.5

0.6

IFN-γ ICS/D14 (blood)

IFN-γ ICS/D23 (spleen)

0.4 1.0 0.2 0.5 0.0

0.0

OT-I

+

+

+

+

-

+

+

+

OVA

-

+

+

+

-

-

+

+

PCI

-

2h Treatment

18h

-

-

2h Treatment

+ + 18h

F) Cytokine (ELISA, OD)

band pass (BP) excitation filter with a beam splitter at 495 nm and a 525/50 nm BP emission filter. TPCS2a fluorescence was obtained by using a 395–440 nm BP excitation filter with a beam splitter at 460 nm, and a 620 nm long pass filter. Micrographs were recorded with a digital AxioCam MRm camera and processed and analysed by use of the Axiovision Software (Carl Zeiss).

145

2.0

0.8

IFN-γ ELISA/D23

1.5

0.6

1.0

0.4

0.5

0.2

IL-2 ELISA/D23

0.0

0.0

OT-I

-

+

+

+

+

-

+

+

+

OVA

-

-

+

+

+

-

-

+

+

+

PCI

-

-

-

2h

18h

-

-

-

2h

18h

Treatment

+

Treatment

Fig. 2. PCI adjuvates stimulation of CD8 T cell immune responses. C57BL/6 mice were spiked with 2 × 106 OT-I cells and the frequency of SIINFEKL-specific cells were measured in the recipients after 18 h by MHC I-SIINFEKL pentamer staining and flow cytometry (A). The mice were then immunised with 100 μg OVA or with 100 μg OVA and 25 μg TPCS2a; control mice were left untreated. After 2 or 18 h, the TPCS2a-treated mice were illuminated for 6 min (4.86 J/cm2). On days 6 (B) and 23 (C), mice were bled and stained with MHC I-SIINFEKL pentamer, anti-CD8 and anti-CD44 antibodies and analysed by flow cytometry. Bars show the frequency of triple positive cells relative to the total number of CD8 T cells. (D) Dot blots of pentamer- and CD44-positive cell from blood analysed by flow cytometry on day 6. Cells were gated on CD8 lymphocytes. (E) On day 23, blood (left panel) and splenocytes (right panel) were re-stimulated overnight with SIINFEKL and analysed for CD8, CD44 and IFN-γ by intracellular staining (ICS). (F) Splenocytes were re-stimulated with SIINFEKL for analysis of IFN-γ (left panel) and IL-2 (right panel) by ELISA.

of CD44 could be observed on SIINFEKL-pentamer staining CD8 T cells from untreated mice. On day 14, the mice were bled and the PBMCs cells re-stimulated with SIINFEKL overnight. After staining for surface CD8 and CD44 and intracellular IFN-γ, the cells were acquired by flow cytometry and the frequency of triple-positive cells within all CD8-positive cells was calculated. OVA-immunised mice had a 4-fold increased frequency as compared to control mice that had received OT-I transfer only (Fig. 2E, left

