Extracorporeal Photopheresis in Graft-versus-Host Disease

Extracorporeal photopheresis (ECP) is a leukapheresis-based procedure used for the treatment of graft-versus-host disease (GvHD). Despite improvements in pharmacologic prevention and therapy as well as in graft manipulation technologies, GvHD continues to be a major source of disabling morbidity and mortality after allogeneic hematopoietic stem cell transplantation (HSCT) [1, 2]. GvHD is an immune-mediated, potentially life-threatening syndrome in which host tissues are attacked by donor immune cells [1, 2]. The disease is classified as acute or chronic GvHD (aGvHD, cGvHD) on the basis of different manifestations and at least in part also pathophysiologic mechanisms. Corticosteroids remain the established standard first-line treatment for GVHD, but only up to 50–60% of patients respond to this therapy and the majority of patients require additional salvage and/or second-line therapy [1, 2]. Multiple second-line treatments have been explored, but they are often not able to control GvHD activity and harbor several side effects, most importantly profound systemic immunosuppression with an increased risk of infection, disease relapse, or unacceptable late effects. Based on the substantial efficacy and the excellent safety profile in the absence of severe systemic immunosuppression, ECP has established itself as a major treatment form for steroid-refractory GvHD.

Here we review the current literature on ECP as a treatment option for patients with aGvHD as well as cGvHD.

The idea of ECP originates from the so called “PUVA” therapy, in which psoralen is taken orally to sensitize the skin and thereafter the skin is exposed to UVA light. The historical basis of this principle dates back to ancient Egypt, where people with vitiligo ingested a plant (Ammi majus) found on the banks of the Nile river, bathed in the sun, and noticed a recovery in melanin production. Nowadays, PUVA is used in the treatment of a variety of mainly dermatologic diseases, including psoriasis and neurodermitis [3].

The first clinical application of ECP was for the treatment of Sezary syndrome, the leukemic form of cutaneous T-cell lymphoma (CTCL). Edelson et al.[4] developed ECP from therapeutic leukapheresis as a “PUVA-therapy for lymphoma cells.” After their landmark study in 1987 [4] ECP received approval by the US Food and Drug Administration (FDA) in 1988. The next clinical application of ECP was for the treatment of systemic scleroderma, lupus erythematosus, rheumatoid arthritis, and other autoimmune diseases [5-7]. Chronic GvHD as a main disease entity for ECP was studied and first reported in 1994 [8]. Since then, ECP has been studied extensively in GvHD.

ECP involves the following steps: (1) leukapheresis, (2) photoactivation with 8-methoxypsoralen (8-MOP)/ultraviolet-A (UVA), and (3) reinfusion [9, 10].

For separation of mononuclear cells (MNC) from plasma and nonnucleated cells, whole blood is collected via a peripheral vein or permanently implanted catheter. The collected MNC are then exposed to UVA irradiation in the presence of the photosensitizing agent 8-MOP prior to reinfusion into the patient. Two different methods are established for this procedure (Table 1): the “closed system” (also called the in-line method) and the so called “open system” (also known as the off-line method). The closed system is a “1-step method” in which cell-separation, 8-MOP infusion, photoactivation, and reinfusion into the patient are fully integrated in one dedicated ECP device. The open system is a “2-step process” in which MNC are first collected by a separate routine apheresis device, whereafter the MNC concentrate is transferred to a second device for UVA irradiation (such as MacoGenic® by Maco Pharma or the UV-A PIT System® by PIT Medical Systems GmbH) and subsequent reinfusion.

Regulatory approval of the different systems varies depending on the respective health agency and country. The closed systems (i.e., Therakos® CELLEX), including their kits and applied drugs (i.e., UVADEX^TM), are moreover approved for ECP in many countries. The open systems on the other hand can be operated with a combination of different devices, which have not been validated for use together. Although these devices may be European CE marked or have FDA approval on their own, they are not specifically approved for combined application in photopheresis.

Closed systems rely on heparin for anticoagulation, which can be contraindicated in heparin-induced thrombocytopenia as well challenging due to its longer half-life in patients with a high bleeding risk (such as patients with a low-platelet count and systemic anticoagulation) and active bleeding. In this setting the use of ACD-A is preferable due to its shorter half-life, which is the standard anticoagulant for the open systems. Some selected centers also have gathered experience with the use of ACD-A in the closed system Therakos Cellex®; however, this application remains off-label at the moment. The mean duration per procedure is about 1–2 h, depending on the system used and according to the aimed target cell number.

