Interplay between iron metabolism, inflammation, and EPO-ERFE-hepcidin axis in RDEB-associated chronic anemia

Recessive dystrophic epidermolysis bullosa (RDEB) is a genetic disease caused by mutations in the COL7A1 gene, resulting in the loss of function of collagen VII anchoring fibrils.¹ This structural defect leads to severe skin and mucosal fragility, causing patients to suffer repetitive cutaneous wounds and mucosal damage. These lesions cause blood loss; persistent inflammation; infection risk; iron deficiency (ID); and, ultimately, anemia, which further exacerbates the overall disease burden.^1-6 

Inflammation induces hypoferremia as an innate defense mechanism to restrict iron availability to pathogens.⁷ Although this strategy limits microbial growth, it simultaneously impairs erythropoiesis by reducing iron availability for red blood cell production. Furthermore, inflammatory cytokines exert a direct negative effect on red blood cell production.⁸ Interleukin-1β (IL-1β) skews hematopoiesis toward myelopoiesis, prioritizing the production of immune cells over erythrocytes to meet the increased demand for host defense.^9,10 Tumor necrosis factor (TNF) contributes to this shift by directly inhibiting the differentiation and survival of erythroid progenitors,^11,12 whereas interferon-γ (IFN-γ) shortens erythrocyte life span, further compromising erythroid output.^13,14 The sustained activation of these mechanisms ultimately leads to anemia of inflammation (AI), also referred to as anemia of chronic disease. AI is associated with highly prevalent conditions including cancer, chronic kidney disease, heart disease, HIV infection, autoimmune diseases, and other inflammatory disorders.^8,15,16 Similarly, RDEB illustrates the complex interplay between inflammation, iron dysregulation, and increased erythropoietic demands, making it a valuable model for studying AI in the context of chronic systemic damage.

Hepcidin, erythropoietin (EPO), and erythroferrone (ERFE) are considered the primary systemic regulators that balance iron metabolism and erythropoiesis during inflammation and anemia.^17-19 The secretion of these hormones is controlled by cytokine signaling, hypoxia, and body iron status. Hepcidin, predominantly secreted by hepatocytes and, to a lesser extent, by myeloid cells in response to IL-6, IL-1, or Toll-like receptor signaling, is the central mediator of inflammatory iron restriction.¹⁷ By binding to the iron exporter ferroportin, hepcidin induces its internalization and lysosomal degradation, thereby reducing duodenal iron absorption and promoting iron sequestration within macrophages and hepatocytes.¹⁷ In addition, Toll-like receptor signaling independently suppresses ferroportin expression in response to infection or tissue damage, linking these inflammatory cues directly to hypoferremia, even in the absence of elevated hepcidin levels.²⁰ 

As described earlier, inflammatory signals and the restriction in iron availability impair erythropoiesis, contributing to the development of AI. In response to hypoxia, EPO secretion increases, which promotes tissue repair and the expansion of erythroblast while attenuating inflammatory signaling.¹⁸ Erythroblasts secrete ERFE,^21-23 which disrupts bone morphogenetic protein signaling, reducing hepcidin secretion by hepatocytes and thereby allowing iron release to support erythropoiesis. However, chronic inflammatory signals can unbalance this homeostatic loop by suppressing EPO production and its downstream signaling.^7,8 Thus, in chronic pathologies, this dysregulation causes inflammatory iron restriction to progress into absolute ID, making it common for AI to coexist with ID anemia (IDA).^8,15 Under conditions of ID, the production of EPO in response to hypoxia is further diminished as the body attempts to conserve limited iron resources for essential functions beyond erythropoiesis. This suppression impairs the maturation of red blood cells,²⁴ thereby complicating the restoration of hemoglobin levels when AI and IDA occur simultaneously.¹⁸ 

ID and inflammation are considered the main contributors to anemia in RDEB, yet the role of specific mediators and the reasons behind the limited efficacy of current therapies remain incompletely understood.^2-4,6 In RDEB, continuous wounding leads to unrelenting inflammation that complicates anemia recovery. However, because of the constant risk of infection, anti-inflammatory drugs are not the first line of treatment for RDEB. This situation underscores the need for alternative therapeutic strategies that specifically target iron metabolism and erythropoiesis. A deeper understanding of these under an inflammatory context could provide a foundation for the development of novel therapies for RDEB and other conditions where anemia is a prevalent comorbidity.