M. Håkerud et al. / Journal of Controlled Release 174 (2014) 143–150

Since immunisation with PCI did not produce responders in all animals tested (typically 3–4 out of 5), we further tested the effect of the time interval between TPCS2a administration and illumination on the stimulated immune response. Intervals of 6–8 h or of 42 h did not provide an adjuvant effect of PCI (data not shown). However, an interval of approximately 18 h repeatedly gave an adjuvant effect of PCI. This was observed without exceptions in four independent experiments. In order to increase the sensitivity of the experimental system used, we tested the adjuvant effect of PCI at limiting doses of OVA. Intradermal immunisation with 10 μg OVA alone produced no proliferation and activation of SIINFEKL-specific CD8 T cells in blood (data not shown). Several experiments with TPCS2a at 7.5 to 250 μg then suggested that increasing TPCS2a doses also increased the measured OVA-specific immune response. Representatively, PCI with 25 μg TPCS2a caused 40% good responders, 40% week responders and 20% non-responders as measured for SIINFEKL-specific CD8 T cells in blood on day 7, while PCI with 250 μg TPCS2a produced 100% good responders (Fig. 3A). On day 14, the splenocytes were tested by flow cytometry for IFN-γ production. Immunisation with OVA alone gave week responders in all mice tested, whereas immunisation with OVA and PCI caused better responders in nine out ten (90%) mice tested (Fig. 3B). Again, PCI with 250 μg TPCS2a showed 100% responders and the highest frequency of IFN-γ producing cells. Whereas intracellular staining and flow cytometry qualitatively measure whether cells can produce cytokines, ELISA measures how much cytokine the cell can produce. We therefore restimulated the splenocytes with SIINFEKL in vitro and analysed IL-2 (Fig. 3C) and IFN-γ (Fig. 3D) after 24 and 72 h, respectively. Immunisation with OVA alone induced weak, but clearly measurable IL-2, but no IFN-γ secretion. Immunisation with OVA and PCI at 25 μg TPCS2a did not cause an increase in IL-2, but a strong increase in IFN-γ secretion as compared to immunisation with OVA alone. At 250 μg TPCS2a, a fourfold increase in IL-2 secretion and a 17-fold increase in IFN-γ secretion were detected. However, while TPCS2a-PCI of OVA had a dose-dependent adjuvant effect with regards to the immune response measured, higher TPCS2a doses also caused more local inflammation with transient erythema and occasionally lesions on days 1–3 after illumination (data not shown). In most of the immunisation experiments performed, the PBMCs and splenocytes were also re-stimulated with the whole OVA protein. Here, the effect of PCI on immunisation with respect to IFN-γ and IL-2 secretion was less prominent (data not shown) probably due to a polyclonal response to other epitopes than SIINFEKL, especially CD4 T-cell epitopes. Moreover, in mice that received OVA and light, but not TPCS2a or in mice that received OVA and TPCS2a but

C) 4

0.4

P<0.01

0.3

IL-2 (OD)

3 2

0.2

1

0.1

0

0.0

B)

D) 0.5

1.0

P<0.05 0.4

0.8

IFNγ (OD)

3.2. Timing of illumination and the dose of photosensitiser affect the adjuvant effect

A) MHC-SIINFEKL pentamer (%; CD8+, CD44+)

panel). The increase in IFN-γ-producing CD44-positive cells after PCI treatment was 6-fold (illumination at 2 h) and 15-fold (18 h). On day 23, mice were euthanized and splenocytes cultured overnight with SIINFEKL. The cells were then analysed for intracellular IFN-γ by flow cytometry (Fig. 2E, right panel) or for the secretion of IL-2 (24 h) and IFN-γ (72 h) by ELISA (Fig. 2F). The intracellular IFN-γ staining showed background frequencies of CD44-positive IFN-γ producing cells in splenocytes from OVA-immunised mice that did not receive parallel PCI treatment (Fig. 2E, right panel). Clearly higher frequencies of IFN-γ producing cells were detected in splenocytes from mice that received PCI-treatment. Again, 18 h interval between immunisation and illumination was most beneficial. Splenocytes from all OVA-immunised mice showed significant production of both of IL-2 and IFN-γ when compared to non-immunised OT-I recipients. Although not statistically significant, there was a clear tendency for increased cytokine secretion in splenocytes from mice that were immunised with PCI treatment.

IFN-γ (% of CD8)

146

0.3 0.2 0.1

0.0 OVA TPCS2a Light

0.6 0.4 0.2

+ -

+ 25 6

+ 250 6

0.0 OVA TPCS2a Light

Treatment

-

+ -

+ 25 6

+ 250 6

Treatment

Fig. 3. C57BL/6 mice were spiked with 2 × 106 OT-I cells. After 8 h, the mice were immunised with 10 μg OVA, with 10 μg OVA and 25 μg TPCS2a, or with 10 μg OVA and 250 μg TPCS2a. On day 11, the mice were euthanized and the splenocytes analysed for (A) MHC I-SIINFEKL pentamer, CD44 and CD8 staining, (B) CD8 and CD44 and intracellular IFN-γ, as well as secretion of IL-2 (C) and IFN-γ (D) measured by ELISA. Bars show the frequency of triple positive cells relative to the total number of CD8 T cells.