In low-weight patients (children), red blood cell priming is necessary for both systems (i.e., in the Therakos Cellex® device in patients weighing <40 kg, and in the Spectra Optia® in patients weighing <25 kg). As an alternative, “mini photopheresis,” which only processes 100–200 mL of whole blood and is well tolerated, can be an option in low-weight patients [11].

Overall, both systems have proven their efficacy and are used widely in routine clinical practice, but studies for a direct head-to-head comparison will likely not be performed. Each center has to balance the benefits versus the drawbacks of each particular system, i.e., for centers with small patient numbers the open system can allow more apheresis procedures when using the apheresis device (i.e., Spectra Optia®; Terumo BCT) not only for ECP but also for other indications. On the other hand, the handling of the closed system is less demanding, it is associated with a lower risk of improper reinfusion (risk of infection), and it has fewer regulatory boundaries.

Although the first study on ECP was completed 35 years ago and since then over 2 million of treatments have been performed, the mechanism of action is still incompletely understood.

Several mechanisms have been suggested that modulate the in vivo immune response. These seem to differ between the different disease entities and currently are thought to revolve mainly around the apoptosis of lymphocytes and differentiation of dendritic cells (Fig. 1).

The first studies postulated that ECP initiates apoptosis in lymphoid cells as a result of 8-MOP binding covalently to DNA, leading to cell cycle arrest [12, 13].

Further studies investigated intensively the increased differentiation of immature dendritic cells [14], which is triggered by contact of these cells with extracorporeal surfaces in the tubing and radiation chamber of the ECP device and with platelets [15]. Dendritic cells are the master switches governing antigen-specific T-cell immunity and T-cell tolerance, and thereby they are central and critical components of the ECP immunotherapeutic response. In fact, experts have judged ECP as a dendritic antigen-presenting cell-based therapy [16].

The cytokine profile in the peripheral blood is modified after reinfusion of 8-MOP and UVA-treated cells [17], showing a switch from proinflammatory to anti-inflammatory cytokine production.

During ECP T-cells are also activated, leading to generation of regulatory T-cells which play a pivotal role in the downregulation of immune reactions. In aGvHD patients in particular, a significantly higher number of regulatory T cells is detectable in peripheral blood after ECP [18]. Gatza et al. [19] also showed that injection of ECP-treated splenocytes from mice developing GvHD triggers IL-10-producing regulatory T cells, which are able to reverse experimental GvHD [19]. Altogether, the current data suggests that ECP might be able to trigger an immune tolerance.

However, the generation of regulatory T cells does not explain how ECP selectively targets pathogenetic T cells in the setting of CTLC [20]. In this setting it has been rather suggested that ECP induces an antitumor response directed toward tumoral T cells [21].

Overall, the full concept and mode of action of ECP remain unclear and a deeper understanding of the mechanism of action would help to optimize treatment regimens.

ECP has been established in several indications that are listed by the latest American Society for Apheresis (ASFA) guidelines based on a stringent review of up-to-date literature, an analysis of the quality of evidence, and the strength of recommendations from this evidence (Table 2) [10]. Further recommendations were made recently by the section “Preparative and Therapeutic Hemapheresis” of the German Society for Transfusion Medicine and Immunohematology [22]. While the best evidence of ECP effectiveness exists for from the treatment of CTCL, in the following section we will focus on the available evidence for the use in GvHD.

ECP is used in 2 clinical settings for HSCT patients, i.e., prevention and treatment of GvHD. Although ECP is a widely recognized treatment modality for GvHD, the published data is limited to nonrandomized trials and case reports/series. Only 2 randomized clinical trials have evaluated the efficacy of ECP in cGvHD. The limited data and the heterogeneity of the patients and stem cell sources, combined with significant differences among practices between centers, pose a challenge in the evaluation of the currently available literature on the efficacy of ECP in GvHD. Furthermore, grading of GVHD has evolved in recent years and assessment of the disease itself remains challenging [23, 24]. Nonetheless, several expert groups have published consensus recommendations on treatment indications, regimens, and response assessments in patients undergoing ECP for GvHD [9, 10, 25, 26] (Table 3). These recommendations are similar and have not been prospectively studied in a head-to-head comparison. They are mainly based on expert opinion (i.e., survey in transplant centers) [27], experiences from other indications (i.e., CTLC), the frequency and interval of ECP often depending on the patient’s follow-ups, the logistics and capacity of the apheresis unit, personnel, and reimbursement.

