Summary
Extracorporeal photochemotherapy (ECP) has been used worldwide in more than 150 centers since its first application in the treatment of cutaneous T-cell lymphoma (CTCL) by Edelson et al. [1] in 1987. It was the first selective immunotherapy for a cancer (CTCL) approved by FDA in United States. Based on its efficacy and safety in the treatment of CTCL, ECPs application was extended on other human pathologies: rheumatoid arthritis [2], systemic sclerosis [3], systemic lupus erythematosus [4], pemphigus vulgaris [5] and more recently in solid organ graft rejection especially in acute cardiac graft rejection [6] and in graft versus host disease [7].
In 1990 we proposed a new technique [8] in order to improve the quality of ECP treatment and we applied it for the first time in the treatment of refractory rheumatoid arthritis [9]. Recently Edelson [10] proposed a new method of ECP transimmunisation in which UVA/8-methoxypsoralen (8-MOP) treated cells are incubated over one night before reinfusion to the patient.
The exact mechanisms of action of ECP are still elusive. During the collection of peripheral mononuclear cells, cells environment modifications have been suggested to increase monocytes activation and possibly to induce their differentiation into dendritic cells (DCs) [10]. UVA irradiation in the presence of 8-MOP is presumed to induce cell membrane damage, DNA crosslinking and binding to a variety of cytosolic proteins leading to apoptosis [11], modification of membrane antigenicity [12] and antigen presenting cell (APC) activation. While it is unclear what exactly occurs in vivo, it is thought that DCs play a critical role by inducing an immunological response against pathogenic cells [13]. The immature DC, activated by ECP, phagocytizes and internalizes the apoptotic cells, processes the antigens and increases the synthesis of classes I and II major histocompatibility complex (MHC) molecules [14], [15]. The peptides associated with class II MHC are presented to the CD4+ T helper cells. The final maturation of DC is completed in vivo with the help of these activated T helper cells using a variety of mechanisms including CD40 ligation [16]. Finally, the mature DCs fully loaded with pathogenic T-cell peptides migrate to secondary lymphoid organs, stimulate the naive CD8+ T cells and induce a cytotoxic response directed against pathogenic clones (Sezary cells). DC can also control an existing pathogenic immune response by down regulation. Recruitment and involvement of other immune cells in the mechanism of ECP have been suggested and merit more studies.
In some cultures, the sun was venerated as a source of life in ancient time. Five thousand years ago, the ancient Egyptians ingested the leaves of ammi majus, a plant that grew by the Nile river coasts, to treat depigmented areas of skin. They also knew that the sunlight was essential for this treatment. It was probably the first PUVA therapy for Vitiligo (but they did not know how to patent it).
In 1948, El Mofty [17] isolated and characterized the active ingredient of the ammi majus plant, 8-MOP; using the purified compound he showed that exposure of the skin to sunlight after its ingestion led to repigmentation of depigmented skin.
Many years later (in 1979) Parrish et al. [18] developed PUVA therapy for the treatment of Psoriasis (UVA skin radiation 2 h after psoralen ingestion).
Gilchrest [19] showed that PUVA was effective in the treatment of CTCL.
Finally Edelson reasoned that the direct exposure of the diseased T cell of CTCL patients to 8-MOP and UVA might enhance their therapeutic efficacy. The first generation prototype system of photopheresis was born. The first publication of Edelson et al. [1] in 1987 showed that photopheresis could induce a full remission in some of CTCL patients and improvement in others. Heald et al. [20] in 1993 reported the follow-up of the patients treated first by Edelson in 1987. The median survival of ECP treated patient was longer than the historical control (60 month compared with 33). Most of the other reports suggest that survival may increase but none of them has been randomised.
The first device was manufactured by Therakos company (Johnson & Johnson, PA, USA) see article of F. Schooneman in this issue.
At first, the aim of the treatment was to destroy maximal quantity of the blood CTCL cells by apoptosis induced by UVA and 8-MOP, only 5–10% of the whole blood mononuclear cells (MNCs) were treated in each procedure. Interestingly Edelson found that after elimination of the treated cells in blood, the number of non-treated malignant cells in blood and also in skin decrease considerably, suggesting that the treatment induce a systemic anti-tumoral action, this was confirmed by the other American and European centers.