not light, no beneficial effect of light or TPCS2a, respectively, was observed (data not shown). Only at very high doses of TPCS 2a , some adjuvant effect was seen, but this has been ascribed to environmental light that activates the photosensitiser; it was not possible to keep the animal in complete darkness for the time of the experiments. 3.3. The adjuvant effect of PCI is long lasting The longevity or the memory of the observed CD8-positive immune responses were tested in mice immunised as described above using 20 μg OVA with or without 200 μg TPCS2a. After 54 days, the mice were euthanized and the splenocytes analysed directly for the frequency and function of SIINFEKL-specific CD8 T cells. As shown in Fig. 4A, the frequencies of measurable SIINFEKL-specific CD8 T cells in mice treated immunisation with OVA or with OVA and PCI were not different from untreated mice. However, re-stimulation with SIINFEKL overnight revealed that PCI-treatment enabled stimulation of antigen-specific CD8 memory cells, which reacted by secretion of the effector cytokine IFNγ. This was observed both by intracellular staining and flow cytometry (Fig. 4B) and by ELISA (Fig. 4C). By both assay, a statistically significant difference was observed between OVA alone and OVA–PCI treated mice (P b 0.01). 3.4. PCI mediates cytosolic delivery of antigen To study the mechanism by which PCI mediated the adjuvant effect, murine J774.1 cells, an antigen-presenting macrophage cell line, were incubated with Alexa488-labelled OVA with or without parallel PCI treatment. As shown in the fluorescence micrograph of Fig. 5A, in cells

M. Håkerud et al. / Journal of Controlled Release 174 (2014) 143–150

O

O

Treatment

Treatment

0.4 0.2

VA

PC N I äi ve

0.0 O

O

VA

0.0

VA

O

0.1

PC N I äi ve

0.0

0.2

P=0.041

0.6

VA

0.1

0.3

0.8

O

0.2

PC N I äi ve

0.3

P=0.044

0.4

VA

0.4

0.5

VA

MHC-SIINFEKL (% of CD8)

0.5

C) IFNγ+ (ELISA, OD)

B) IFNγ+CD44+CD8+ (%)

A)

Treatment

147

OVA localised close to the plasma membrane of the cells. After parallel PCI of OVA treatment, the cytosol showed a more generalised diffuse green fluorescence suggesting light-triggered endosomal escape of the antigen to the cytosol. Moreover, the fluorescence intensity in PCI-treated cells was generally much stronger than in cell treated with OVA only. This indicates that PCI also facilitated better antigen uptake in general. Since the photosensitiser TPCS2a is auto-fluorescent, it enabled the study of the relative localisation of antigen and photosensitiser after incubation of J774 cells with Alexa488-labelled OVA (green) and TPCS2a. Again, a diffuse distribution of antigen and TPCS2a throughout the cytosol was observed (Fig. 5B). The merge of the TPCS2a and OVA-Alexa488 images revealed that antigen and photosensitiser are co-localised in the cytosol.

6

Fig. 4. C57BL/6 mice were spiked with 2 × 10 OT-I cells. One day later, the mice were immunised with 20 μg OVA, with 20 μg OVA and 200 μg TPCS2a, or left untreated. On day 54, the mice were euthanized and the splenocytes analysed by flow cytometry for (A) MHC I-SIINFEKL pentamer and CD8 staining, or (B) intracellular IFN-γ and CD8 and CD44 staining. Bars show the frequency of triple positive cells relative to the total number of CD8 T cells. (C) Secretion of IFN-γ into 96-h splenocyte cultures was measured by ELISA.

treated with OVA alone, antigen uptake was observed, and the antigen was located in concise spherical-shaped bodies, suggesting that the antigen was contained in vesicles, e.g. endosomes. The Alexa488-labelled

A)

OVA

OVA + PCI

3.5. PCI-based vaccination can prevent growth of malignant B16 cells We next investigated if the PCI-based vaccination effected the growth of malignant B16 cells in vivo. Mice received SIINFEKLexpressing B16 cells by intradermal injection 4 days after vaccination, and the growth was measured on day 14 post injection of the cells. In non-vaccinated mice, the transfer of 2 × 106 OT-1 cells, as described for the immunisation studies above, totally prevented B16 growth (data not shown). Therefore, the number of transferred cells was reduced to 1 × 104 OT-I cells. When compared to untreated controls, a significantly reduced B16 growth was observed in mice that received PCI-based vaccination with OVA (P b 0.05 by Kruskal–Wallis test) but not after vaccination with OVA alone (Fig. 6A). When all data were transformed to binary data (0 = no growth; 1 = growth) and analysed by the Chi-square test, PCI-based vaccination had a significantly stronger suppressing effect on B16 growth than vaccination with OVA alone (P = 0.048). Fig. 6B shows representative micrographs of neoplasms on day 14 from differently treated mice. 4. Discussion Lysosomal degradation is the most common fate of vaccines and microbes [2]. As a consequence, nearly all vaccines work through

600 400 200

I C

O

Treatment

nt r

0

-P

Nomarski

*

800

U

Merge

n.s.