The preventive role of ECP for GvHD has only been tested in a handful of studies, either in the application of ECP within the conditioning regimen or as GvHD prophylaxis after HSCT.

The first investigations were performed by Miller et al. [28] in 2004 by integrating ECP in a reduced-intensity conditioning (RIC) regimen together with pentostatin and low-dose body irradiation (600 cGy). Nine percent of the patients developed aGvHD higher than grade 2, while 43% of the patients developed cGvHD, with 12% showing extensive disease. The study population was heavily pretreated and many cofactors (concomitant conditioning agents, GvHD prophylaxis, donor, and bone marrow source) possibly influenced GvHD rates, thus making it difficult to draw further conclusions from this study.

Shaughnessy et al. [29] applied ECP before initiating a myeloablative conditioning regimen consisting of cyclophosphamide and fractionated total-body irradiation (10–13.5 Gy) [29]. They hypothesized that the use of ECP 4 days before the beginning of myeloablative conditioning could modulate the host antigen-presenting cell function, thereby reducing the incidence of GvHD. They matched their outcomes to historical controls with similar conditioning regimens and GvHD prophylaxis without ECP, thereby showing that the addition of ECP resulted in a significantly lower risk of aGvHD grades 2–4.

In a further study on GvHD prophylaxis, ECP was tested together with Etanercept administered twice weekly for 8 weeks [30]. ECP was performed weekly from day +28 post-HSCT to approximately day +70, followed by every other week up to day +100 and thereafter monthly to day +180. Further GvHD prophylaxis consisted of tacrolimus and mycophenolate mofetil. The results showed that aGvHD developed in 57% of the patients and there was a high incidence of steroid-dependent cGvHD once the prophylaxis was stopped, showing that this strategy was not effective.

Recently, Michallet et al. [31] performed a prospective multicenter phase 2 trial for prophylactic ECP after reduced-intensity conditioning. ECP took place twice per week during the first 2 weeks and then once a week for the next 4 weeks for a total of 8 ECP courses. The cumulative incidence was 15% for aGvHD and 22% for cGvHD, pointing towards encouraging results.

Experimentally, preliminary data also indicates that ECP could have a role in processing of donor leukocytes pretransplant. In a murine model, the exposure of tissue donor leukocytes to UVA-activated 8-MOP and their infusion into HLA-mismatched tissue recipients enhanced tolerance to transplanted heart, liver, and kidney [16]. This is a novel approach since ECP is conventionally performed after HSCT processing of donor cells. However, open issues remain a potential effect on relapse rate in malignant diseases. Further clinical trials are needed to confirm this otherwise elegant strategy, but it could prove to be an elegant prevention strategy for inducing tolerance in the transplant setting.

Overall, ECP for the prevention of GvHD has not found its way into routine use and the effect remains to be further validated in future studies.

While corticosteroids remain the backbone of first-line treatment of aGvHD, options for second-line therapy vary, with no clear advantage of one over the other. Furthermore, very little prospective controlled trial data is available to date. The choice of second-line therapy is not standardized, often relies on center-specific experiences, may be partly reimbursement driven, and can differ even within a center depending on an individual physician’s preference. The lacking standard definition of refractory aGvHD further complicates the decision and research process. The role of ECP for second-line therapy in aGvHD is less well supported by published data in comparison to cGvHD, but several reports have prompted its increased use (summarized in Table 4).

Considering that response assessment was not uniform across studies the interpretation of the results is limited. Overall response rates for steroid-refractory aGvHD range from 52 to 100%, with responses in 66–100% for skin, 40–83% for the gastrointestinal tract, and 27–71% for liver aGvHD. A multicenter comparative analysis of ECP (n = 57) versus anticytokine therapy with inolimumab or etanercept (n = 41) as a second-line treatment for steroid refractory aGvHD reported a higher overall response in the ECP group (66 vs. 32%; p = 0.001), along with a significant survival advantage for patients receiving ECP [32].