CTCL is a clonally derived skin-invasive malignancy of CD4+ T cells with mature helper T-cell phenotype. In advanced stage it is associated with depressed cell-mediated immunity, increased secretion of Th2 cytokines and decrease level of Th1 cytokine secretion and NK cell activity [21]. In 1988 photopheresis was approved by FDA in United States as the first selective immunotherapy for a cancer (CTCL).
Photopheresis consisted of oral administration of 8-MOP, 2 h before leukapheresis. The MNC of the patient (including tumoral cells) was then collected by the UVAR apparatus (Therakos) using a discontinuous flow, and by the same apparatus the MNC were exposed to UVA radiation. At the end of the procedure, whole UVA treated cells were reinfused to the patient. The schedule of treatment was two consecutive days photopheresis each month for several months or years.
This was the only method available for many years. In 1990, in France we proposed another method [8] that we called ECP. In ECP, MNC are collected by Cobe Spectra (Gambro, Denver, CO, USA), then the collected cells are transferred sterilely into a special UV transparent bag (Macopharma, Tourcoing, France) and a hydrosoluble 8-MOP is directly introduced into the MNC containing solution. This was the first time that an injectable 8-MOP was prepared for ECP use [22]. UVA exposure is performed by Vilbert-Lourmat device (Vilbert-Lourmat, Torcy, France). Then the UVA treated cells are reinfused to the patient.
ECP improved photopheresis method in many point of view [8]; 8-MOP concentration in the irradiation bag is well controlled (we use a concentration of 200 ng/ml of 8-MOP) whereas orally administration of 8-MOP is associated with an important variability of its plasmatic concentration (0–500 ng/ml). 8-MOP administration in the bag does not expose the patient to the side effects of oral administration. The total number of whole leukocyte and MNC harvested is considerably higher in ECP. The MNC are less contaminated by polymorphonuclears (less than 2%) and red blood cells (less than 3% of hematocrit). The quality of UVA irradiation seems to be also better because of the energy monitoring of the Vilbert-Lourmat system and a better control of red cell concentration in the MNC bag. Finally the price of ECP is lower than photopheresis. These advantages led many investigators in Europe to use this new method.
We performed ECP with a different schedule, two procedures for three weeks then two procedures for four weeks and decreasing the frequency to finally one procedure per month.
The first clinical experience of this new method (ECP) was successfully performed in patients with rheumatoid arthritis [2].
Efficacy of ECP in the treatment of Sezary and CTCL supplemented by experimental trials with its low-side effect profile encouraged investigators to use it in other pathologies. Acute cardiac graft rejection was treated successfully by Costanzo Nordin et al. [54]. Other investigators confirmed later that ECP is a safe and efficient adjuvant therapy for the episodes of acute cardiac graft rejection especially when the second line treatment fails. Barr et al. [23], confirmed in a multi-centric randomised European prospective study in the patients undergoing cardiac graft, the efficacy of ECP in the prevention of rejection episodes in cardiac transplantation.
We reported the first successful treatment of acute lung graft rejection by ECP [24] and the use of ECP in the treatment of obliterans bronchiolitis, which is a manifestation of lung chronic graft rejection, in another 11 patients [25], [26].
Other solid organ allograft rejections were also treated successfully by ECP like renal graft rejection [27], and finally graft versus host disease [28].
Experience has been gained in autoimmune diseases such as rheumatoid arthritis [2], systemic sclerosis [3], systemic lupus erythematosus [4], pemphigus vulgaris [5], psoriatic arthritis [29], and recently in atopic dermatitis [30], Crohn disease [31] and chronic erosive lichen planus [32].
Animal experimentation has been used to study the mechanisms and efficacy of ECP in the treatment and prevention of autoimmune diseases and pathogenic allo-immune responses. First, Cohen [12] described the concept of “T-cell vaccination”. It exists some physiological mechanisms of non-expression of auto-immune diseases despite the presence of the auto-reactive T lymphocytes in normal individuals. Experimental models have shown that this regulatory system depends on many different actors such as the presence of T lymphocyte clones whose activity is directed against auto-reactive T lymphocytes. It has been demonstrated that the induction of such reactions “against autoimmunity” can be performed in vitro by different methods of manipulation of autoreactive lymphocytes; ECP is one of the “T-cell vaccination” methods.