1000

VA

A)

O

TPCS2a

VA

OVA

Tumour volume (mm3)

B)

B)

Fig. 5. PCI-mediated cytosolic release of antigen. (A) J774.1 cells were incubated 4 h with 25 μg/ml OVA-Alexa488 (left panel) or 18 h with 0.05 μg/ml TPCS2a and a subsequent 4 h with 25 μg/ml OVA-Alexa488 only (right panel). The cells were again washed and illuminated (4 J/cm2). Cellular uptake and distribution of OVA-Alexa488 was analysed by fluorescence microscopy 90 min after illumination. (B) J774 cells were incubated with 1.0 μg/ml TPCS2a and 25 μg/ml OVA-Alexa488 as above and analysed for cellular uptake, distribution and co-localisation of OVA-Alexa488 (green) and TPCS2a (red) by fluorescence microscopy. Yellow colour after merge of the fluorescence images suggests colocalisation of the two compounds. A differential interference contrast microscopy (Nomarski) of the same section is also shown. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Untreated

OVA

OVA + PCI

Fig. 6. Effect of PCI-based vaccination on malignant B16 growth. C57BL/6 mice were spiked with 1 × 104 OT-I cells. One day later, the mice were immunised with 10 μg OVA, or 10 μg OVA and 200 μg TPCS2a, or left untreated. On day 4 after immunisation, the mice received an intradermal injection of 5 × 105 SIINFEKL-expressing B16 cells. Two weeks thereafter, the size of the solid B16 neoplasm was measured and the volume calculated (A) and photographs were made (B). n.s.: not significant; *: p b 0.05 as analysed by Kruskal–Wallis test.

148

M. Håkerud et al. / Journal of Controlled Release 174 (2014) 143–150

stimulation of B cells for secretion of antibodies [8]. Such stimulation is a result of antigen presentation via the MHC class-II pathway. In fact, MHC class-II antigen presentation is a default pathway for most exogenous materials, including vaccines, particles and bacteria. However, many pathogens are facultative intracellular organisms that replicate in the cytosol of cells, especially in macrophages [9]. Some viral pathogens can even produce immunoevasins that interfere with MHC class-I antigen presentation and prevent immunity [10,11], while Mycobacterium tuberculosis is able to avoid degradation in macrophages by preventing phagosomal maturation and phagosome–lysosome fusion [12]. Since antibodies have no access to the intracellular compartments, they play almost no role in the protection against such pathogens. Likewise, cytokines secreted by CD4 T cells only have indirect effect on the survival and replication of intracellular pathogens, for example by TNF-α and IFN-γ-mediated activation of infected macrophages [13]. In contrast, antigen-specific cytotoxic CD8 T cells have been shown very effective in eliminating infected cells [13,14] as well as tumours [5]. Tumours are often ignored by the immune system since immune surveillance is based on recognition of non-self-antigens and of pathogen patterns [15]. Unless specific release mechanisms have been incorporated to facilitate cytoplasmic delivery, MHC class-II antigen presentation is the fate of any vaccine. One important goal of modern vaccine development is therefore to find methods that bypass MHC classII and enables MHC class-I antigen presentation. Briefly, this can be achieved in either of two ways. Firstly, antigens could be directly shuffled across the plasma membrane into the cytosol with its proteasome that can generate MHC-class-I-binding peptides for delivery in the endoplasmic reticulum. Secondly, antigens could be taken up by conventional endocytosis (solutes) or phagocytosis (particles), but by disruption of the endosomes prior to their fusion with lysosomes, the antigen could be released into the cytosol and again processed by the proteasomes. In the current study in mice, the latter strategy was followed. By combining the vaccine antigen with a photosensitiser, we aimed at disrupting the photosensitiser- and antigen-containing endosomes through light activation. This study demonstrated that immunisation with a combination of PCI and a protein antigen resulted in strongly improved stimulation of CD8 T cells as measured by proliferation and pro-inflammatory cytokine secretion, and these immune responses hampered growth of a CD8sensitive malignant cell line in mice. The immune responses were measured with respect to the MHC class I (H-2Kb)-binding peptide SIINFEKL, by pentameter staining and flow cytometry and by restimulation of splenocytes with SIINFEKL. Hence, the beneficial effect of PCI on SIINFEKL-specific immune responses suggests that OVA indeed undergoes an endosome-to-cytosol transfer and that the protein is correctly digested by the proteasome for loading of SIINFEKL to MHC class I. The PCI-mediated cytosolic release of antigen from the endosome was also confirmed by live-cell fluorescence microscopy. Untreated phagocytic cells incubated with fluorescently labelled OVA sequestered the antigen in endocytic vesicles during the experimental period. In contrast, photosensitised and light-exposed cells caused a re-localisation of both OVA and photosensitiser from endosomes to cytosol. Importantly, the cells as such were intact so that the antigen entered the cytosol via PCI-mediated disruption of the endosomes and not due to a disruption of the plasma membrane. The process by which PCI can mediate endosomal release of antigens for MHC class-I antigen presentation is illustrated in the cartoon of Fig. 7. During the last two decades, several approaches have been suggested for the delivery of antigens and drugs to the cytosol. Especially in gene delivery, the need to target the nucleus with DNA or oligonucleotides has triggered much research in the area of targeted intracellular delivery [16]. Some strategies have exploited properties of certain viruses and bacteria [17–21], for instance cell-penetrating peptides from HIV-1 (TAT) and Herpes simplex