Overall, the results for ECP are not superior to results reported for alternative salvage therapies for steroid refractory aGvHD. However, as discussed before, the profile of adverse effects, steroid sparing, and organ specific responses need to be included in decision making and efficacy assessment.

As in aGvHD, corticosteroids are the standard of care for first-line treatment of cGvHD, whereas the standard for the optimal second-line treatment is not clearly defined. Compared to aGvHD, in cGVHD ECP has been studied more extensively with variable responses, although the majority of data relies on uncontrolled trials or retrospective studies. There are 3 prospective trials [33-35] and 1 recently published prospective randomized study on ECP in cGvHD based on NIH 2015 consensus criteria [32]. Hence, several expert groups have reached out to the consensus that ECP has an established place as second-line or adjuvant therapy in cGvHD.

Looking at the published data so far, the only randomized controlled trial that investigated the efficacy of ECP in patients with steroid-refractory cGvHD reported no statistically significant difference in total skin scores at 12 weeks of ECP in combination with salvage therapy versus salvage therapy alone [34]. However, patients in the ECP group showed a 40% complete and partial response compared to 10% in the non-ECP group. As a further advantage, corticosteroids could be tapered more rapidly in the ECP group.

A cross-over study after 24 weeks of ECP showed responses of 33% in skin and up to 70% in extracutaneous tissues, emphasizing the importance of prolonged ECP for an optimal therapeutic effect [35].

A systematic review showed pooled response rates for skin, liver, ocular, oral, lung, gut, and musculoskeletal steroid refractory cGvHD of 74, 68, 60, 72, 48, 53, and 64%, respectively [36].

Most recently, the first randomized controlled trial was published in the first-line setting for moderate to severe cGvHD [32], reporting results after 28 weeks of ECP treatment together with standard of care (corticosteroids) versus standard of care alone. There were no significant differences in the overall response rates between the groups (74% in the ECP group vs. 60.9% in the standard-alone group, by blinded assessors), but a trend of improvement of quality of life in the ECP group was observed. Further clinical studies are needed to understand the role of ECP in the first-line setting.

Numerous small studies have been published to address the role of ECP in cGvHD, showing varying disease responses. Table 5 summarizes the studies on the effect of ECP in second-line treatment of cGvHD, though response assessment was also not uniform across all studies.

Current consensus recommendations point out that treatment should be performed on 2 consecutive days every week or every 2 weeks for at least 8–12 weeks or until a response is noticeable (Table 3). The most recent study by Cid et al. [32] showed that their center-specific off-line ECP schedule was also efficacious, with overall response rates of 57% for aGvHD and 38% for cGvHD. The authors treated patients with aGVHD twice a week for the first 2 weeks, followed by once a week for 2 more weeks. After the first month of treatment, the patients received treatment once every 2 weeks for a minimum total of 16 ECP procedures. Patients with chronic GVHD were treated once a week for 4 weeks followed by once every 2 weeks for a minimum total of 14 ECP procedures. This data highlights that the optimal treatment schedule is still not clear, and most likely several regimens have to be tested ideally in a head-to-head comparison.

ECP therapy is characterized by an excellent safety profile, with almost no serious adverse events reported so far as long as patients are carefully selected [37]. The main side effects are mild and comprise the increased photosensitivity from 8-MOP and patient-related issues due to repeated venous puncture and volume shifts during the procedure. There are no reports of an increased infection risk or disease relapse, although a warning has been issued for splenectomized patients.

Due to increased photosensitivity, patients are instructed to wear eye and skin protection for 24 h after ECP treatment and ECP should not be performed in patients with aphakia due to an increased risk of retinal damage [38].

ECP relies on either repeating peripheral venous puncture or long-term venous access (peripheral or central), which can result in local hematoma, arterial puncture, phlebitis, venous thromboembolism, catheter-related infection, and pneumothorax.

With the closed Cellex® system in particular, venous thromboembolism has been reported, as noted in a 2018 MedWatch Safety Alert issued by the FDA [39]. These events are, to date, not fully understood; nonetheless, physicians should inform their patients on the potential risk during ECP treatment.

Other side effects observed during ECP are primarily related to volume shifts since hypotension may occur during any treatment involving extracorporeal circulation. This is of major importance for low-weight patients including pediatric patients, since priming of the machine may be necessary.