This concept of T-cell vaccination is valuable, not only for auto-reactive, but also for all activated T lymphocytes including allo-reactive T cells. Experimental auto-immune encephalomyelitis (EAE) can be induced in rats by the injection of myelin basic protein associated with Freund adjuvant with which the animals develop an acute paralytic syndrome. The T cells of these animals, when injected to normal syngeneic rats, can transmit the same paralytic syndrome. Ben Nun et al. [33] have demonstrated that it was possible to prevent the EAE if the recipient animal had received previously (by IV injection) membrane-modified anti-myelin basic T-cell clones. This prevention was specific for the EAE. Membrane modifications can be achieved by ECP.
For the prevention or treatment of the auto-immune diseases by ECP, animal strain susceptible to develop an auto-immune disease spontaneously or after stimulation have been used, spontaneous lupus disease of the MRL/1 mice can be prevented by ECP [34]; adult MLR/1 spontaneously develop an auto-immune disease with lymph-node enlargement, splenomegaly and presence of anti-DNA antibodies, because of the existence of abnormal auto-reactive T-cell lymphocyte proliferation. This pathology can be prevented if the young mice have previously received the T cells treated by ECP of an adult ill mouse. It seems that the treated mice recognise the T-cell auto-reactive population as abnormal and therefore suppress it. If this treatment is applied to old mice, the DNA auto-antibodies rate decrease and lymph-node hyperplasia no more appears.
ECP prevents also the collagen induced arthritis in DBA/1 mice [35].
The prevention or treatment of graft rejection (allo- or xeno-grafts) has been studied in several animal species. The first experiment for the prevention of allo-graft rejection was performed by Perez et al. [36]: they used ECP for the prevention of skin allo-graft rejection in BALB/c mice: the skin of CBA/j mice, engrafted to non-pre-treated BALB/c mice, was rejected after 7 days, but BALB/c mice pre-treated by ECP rejected the skin allo-graft only after 42 days.
Pepino et al. [37] used ECP successfully for the prevention of cardiac allo- and xeno-graft rejection in primates. Baboon apes were heterotopically engrafted with cynomolgus hearts. The recipients had previously received corticosteroids and CsA during two days, before engraftment. Three days after engraftment, a series of six baboons were treated by PCE, two times a week. Active lymphocytes were obtained by xeno-stimulation of the Baboon apes by the injection of cynomolgus blood cells. The graft survival was prolonged to 38 days in the ECP group versus 26 days in the control group (corticosteroids and CsA without ECP).
In our own experiments, we showed that ECP can prevent cardiac heterotopic allo-graft rejection in rats.
Heterotopic heart transplantation was performed between Lewis rat strain as recipient and dark agoutti rat strain as donor. The rejection was delayed to 50 days in the ECP pre-treated rats versus 13.5 days in the control group [38].
Mechanisms of action of ECP have been studied in vitro, in animal model experimentation and in clinical studies. The majority of the clinical studies have been performed during the treatment of CTCL and Sezary with more than 200,000 procedures worldwide. Despite increasing available data, ECPs mechanism remains still elusive. A better understanding of its mechanisms will be useful to assess the justified indications, to modify the existing techniques and also to select synergistic associated treatments.
During the leukapheresis step of ECP several cell environment changes occur (temperature decrease, centrifugation, cell shape changes, plastic contact, and pH modification). These conditions may have a variety of effects such as to inhibit mRNA and protein synthesis. Several of these have been suggested to increase monocyte activation and possibly drive DC differentiation [39]. The second step of ECP is the cell irradiation by UVA in presence of 8-MOP which is presumed to induce cell membrane damage, DNA crosslinking [40] and binding to a variety of cytosolic proteins leading to apoptosis [41], modification of membrane antigenicity and APC activation. Two initial important events after reinfusion of phototreated cells are apoptosis and monocyte differentiation to DC. Apoptosis appear about 24 h later in T lymphocytes by DNA crosslinking, mitochondrial dysfunction, caspases activation and other cell damages by ECP but the monocytes are resistant to apoptosis [41]. In the other hand, the number of immature DCs increases after ECP.