Antigen-presenting cell (APC)

APC with sensitised endosome

Vaccine Ag PS

Proteasome Nucleus

PS wash-out period (18h in vivo)

Light

Light-triggered endosomal break-up and MHC class-I processing and Ag presentation Fig. 7. Schematic illustration of PCI-mediated endosomal escape of antigen and stimulation of CD8 T cells. The photosensitiser (PS) and protein antigen (Ag) bind to the plasma membrane of antigen presenting cells (APCs) such as dendritic cells and macrophages. The PS and Ag are taken up by endocytosis, whereby the PS is contained in the endosomal membrane and the Ag is contained in the endosomal lumen. As a result of endocytosis, the PS is inverted so that its polar head is oriented towards the endosomal lumen; this, in addition to its amphiphilic nature, prevents it from being washed out of the endosome. After a PS wash-out period (2–4 h in vitro and ca. 18 h in vivo), residual PS contained in the outer plasma membrane is removed; this prevents disruption of the plasma membrane upon light exposure. Light is then exposed to the cells. The PS-containing endosomes break up and the Ag is release into the cytosol for access to the proteasomes and the machinery of MHC class I presentation.

virus (HSV VP-22) [20–23]. Another important method to enable cytoplasmic delivery of antigens or drugs is based on pH-sensitive endosomolytic bacterial hemolysins, e.g. perfringolysin O and listeriolysin O [17,18]. Some hemolysins are pore-forming toxins, typically produced by gram positive bacteria within the genera Clostridium, Streptococcus, Bacillus and Listeria. Biodegradable particles have also been extensively investigated for use in cytosolic delivery. Fusogenic and pH-sensitive liposomes, especially cationic liposomes, that undergo a phase transition in the acidic environment of late endosomes were the first particles to be studied and did show potential applications [24,25]. Also liposomes based on reconstituted influenza virus envelopes (virosomes) containing the viral hemagglutinin A have been used for cytoplasmic delivery due to the fusogenic properties of hemagglutinin [19,26]. However, in order to prolong therapeutic drug concentrations, drug-delivery systems based on synthetic polymers have been suggested, and some polymeric particles have been shown to cause a destabilisation of the endosomal membrane, thereby triggering their release into the cytosol [16,27]. Both for liposomes and polymeric nanoparticles, the utilised materials and the particle surface play a crucial role for the efficacy of cytoplasmic delivery, and much research deals with the tuning of such materials in the different targeted cell types or intracellular structures. In order to trigger the release of liposomal contents, also light activation has been utilised in previous studies. In 1994, photosensitive liposomes for delivery of drugs to cytosol were described [28] and similar drug formulations are currently still subject to research [29].