The modulation of the EPO-ERFE-hepcidin axis offers a promising option for anemia treatment for patients facing complex, multisystem conditions such as RDEB.¹⁵ However, our understanding of how these hormones function in the context of IDA and AI remains limited. In this study, we conduct a comprehensive analysis to quantitatively assess the interrelationships among inflammatory cytokines, the EPO-ERFE-hepcidin axis, and their combined impact on iron metabolism and erythropoiesis in patients with RDEB. Through this approach, we seek to provide a rational basis for the development of targeted therapeutic strategies.

This study was approved by the ethics committee at Hospital Universitario La Paz (HULP; code: PI-4690). Written informed consent was obtained from all participants and/or legal guardians.

The design of this observational, noninterventional, cross-sectional study has been described previously²⁵ and is summarized in the supplemental Methods of this article.

Blood samples were collected only when the patient’s health status allowed. Laboratory parameters were analyzed in the Central Clinical Laboratory at HULP using standard reference ranges (supplemental Table 1). EPO quantification was measured at hospital facilities, using the Atellica IM EPO assay, a 1-step sandwich immunoassay designed for in vitro diagnostic.

Nonstandardized clinical measurements were performed at the Hemostasis Laboratory of HULP. Cytokine levels were determined using commercial kits (Merck) as previously described,²⁵ including IFN-γ, IL-1β, IL-4, IL-6, IL-10, and TNF. Commercial validated competitive enzyme-linked immunosorbent assays were used to measure hepcidin (EIA-5782 DRG Instruments GmbH) and ERFE (SKU# ERF-001, Intrinsic Lifesciences). The control range was determined in an age- and sex-matched cohort of healthy controls (HCs). Controls and patient samples were processed in parallel to ensure consistency and minimize variability. Values above the 95th or 90th percentile of HCs were considered elevated.²⁵ The specific number of samples available for each parameter was different because of pediatric volume constraints and the health status of each patient (see specific n value in each data set). Extremely high cytokine levels from 1 patient were excluded from analysis because of suspected analytical error.

Statistical analysis and visualization were performed as previously described²⁵ and are summarized in the supplemental Section.

The procedures followed in this study were in accordance with the ethical standards of the Declaration of Helsinki of 1975, as revised in 1983, and approved by the responsible committee on human experimentation at Hospital Universitario La Paz (Code: PI-4690). Written informed consent was obtained from all participants and/or their legal guardians, which includes consent for the publication in scientific media.

Demographic (age and sex) and severity characteristics (epidermolysis bullosa disease activity and scarring index [EBDASI] scores, and percentage of body surface area [BSA] affected) of the cohort studied were previously described²⁵ and are summarized in supplemental Table 2.

Anemia was present in 50% (42/84) of the RDEB cohort, with mild anemia in 10 of 42, moderate in 30 of 42, and severe in 2 of 42 cases. No sex-based differences were observed (supplemental Table 3). The youngest patients with anemia were aged 4 years (n = 2). The prevalence of anemia escalated from 5 years of age (P < .0001), with no significant difference between children and adults thereafter (Figure 1A). After this study, 5 patients passed away, including 2 of 2 with severe anemia and 3 with hemoglobin levels between 9.4 and 10.3 g/dL, suggesting anemia as a potential mortality risk factor in RDEB, as reported in other chronic conditions.²⁶ Hematological parameters, including reticulocyte production index (RPI), are described in the supplemental Table 4 and supplemental Figure 1.

The EBDASI total score strongly negatively correlated with hemoglobin levels (Figure 1B), and anemia prevalence was notably higher among patients with severe disease (78%) but absent in those with mild disease (supplemental Table 3). Hemoglobin levels showed a strong negative correlation with the percentage of BSA affected and C-reactive protein (CRP) levels, which are used as damage and inflammatory markers, respectively (Figure 1C-D). Anemia prevalence increased significantly in patients with ≥20% of BSA affected (from 4% to 67%; P < .0001) or with a CRP of ≥15 mg/L (from 0% to 79%; P < .0001). To evaluate their combined predictive value for monitoring anemia risk in RDEB, patients were assigned 1 point for each of these 3 criteria: age of ≥5 years, percent BSA affected of ≥20%, and CRP ≥ 15mg/L. The maximum score of 3 points was associated with a specificity of 78.6%, a sensitivity of 92.9%, and a positive predictive value of 83% to identify patients with anemia (Figure 1E).

We have previously reported a rise in inflammatory cytokines in this RDEB cohort, predominantly in patients with moderate or severe disease.²⁵ Herein, only IL-6 and IL-10 levels were higher in patients with anemia than in those without anemia. However, after adjusting for disease severity as a potential confounding factor and comparing patients with anemia with patients without anemia within similar EBDASI categories (moderate and severe), these differences were no longer significant (Figure 2). Nevertheless, patients with mild EBDASI, all free from anemia, exhibited the lowest levels of all analyzed cytokines except for TNF²⁵ (Figure 2). Notably, a patient on tocilizumab treatment, with TNF (32 pg/mL) as the only elevated cytokine (besides blocked IL-6), remained without anemia despite high CRP levels, a severe EBDASI score, 70% of BSA affected, and renal disease.