DCs are the professional APCs; they are located strategically at the interface of potential pathogen entry site and take up antigen, move into secondary lymphoid tissues and activate both helper and cytotoxic T cells (naive and primed). They also interact with B cells and NK and NKT cells [16].
Apart from pathogen protection, they have roles in central and peripheral tolerance and anti-cancer immune responses. They are implicated in cellular and humoral immune responses [42]. They have also the capacity of high expression of MHC classes I and II molecules, adherence molecules, co-stimulatory molecules and secretion of chemokines and IL-12. Recently another function of DCs, the secretion of exosomes has been discovered [43]. Finally DCs, have an important capacity of migration and functional maturation (Fig. 1).
ECP increases the quantity of antigenic peptides and MHC class I (200–300%, 20 h later) at the cell surface [15].
In leukapheresis product of ECP, different type of DC or their precursors can be identified; Immature blood DCs (CD11c+) and the precursors of DCs: pDC1; giving rise to monocyte derived DCs cell (MoDC), or type one DC (DC1) after 5–7 days culture of monocytes with IL4 and GM-CSF, and pDC2 (CD4+CD3−CD11c−plasmacytoid derived DC) after IL3 stimulation [44]. The main source of DC is therefore the monocytes, because the number of immature DC in the blood is very low and plasmocytoid DC2 are mainly located within T-cell areas of lymphoid tissues and are specialized to recognise self-antigens or blood-born pathogens, in the blood they are only in passage.
It is known that DC1 induce Th1 immune response and DC2 induce Th2 immune response, and a good anti-tumoral immune response (against CTCL) is the Th1 type, specially by stimulation of CD8 cytotoxic T-cells specific for tumor associated antigen (CTL) [45]. CTCL and Sezary cells are CD4+ T lymphocytes with Th2 activity [21]. In vitro stimulation of purified blood MNC (PBMC) of these patients by PHA show an increase of Th2 cytokines (IL4, IL5, IL10) and a decrease of Th1 cytokines (IL2, IFNγ), and after stimulation by LPS, it show an decrease of IL12 (secreted by DC1). Many investigators report that in Sezary patients clinical and haematological improvement after ECP are associated with a shift in Th1/Th2 balance and the increase of Th1 cytokines and IL12 [21]. Natural killers (NKs) are also decreased in Sezary patients blood and ECP can increase its number [46]. Thus ECP can enhance an anti-tumoral response by Th1 type specific immune response and by NKs.
After ECP, the activated immature DCs or MoDCs phagocyte the apoptotic Sezary cells and process tumoral antigens and begun their maturation. MHC classes I and II synthesis increase, TAA loaded MHC II, co-stimulatory molecules and receptors for migration chemokines are expressed on the cell surface, DC migrate to secondary lymphoid system and present MHC class II loaded TAA to T helpers (CD40/CD40L ligation is essential) then TAA is associated to MHC I in cytosolic compartments (crossing over) and expressed on cell surface. CD8 cytotoxic T cell (naive and primed) are stimulated by DCs (co-stimulation by CD80 and CD86 are essential) as well as NKT cells. NK proliferation is also enhanced by DCs chemokines (Fig. 2). The relevant antigens may be T-cell receptor (TCR) derived. Each malignant clone of T cells expresses a TCR unique for its clone (see article of Oro in this issue).
A given DC subset has a remarkable plasticity in directing different types of T-cell responses under different environment and activation factors [45]. DC1 at mature stage induce Th1 differentiation, but at an immature stage induce IL-10-producing CD4+ and CD8+ regulatory T cell (Fig. 3).
The ability of ECP to initiate an immune response against a single clone, or multiple clones of normal T cell may explain in part the capacity of ECP to down regulate aberrant immune response in autoimmune diseases and graft rejection and GvHD. Classic hypothesis for the mechanism is the induction of an anti-idiotypic response by CD8 suppressors based on prevention of EAE by ECP experience [32]. Recent data indicate the presence of allo or auto immune response (CTL, Th1) in these pathologies. EPC down regulates the immune response, induces tolerance by regulatory T cells and controls allo or auto immune responses.