M. Håkerud et al. / Journal of Controlled Release 174 (2014) 143–150

While many of the described methods for triggered endosomal release of drugs and antigens are dependent on the acidification of the endosomal compartment, PCI is independent of the pH drop. An obvious drawback of pH-sensitive delivery systems is that peptides, proteins and DNA may degrade in the acidic endosomes prior to their cytoplasmic release. With PCI, it is possible to trigger release of antigen prior to acidification of the endosome. The combination of a photosensitiser and a therapeutic drug for the light-triggered release from endosomes and lysosomes has been described as PCI for the delivery a variety of compounds, e.g. proteins, peptides, oligonucleotides, genes and low molecular weight drugs [30–32]. The delivery to tumour cells is pre-clinically documented and is now in a Phase-II clinical trial for the delivery of bleomycin to patients with head-and-neck squamous cell carcinoma [33]. The current study follows up on these studies by applying it to antigens, for the cytoplasmic delivery of these with the goal of stimulating cytotoxic CD8 T-cell responses. The study is also in debt to other studies which showed that the photodynamic therapy (PDT) could induce anti-tumour immunity. Already in 1994, Canti et al. demonstrated that the PDT on fibrosarcoma-bearing mice caused improved survival, and a re-challenge with the parental tumour, but not with tumours of other specificities, resulted in tumourspecific immunity and resistance of tumour growth [34]. These data have later been supported by other groups, both by using in situ PDT treatment of solid tumours and by using autologous vaccination with tumour cells being PDT treated in vitro [35–40]. However, while PDT is often non-curative due to PDT-induced dysfunction of immune cells and immune suppression, combination approaches have been suggested [37], including the use of additional adjuvants such as CpG oligonucleotides [36]. One similarity of PDT- and PCI vaccines is a potential dependency on cell death, at least for in vitro generated vaccines [6,39]. However, PCI vaccines represent a further development of PDT vaccines, with the advantage of more direct and specific triggering of cytotoxic CD8 T-cell responses than PDT vaccines, which mediated a broad-range immune response including both CD4 and CD8 T cells as well as B cells [41]. 5. Conclusion We recently demonstrated the potential for PCI-mediated vaccination for stimulation of CD8 T-cell responses using photosensitised dendritic cells for antigen presentation in vitro [6]. While previous studies have demonstrated immunological effects of conventional PDT, the current study is the first to apply photosensitisation and immunisation with a purified antigen directly in vivo. While it demonstrated a first proof-of-principle for PCI-mediated vaccination, including protective immunity that hampered the growth of malignant cells in vivo, further studies in preclinical and therapeutic tumour models are required to verify and extend the observed effects reported here. However, next to the suggested beneficial effect on tumour growth, one can envisage the application of PCI also in vaccination against CD8 T-cell dependent infectious diseases such as malaria, AIDS, and hepatitis B infections, against which we currently have no or only poorly effective vaccines. For the utilisation of PCI-based vaccines in the treatment of cancer, e.g. in skin basalioma and spinalioma, melanoma, osteosarcoma, squamous cell carcinoma, and head & neck cancer, both the chemotherapeutic effect of the photosensitiser and the immunotherapeutic effect of its combination with tumour antigens can be anticipated. Acknowledgements The project was financially supported from Fonds für Medizinische Forschung at the University of Zurich, the Novartis Foundation for Medical–Biological Research, and the Norwegian Research Council.