Absolute contraindications for ECP exist for patients who cannot tolerate methoxsalen, show hemodynamic instability, have an uncontrolled systemic infection, and are affected by coagulation disorders with a high bleeding risk.

ECP centers should be quality assured. ECP requires well-maintained and validated machines, specifically trained staff, and a quality management system to ensure that the clinical standards for safe, high-quality procedures are achieved and maintained. A quality control of the collected cells is recommended (i.e., cell count and cell type) and can be collected not only by users of the open system but also by users of the closed system [40].

Patients affected by cutaneous GvHD often show sensitive und vulnerable skin features, and thus multiple insertions of peripheral intravenous catheters – as are needed for long-term ECP – can be cumbersome. In addition, peripheral vein access can be lost after several ECP cycles because of contractures, skin sclerosis, or steroid effects. In this patient group with poor peripheral venous access we observed a favorable outcome by insertion of central venous catheters, either by tunneled hemodialysis catheters or by special ports.

ECP candidates with GvHD commonly present with abnormalities in their peripheral WBC count (leukopenia as well leukocytosis). To date, there is no consensus addressing the limitation of ECP in cases of leukopenia, lymphopenia, monocytopenia, or leukocytosis. Some centers do not perform ECP in cases with a very low WBC (<1.5 G/L). Specific limits are arbitrary and have not been validated. One study demonstrated that peripheral and buffy coat lymphocyte and monocyte counts strongly correlate with each other [41], suggesting that the peripheral leukocyte count can function as a surrogate for the cell dose treated per ECP. However, taking into account the limited data, placing a WBC dependency on the decision to start ECP still needs to be left in the responsibility of the individual medical center.

GvHD patients can be affected by thrombocytopenia, either caused by GvHD itself or due to other causes. Severe thrombocytopenia (<50 G/L) can be challenging considering that patients receive anticoagulation with heparin or citrate during the ECP procedure. In our center the following scheme for ECP with heparin (10,000 E) is standard for this patient group with a high bleeding risk: platelets <20 G/L: 1 platelet concentrate directly before the start of ECP and protamine 3,000 E directly after ECP; platelets <50 G/L: protamine (3,000 E) directly after ECP.

This scheme has not been scientifically evaluated so far; however, to date, we have not observed severe bleeding events in GvHD patients with this approach. Alternatively, citrate can be considered with selected devices due to its shorter half-life (see Techniques and Systems).

Besides thrombocytopenia, GvHD patients also occasionally receive long-term systemic anticoagulation due to other indications (e.g., previous venous thromboembolism and pulmonary embolism). Concomitant anticoagulation during ECP can be challenging and has to be evaluated for every patient individually based on the properties of the prescribed anticoagulant, renal function, and intake of further concomitant therapies with potential for interactions.

Since the first trial on the use of ECP in CTCL was performed by Edelson et al.[4] in 1987, promising results for this therapeutic procedure in various conditions have been published. Today, ECP is an established immunomodulatory treatment in several disease entities.

Current data in GvHD is particularly encouraging, suggesting that ECP leads to no or only minor impairment of the individual immune response to pathogens and the underlying disease in comparison to other GvHD treatments. Due to the beneficial results and the favorable safety profile, ECP is continuously investigated in several experimental and clinical studies regarding the underlying pathophysiological mechanisms and clinical effects.

The current “one size fits all” strategy is rudimentary and may not be the optimal treatment approach [16]. With increasing knowledge, numerous research questions arise in ECP and require validation, i.e., How many lymphocytes and dendritic cells are required for an efficacious therapy? How much whole blood should be processed (i.e., mini-ECP) ? Are other blood components (e.g., platelets and coagulation factors) needed for the best efficacy? What is the optimal dose of 8-MOP? How long should photoactivation of cells take place? Is UVA light the best energy source for inducing the mechanisms of action? How often and in which sequence should procedures be performed for maximizing the therapy effect? Which patients would profit most from ECP treatment (personalized therapy)?

To answer these questions, further experimental research and prospective controlled multicenter trials are needed, ultimately defining the ideal role of ECP in the treatment of GvHD.

The authors have no ethical conflicts to disclose.

The authors have no conflict of interests to declare.

The authors declare that no funding relevant to this paper exists.

B.D. and A.H. reviewed and interpreted current literature and drafted, wrote, and approved this paper. A.B., L.I., G.S., and J.H. interpreted the data and revised and approved this work.