Background
Vaso-occlusive crises (VOC) are important features of sickle cell disease (SCD), which is characterized by ischemic injury of potentially all major organs in the body [
1‐
3]. Tissue damage initially results from hypoxia and then from oxygen re-exposure, that causes massive production of reactive oxygen species (ROS), local activation of endothelial cells, platelets, neutrophils, monocytes, resident tissue macrophages and perivascular mast cells [
4]. Pro-inflammatory mediators derived from activated cells create a positive feedback that amplifies the vascular/tissue damage and inflammation [
3,
4]. In about 50% of the adult patients, ischemic lesions evolve to avascular osteonecrosis often located in the femoral and humeral heads [
5,
6].
Another important aspect of SCD is the chronic hemolysis occurrence [
3,
7]. Intra-erythrocyte ADP and ATP released in the extracellular space contribute to vaso-occlusion and inflammation [
7‐
9]. Similarly, free hemoglobin is a powerful NO scavenger and promotes ROS accumulation [
3,
10]. In addition, the presence of the prosthetic heme group in plasma activates the coagulation system and innate immunity [
11]. Heme acts as a ligand of the Toll-like receptor 4 (TLR4), with activation of two main pathways in endothelial cells [
12]. The first one is the protein kinase C-mediated mobilization of Weibel–Palade bodies (WPBs) to the cell surface, simultaneously releasing the pro-coagulant von Willebrand Factor and loading P-selectin onto the cell membrane to promote leukocyte-endothelial binding and stasis [
11]. The second pathway is MyD88-mediated NF-κB activation and subsequent transcription of several responsive genes, notably those encoding pro-inflammatory mediators, such as: IL-1, IL-6, IL-8, VCAM-1, ICAM-1, P-selectin and E-selectin [
4,
11‐
13].
For decades, nonsteroidal anti-inflammatory drugs (NSAID) have been recommended for treatment of light and moderate pain in SCD patients [
1,
14]. Similar to NSAID, low-dose methotrexate (MTX) exhibits anti-inflammatory activity that has been successfully used for therapy of several chronic inflammatory diseases, such as rheumatoid arthritis [
15], psoriasis [
16], uveitis [
17], juvenile dermatomyositis [
18], localized scleroderma [
18], Crohn’s disease [
18], Wegener granulomatosis [
19], and sarcoidosis [
20]. This study tests the hypothesis that the inflammatory component of sickle cell disease can be responsive to methotrexate treatment, with clinical improvement in frequency and pain intensity of VOC episodes.
Discussion
Ten out of fourteen patients exhibited some degree of response to methotrexate, ranging from mild pain improvement (Patient #14) to dramatic outcomes (Patient #8, who had a 75% MPI drop and over 90% reduction of his osteonecrosis pain). The best responders were often among those with osteonecrosis (Patients #1, #3, #4, #8, and #9). Furthermore, pain relief was accompanied by functional gain, most notably in those who had osteonecrosis. It is worth mentioning that the SF-36 physical functioning subscale used in this study has also been applied as a stand-alone assessment tool in similar conditions, that may involve physical limitation and pain, including peripheral arterial disease and hip fracture [
30,
31]. The disappearance or substantial reduction of osteonecrosis-associated pain in five out of seven patients was remarkable, and occurred even when vaso-occlusion crises have not subsided. Three of the five individuals with osteonecrosis who responded to methotrexate (Patients #1, #3, and #4) reported persistent pain before the MTX treatment, with virtually no pain-free intervals. A therapeutic alternative for this complication is highly desirable, given that its low responsiveness to opioids leaves little to offer to the patients other than surgical replacement of the affected bone area [
6,
27,
32].
The use of MTX was associated to the patient’s perception of occurrence of longer uninterrupted pain-free periods in between crises. It is not surprising that any pain-free period produced by MTX use would be seen as a tremendous gain, notably for those that previously had continuous chronic pain.
Although not evaluated in this study, the MTX beneficial effect is likely to be independent of a mechanism of reduction of sickling episodes similar to that induced by hydroxyurea [
33,
34]. Methotrexate did not have, in this study, any significant effect on hemoglobin F levels and reticulocyte counts.
Rheumatoid arthritis and sickle cell disease are similar in that they both have an inflammatory component [
4,
11‐
13,
15]. Rheumatoid arthritis is an autoimmune disorder associated with significant morbidity, caused by immune-mediated destruction of synovial surfaces and progressive joint damage [
15]. The use of low-dose aminopterin as a tissue reactive suppressor was attempted for the first time close to 75 years ago in seven arthritis patients, with symptom improvement reported in 6 individuals [
35]. At the time, the immune modulatory effect of aminopterin was not known and its effectiveness on pain control despite the lack of analgesic activity was puzzling. Aminopterin was soon replaced by the closely related molecule methotrexate that became the mainstay of the rheumatoid arthritis treatment [
36]. The potentially beneficial effect of methotrexate in sickle cell disease can be inferred from a few clinical reports in which it was used to treat co-existing juvenile rheumatoid arthritis [
37,
38]. MTX is an antifolate that acts by inhibiting folate-dependent enzymes, one of the most affected being AICAR (5-aminoimidazole-4-carboxamide ribonucleotide) transformylase, resulting in accumulation of AICAR [
39]. AICAR inhibits AMP deaminase directly and the AICAR dephosphorylated form inhibits adenosine deaminase, leading to net tissue accumulation of AMP and/or adenosine [
36].