149

Interest conflicts M.H., P.K.S. and A.H are employees of PCI Biotech, which has field patents on the use of photosensitiser in vaccination. A.H. and P.J. are mentioned as inventors of patents describing the use of PCI in immunisation and vaccination. The other authors disclose no conflict of interest. References [1] R.M. Zinkernagel, On natural and artificial vaccinations, Annu. Rev. Immunol. 21 (2003) 515–546. [2] R.S. Flannagan, V. Jaumouille, S. Grinstein, The cell biology of phagocytosis, Annu. Rev. Pathol. 7 (2012) 61–98. [3] R.A. Koup, D.C. Douek, Vaccine design for CD8 T lymphocyte responses, Cold Spring Harb. Perspect. Med. 1 (2011) a007252. [4] R.K. Tyagi, N.K. Garg, T. Sahu, Vaccination strategies against malaria: novel carrier(s) more than a tour de force, J. Control. Release 162 (2012) 242–254. [5] F. Arce, K. Breckpot, M. Collins, D. Escors, Targeting lentiviral vectors for cancer immunotherapy, Curr. Cancer Ther. Rev. 7 (2011) 248–260. [6] Y. Waeckerle-Men, A. Mauracher, M. Hakerud, D. Mohanan, T.M. Kundig, A. Hogset, P. Johansen, Photochemical targeting of antigens to the cytosol for stimulation of MHC class-I-restricted T-cell responses, Eur. J. Pharm. Biopharm. 85 (2013) 34–41. [7] K. Berg, S. Nordstrand, P.K. Selbo, D.T. Tran, E. Angell-Petersen, A. Hogset, Disulfonated tetraphenyl chlorin (TPCS2a), a novel photosensitizer developed for clinical utilization of photochemical internalization, Photochem. Photobiol. Sci. 10 (2011) 1637–1651. [8] R.M. Zinkernagel, Immunological memory not equal protective immunity, Cell. Mol. Life Sci. 69 (2012) 1635–1640. [9] M.M. Curtis, S.S. Way, Interleukin-17 in host defence against bacterial, mycobacterial and fungal pathogens, Immunology 126 (2009) 177–185. [10] V. Bohm, C.O. Simon, J. Podlech, C.K. Seckert, D. Gendig, P. Deegen, D. Gillert-Marien, N.A. Lemmermann, R. Holtappels, M.J. Reddehase, The immune evasion paradox: immunoevasins of murine cytomegalovirus enhance priming of CD8 T cells by preventing negative feedback regulation, J. Virol. 82 (2008) 11637–11650. [11] M.J. Reddehase, Antigens and immunoevasins: opponents in cytomegalovirus immune surveillance, Nat. Rev. Immunol. 2 (2002) 831–844. [12] P. Johansen, A. Fettelschoss, B. Amstutz, P. Selchow, Y. Waeckerle-Men, P. Keller, V. Deretic, L. Held, T.M. Kundig, E.C. Bottger, P. Sander, Relief from Zmp1-mediated arrest of phagosome maturation is associated with facilitated presentation and enhanced immunogenicity of mycobacterial antigens, Clin. Vaccine Immunol. 18 (2011) 907–913. [13] J.C. Ray, J. Wang, J. Chan, D.E. Kirschner, The timing of TNF and IFN-gamma signaling affects macrophage activation strategies during Mycobacterium tuberculosis infection, J. Theor. Biol. 252 (2008) 24–38. [14] F. Winau, S. Weber, S. Sad, J. de Diego, S.L. Hoops, B. Breiden, K. Sandhoff, V. Brinkmann, S.H. Kaufmann, U.E. Schaible, Apoptotic vesicles crossprime CD8 T cells and protect against tuberculosis, Immunity 24 (2006) 105–117. [15] A.F. Ochsenbein, Immunological ignorance of solid tumors, Springer Semin. Immunopathol. 27 (2005) 19–35. [16] A. Aied, U. Greiser, A. Pandit, W. Wang, Polymer gene delivery: overcoming the obstacles, Drug Discov. Today 18 (2013) 1090–1098. [17] S. Gottschalk, R.K. Tweten, L.C. Smith, S.L. Woo, Efficient gene delivery and expression in mammalian cells using DNA coupled with perfringolysin O, Gene Ther. 2 (1995) 498–503. [18] Z.F. Walls, S. Goodell, C.D. Andrews, J. Mathis, K.D. Lee, Mutants of listeriolysin O for enhanced liposomal delivery of macromolecules, J. Biotechnol. 164 (2013) 500–502. [19] E. Mastrobattista, P. Schoen, J. Wilschut, D.J. Crommelin, G. Storm, Targeting influenza virosomes to ovarian carcinoma cells, FEBS Lett. 509 (2001) 71–76. [20] V. Gopal, Bioinspired peptides as versatile nucleic acid delivery platforms, J. Control. Release 167 (2013) 323–332. [21] H. Margus, K. Padari, M. Pooga, Insights into cell entry and intracellular trafficking of peptide and protein drugs provided by electron microscopy, Adv. Drug Deliv. Rev. 65 (2013) 1031–1038. [22] N.A. Alhakamy, A.S. Nigatu, C.J. Berkland, J.D. Ramsey, Noncovalently associated cell-penetrating peptides for gene delivery applications, Ther. Deliv. 4 (2013) 741–757. [23] M.C. Shin, J. Zhang, K.A. Min, K. Lee, Y. Byun, A.E. David, H. He, V.C. Yang, Cell-penetrating peptides: achievements and challenges in application for cancer treatment, J. Biomed. Mater. Res. A (2013)(in press, PMID 23852939). [24] O.V. Gerasimov, J.A. Boomer, M.M. Qualls, D.H. Thompson, Cytosolic drug delivery using pH- and light-sensitive liposomes, Adv. Drug Deliv. Rev. 38 (1999) 317–338. [25] J.K. Vasir, V. Labhasetwar, Biodegradable nanoparticles for cytosolic delivery of therapeutics, Adv. Drug Deliv. Rev. 59 (2007) 718–728. [26] J. Angel, L. Chaperot, J.P. Molens, P. Mezin, M. Amacker, R. Zurbriggen, A. Grichine, J. Plumas, Virosome-mediated delivery of tumor antigen to plasmacytoid dendritic cells, Vaccine 25 (2007) 3913–3921. [27] E. Mastrobattista, W.E. Hennink, Polymers for gene delivery: charged for success, Nat. Mater. 11 (2012) 10–12. [28] D.E. Bennett, H. Lamparski, D.F. O'Brien, Photosensitive liposomes, J. Liposome Res. 4 (1994) 331–348. [29] L. Paasonen, T. Sipila, A. Subrizi, P. Laurinmaki, S.J. Butcher, M. Rappolt, A. Yaghmur, A. Urtti, M. Yliperttula, Gold-embedded photosensitive liposomes for drug delivery: triggering mechanism and intracellular release, J. Control. Release 147 (2010) 136–143.