The common denominator between these diverse groups of responding patients is the presence of clonally distinctive TCRs on the malignant or allo–autoreactive T-cell clones [47].
Girardi et al. [48] suggest that in the site of interaction, the monocytes take up cell debris of graft, goes to blood circulation, under ECP, these cells induce the immature DCs; DCs process and present the graft specific peptide to Th cells and induce a Th2 response and stop allogeneic Th1 response.
The review of 31 studies where ECP was used in the treatment of acute and chronic GvHD (76 patients with acute GvHD and 204 with chronic GvHD) indicates that ECP is a non-aggressive treatment that may benefit patients with both acute and chronic GvHD who do not respond to standard immunosuppressive therapy [7].
Immunopathogenesis of cGVHD is, in part, Th2 mediated [49], resulting in a syndrome of immunodeficiency and an autoimmune disorder. The most important risk factor for cGVHD is prior history of aGVHD, use of a non-T-cell-depleted graft, older age of donor and recipient. The strategies that prevent aGVHD also decrease the risk of cGVHD. Newer therapeutic strategies under investigation include thalidomide, anti-TNF and B cell depletion with anti-CD20 monoclonal antibody and specially ECP are promising [7].
A French study of cGVHD treated by ECP in children (see article of Kanold in this issue) confirm also the safety and efficacy of this treatment (254 procedures in eight children).
In acute GvHD, Th1 immune response is dominant [50] while in chronic GvHD Th2 response dominates. ECP has been used successfully in the treatment of both acute and chronic GvHD.
Different results have been reported by Gorgun et al. [51]: clinical response to ECP in 10 patients with chronic GvHD was associated with normalization of CD4/CD8 ratio, increase in CD3−/CD56+ NK cells and decrease in circulating DCs, T-cell proliferation (autologous and allogeneic) in MLR. A shift from DC1 to DC2 and a shift from predominantly Th1 (IL-2, IFNγ) to Th2 (IL-4, IL-10) was also observed.
ECP appears to induce tolerance to alloreactive or autoreactive antigen-generated T-cell responses.
The pathogenesis of cGvHD is controversial, either an extension of acute alloreactivity seen in aGvHD, or a dysfunctional immune reconstitution with generation of tissue auto-reactive T-cell clones and dysregulation of CD4+ CD25+ immune modulating T cells.
In murine GvHD models, Th2 phenotype donor CD4-enriched cells prevent GvHD without affecting engraftment (down regulation of Th1 immune response by Th2 cells).
In allograft, DCs induce a T-cell response against foreign (allogeneic) MHC molecules and both CD4 and CD8 alloreactive cells are involved in graft rejection.
Both donor and recipient APCs are involved in graft rejection and the most important APC involved in graft rejection are DCs, either of donor origin (resident in the interstitium of graft) or of recipient entering the graft through the blood circulation. Both these DCs may stimulate recipient T cells. Donor DCs can also migrate into draining lymph nodes and activate naive T cells of recipient by direct pathway.
It is also possible that donor DC2 present the host antigens to donor T cells and induce Th2 responses and attenuate GvHD.
Interestingly, umbilical cord blood stem cells graft containing DC2 but not DC1 subsets is associated with low incidence of acute GvHD.
Donor DCs engraftment in recipient of allogeneic transplant has been associated with prolonged organ transplant survival. Donor DC1 may activate host T cells to produce IFNγ and CTL against graft favoring rejection, but donor DC2 induce host T cell to produce IL-4 and IL-10 which suppress Th1 and CTL response favoring engraftment.
The balance of Th1 cytokines (IL-2, IFNγ) and Th2 cytokines (IL-4, IL-10) governs the extent to which a cell-mediated immune response or a systemic inflammatory response develops. The DC2 inducing Th2 clones expansion secretion IL-4 and IL-10. These cytokines produce a negative feedback on Th1 differentiation and terminate the immune and inflammatory response (Fig. 4).
Th2 cytokines can inhibit the production of pro-inflammatory cytokines IL-1 and TNFα, a shift of Th1 to Th2 of initial response of donor T cells may prevent and treat aGvHD [52].