Extracellular AMP is converted into adenosine by the ecto-5′-nucleotidase CD73 expressed on the endothelial surface [
40]. Adenosine signals through G-coupled receptors (A
1R, A
2AR, A
2BR, and A3) and is a powerful immune modulator to different cells known to participate of the inflammatory response in rheumatoid arthritis and sickle cell disease, such as M1 macrophages and neutrophils [
4,
13,
41]. Adenosine down-regulates pro-inflammatory cytokine production in macrophages and decreases ROS generation and phagocytic activity in neutrophils, which also lose their selectin- and integrin-mediated adhesion to the endothelium [
41]. It is noteworthy that adenosine immune modulation plays a protective role in several animal models of ischemia/reperfusion injury involving different target organs [
9,
42]. A similar injury mechanism has been suggested to occur in sickle cell disease [
4]. One could envisage that MTX-induced adenosine and its AMP precursor would leak into the local microcirculation from multiple cell types and counteract the pro-inflammatory effect of the products released during hemolysis.
A decrease in plasma levels of TNF-α, the master inflammatory regulator, was observed in weeks 6 and 12 of the MTX treatment, with a more pronounced down-regulation in the middle of the study but it did not reach significance in either time point (
P = 0.075 for week 6 and
P = 0.899 for week 12; Additional file
1: Figure S4A). Conversely, up-regulation of two chemokines, CXCL10 and CXCL12, was detected. Although CXCL10 is expected to promote inflammation by attracting CXCR3-positive immune cells to active sites [
43], it was not possible to find an augmentation of inflammatory effector molecules after MTX treatment. A possible explanation is that methotrexate-induced extracellular adenosine blunts CXCL10 chemotactic activity through A
2a receptor activation and heterologous desensitization of CXCR3 [
44‐
46].
The up-regulation of CXCL12 may be a central point to the understanding of the MTX clinical impact on SCD patients. CXCL12 is a powerful angiogenic factor that recruits endothelial progenitor cells from the bone marrow and have regenerative and tissue protective effects in ischemic conditions, such as myocardial infarction [
47,
48]. Importantly, it counteracts and overrides angiostatic stimuli, such as those from the CXCL10/CXCR3 axis [
49]. CXCL12 also attracts mesenchymal stem cells to the inflammation sites, where they differentiate into osteoblasts and chondrocytes halting inflammatory bone destruction [
50]. CXCL12 may decrease in situ inflammation through these highly immunosuppressive mesenchymal stem cells or by direct induction of Th1 repolarization [
51,
52]. Thus, CXCL12 could account for the clinical improvement produced by MTX by mobilizing pro-angiogenic bone marrow cells, thereby limiting the local damage of ischemic episodes and their associated pain. Any consequent reduction of inflammatory bone destruction and perhaps even occurrence of bone regeneration could be particularly beneficial to patients with avascular osteonecrosis.
It is not known how MTX activates the above-mentioned chemokines in SCD, nor if it is a direct or indirect effect. Nevertheless, MTX has shown promise in reducing the often opioid-resistant osteonecrosis-associated pain in this present series. Our findings suggest CXCL12 as a putative marker that could mediate a possible MTX-induced limitation of ischemia–reperfusion damage in sickle cell disease.
The long-term implications of the MTX treatment in sickle cell disease are unknown. There is mounting evidence that deregulated angiogenesis may be an important component of a complex pathophysiology, affecting the course of complications as varied as leg ulcers, proliferative retinopathy, pulmonary hypertension and moyamoya syndrome [
53‐
58]. The CXCL12 up-regulation described in this report will likely have a context-dependent impact on the evolution of the disease. Mobilization of bone marrow cells might have positive implications if they limit inflammation [
51,
52] or if they promote neovascularization of ischemic areas, such as leg ulcers [
58]. However, mobilized fibrocytes might contribute to interstitial lung disease and pulmonary hypertension [
53,
55], and a pro-angiogenic environment might accelerate the development of certain complications, such as proliferative retinopathy [
57]. However, we need to consider that treatment with hydroxyurea has also anti-angiogenic activity, most likely mediated by HIF-1α down-regulation [
56,
57], which might not be helpful to wound healing and revascularization of infarcted areas. Yet, hydroxyurea is the only drug currently approved by the FDA for the treatment of SCD because of its overall benefit to the patients [
33]. Similarly, it remains to be determined if the putative therapeutic value of methotrexate in sickle cell disease overweighs any possible negative effect.
Finally, activation of the A
1 receptor was shown to reduce central nociceptive signaling in the spinal cord and it is possible that the MTX-induced systemic release of adenosine may decrease the neuropathic pain component in SCD patients [
59‐
61].