150

M. Håkerud et al. / Journal of Controlled Release 174 (2014) 143–150

[30] K. Berg, P.K. Selbo, L. Prasmickaite, T.E. Tjelle, K. Sandvig, J. Moan, G. Gaudernack, O. Fodstad, S. Kjolsrud, H. Anholt, G.H. Rodal, S.K. Rodal, A. Hogset, Photochemical internalization: a novel technology for delivery of macromolecules into cytosol, Cancer Res. 59 (1999) 1180–1183. [31] A. Hogset, L. Prasmickaite, P.K. Selbo, M. Hellum, B.O. Engesaeter, A. Bonsted, K. Berg, Photochemical internalisation in drug and gene delivery, Adv. Drug Deliv. Rev. 56 (2004) 95–115. [32] P.K. Selbo, A. Weyergang, A. Hogset, O.J. Norum, M.B. Berstad, M. Vikdal, K. Berg, Photochemical internalization provides time- and space-controlled endolysosomal escape of therapeutic molecules, J. Control. Release 148 (2010) 2–12. [33] An Open-label, Single Arm, Multi-centre Phase II Study to Evaluate Safety and Efficacy of PC-A11 in Patients With Recurrent Non-metastatic Head and Neck Squamous Cell Carcinoma Unsuitable for Surgery and Radiotherapy, 2012. (ClinicalTrials.gov, pp. NCT01606566). [34] G. Canti, D. Lattuada, A. Nicolin, P. Taroni, G. Valentini, R. Cubeddu, Antitumor immunity induced by photodynamic therapy with aluminum disulfonated phthalocyanines and laser light, Anti-Cancer Drugs 5 (1994) 443–447.

[35] M. Korbelik, Cancer vaccines generated by photodynamic therapy, Photochem. Photobiol. Sci. 10 (2011) 664–669. [36] Y. Xia, G.K. Gupta, A.P. Castano, P. Mroz, P. Avci, M.R. Hamblin, CpG oligodeoxynucleotide as immune adjuvant enhances photodynamic therapy response in murine metastatic breast cancer, J. Biophotonics (2013)(in press, PMID 23922221). [37] T.G. St Denis, K. Aziz, A.A. Waheed, Y.Y. Huang, S.K. Sharma, P. Mroz, M.R. Hamblin, Combination approaches to potentiate immune response after photodynamic therapy for cancer, Photochem. Photobiol. Sci. 10 (2011) 792–801. [38] A.P. Castano, P. Mroz, M.X. Wu, M.R. Hamblin, Photodynamic therapy plus low-dose cyclophosphamide generates antitumor immunity in a mouse model, Proc. Natl. Acad. Sci. U. S. A. 105 (2008) 5495–5500. [39] M. Korbelik, B. Stott, J. Sun, Photodynamic therapy-generated vaccines: relevance of tumour cell death expression, Br. J. Cancer 97 (2007) 1381–1387. [40] S.O. Gollnick, L. Vaughan, B.W. Henderson, Generation of effective antitumor vaccines using photodynamic therapy, Cancer Res. 62 (2002) 1604–1608. [41] H. Zhang, W. Ma, Y. Li, Generation of effective vaccines against liver cancer by using photodynamic therapy, Laser Med. Sci. 24 (2009) 549–552.