In ECP, two synergistic phenomena occur in parallel after the reinfusion of phototreated cells to patient: apoptosis of lymphocytes [11] (including CTCL cells), and monocytes differentiation to DC [13]. The tumor associated antigen (TAA) could be presented to immune system by DC after phagocytosis of tumoral apoptotic cells and the process of TAA and its presentation on the DC surface in association with MHC class I. Lymphocyte apoptosis and monocytes differentiation to DC occur about 24 h after ECP procedure in blood and the contact of these tow type of the cells is not optimal in blood circulation. Edelson [10] and his colleges invented once again a modified method of ECP. The process was renamed transimmunisation (TI) by Edelson, since it is based on the transfer of tumor (or other clonally relevant antigens) to cells capable of immunizing the patient against these distinctive antigens.
The TI modification of ECP is a means of inducing DC differentiation from Peripheral blood monocytes in the presence of apoptotic tumor cells. The immature DCs uptake the apoptotic CTCL cells and in the presence of inflammatory cytokines derive to maturation and potent APC. TI employs the overnight incubation of ECP treated leukocytes in a platelet storage bag in autologous plasma and tissue culture media. The monocytes are activated by centrifugation and passage trough the 1 mm channels of UVA irradiator. The plastic adhesion and membrane perturbation derive the activated monocytes to initiate their transformation to DCs (MoDCs). The CTCL cells after ECP treatment begin to undergo apoptosis during overnight incubation. During this incubation an aggressive phagocytosis of the apoptotic cells by the DCs occurs. Up take of apoptotic cells and presence of necrotic tumor cells contribute to DCs maturation. Reinfusion of theses tumor-loaded DCs has the potential to invoke an anti-tumor response in the recipient by the CD8 cytotoxic lymphocytes (CTLs).
The CTL have the capacity to lyse tumor cells that display the relevant peptide through the release of granules containing granzyme and perforin or a Fas-related apoptotic mechanism [10].
Activated monocytes (as well as damaged cells and dying cells) releases the cytokines (IL1β, TNFα and IL6) that promote their differentiation into mature DCs [13].
In the other disorders treated by ECP (autoimmune diseases and graft rejection), the reinjection of immature DCs may play a role in down regulation of pathogen immune response by induction of CD25+, CTL-4 regulatory T cells and IL10 production and expression of CTLA-4 (a molecule that down regulates co-stimulation). The immature DCs that take up and present self Ag can promote tolerance also (by induction of regulatory T cells) [13].
Another factor for induction or suppression of immune response is the affinity of the TCR for Ag (high affinity TCR) induce unwanted immune response while immature DCs presenting self Ag can promote T cells with low affinity TCR and silence pathogen immune response.
At first ECP was used as monotherapy in the treatment of CTCL. But a synergistic effect with other treatments such as interferon-α, methotrexate total skin electron beam therapy, PUVA and retinoides (bexarotene) is possible. Most widely used associated treatment with ECP is IFN-α, Rook et al. [46] reported in 1991 that combination of ECP and IFNα-2b has a synergistic therapeutic effect in the treatment of refractory CTCL patients. It was confirmed in 1997 by Dippel et al. [53].
Evidence is that the combination with ECP is not associated with increased side effects or negative effect in the efficacy of concurrent treatment.
Now, after 25 years experience, extension in more than 150 centres and two method “lifting” we begin to understand how ECP works and what are the best indications are. However, many further studies are still required:
•Most clinical studies already published are limited to a small number of patients, without rigorous comparison with other existing therapies via randomisation; we need to develop such randomised trials to assess the actual role of ECP in the immune modulation of allo and auto-immune pathological processes.
•The procedure of cell treatment needs important modifications to improve its efficacy: these modifications will be introduced according to experimental data giving solid information on the possible mechanisms of action of the technique: they include the use of cytokines, a more relevant selection of cells to be treated, as well as the use of new photoactivable components.
In France, we have recently set up the French Group of ECP supported by the French Society of Apheresis, with the participation of all French centres and a large European and international collaboration. Our goal is to gather all our means to improve our knowledge of this technique. The last national congress of the French Society of Hemapheresis was largely devoted to ECP. In this issue you find some of the subjects presented in this congress.
FULL TEXT