Hemoglobin and serum iron levels exhibited a strong correlation (Figure 3A). All patients with anemia (41/41) and 17 of 42 (40%) individuals who did not have anemia displayed hypoferremia, including all with severe disease and 1 with mild disease. IL-6 levels inversely correlated with iron levels, albeit weakly. In fact, 9 patients maintained normoferremia despite elevated IL-6 levels (Figure 3B). CRP and circulating iron demonstrated a significant negative semilogarithmic correlation, and most patients presented iron levels of <40 μg/dL when CRP was ≥15mg/L (Figure 3C). Transferrin and total iron-binding capacity were reduced in 26 of 41 (63%) and 24 of 41 (59%) patients with anemia, respectively. The transferrin saturation index (TSat) was low in 38 of 41 (93%) patients with anemia but also in 17 of 41 (41%) patients without anemia, indicating iron restriction (supplemental Table 4).

Ferritin levels proved inconsistent with other hematologic parameters when interpreted using standard reference values (supplemental Table 4). Therefore, various approaches have been proposed to assess iron status amid inflammation using ferritin and TSat.²⁷ The World Health Organization’s guideline on iron status assessment incorporates 3 criteria recorded in our study: age (>5 years), CRP (>5 mg/L), and ferritin cutoff points that vary accordingly²⁸ (Figure 3D). Following this guideline (Figure 3E,F), 19 of 42 (45%) patients without anemia exhibited low ferritin levels, which is indicative of ID without anemia (IDWA). Among patients with anemia, all with elevated CRP, 27 of 41 (66%) had ferritin levels below the established reference range and TSat of ≤15%, thus meeting criteria for combined AI and IDA diagnoses (Figure 3F). Conversely, 10 of 41 (24%) patients with anemia had ferritin within range (preserved iron stores). All these patients (10/10) had CRP levels of >50 mg/L (Figure 3E) and accordingly were diagnosed with isolated AI. Notably, they had a TSat of ≤15%, indicative of functional ID (FID), and 7 of 8 (88%) exhibited elevated hepcidin levels (see in the Results section). Finally, 4 of 41 (10%) patients with anemia had ferritin of ≥500 μg/L, of whom 3 of 4 (75%) had TSat of ≤15% (indicating FID) and 4 of 4 (100%) exhibited elevated hepcidin levels. Based on these criteria, these patients with anemia were classified as at risk of iron overload (RIOL) and were also diagnosed with isolated AI.

The analysis of iron status and anemia across age groups (Figure 3F) demonstrated trends in prevalence, with IDWA more frequent among young children (age of <5 years; 31%, 4/13), underscoring the importance of early detection and intervention to prevent progression to AI + IDA. Nevertheless, the highest prevalence of ID (65%, 24/37) was among children and adolescents (age 5 to <18 years), and the RIOL was only found within adults, who had the lowest prevalence of ID (27%, 9/33).

Prevalence of oral iron supplementation and history previous IV iron (IVFe) infusions and blood transfusions are described in the supplemental Results (supplemental Figures 2 and 3). A preliminary analysis of 7 patients who received IVFe infusions within the 6 months before this study, focusing on iron metabolism parameters, inflammatory markers, and the EPO-ERFE-hepcidin axis, is also included in supplemental Table 5.

Analysis of bioactive hepcidin in 77 patients and 72 HCs revealed significantly higher levels in patients with anemia than HCs, but not among patients without anemia (Figure 4A). Defining a hepcidin cutoff is challenging because of circadian and postprandial variability. Although the 95th percentile is commonly used to define elevated levels, we opted for the 90^th percentile (22 ng/mL) because patients were fasting overnight. We found that 21 of 77 (27%) patients had elevated hepcidin: 15 of 38 (39%) had anemia (14 with severe EBDASI score), and 6 of 39 (15%) did not have anemia (3 severe, 2 moderate, and 1 mild RDEB cases; Figure 4A,B). All patients with increased hepcidin had CRP of ≥25 mg/L and/or elevated IL-6 levels (Figure 4C). However, they represented only a third of those with these raised inflammatory markers.

Hepcidin levels varied across age groups, being higher in adults, whereas hepcidin-to-ferritin ratios were similar across groups, suggesting an indirect effect of age-related increases in ferritin (supplemental Figure 4). Hepcidin did not correlate with EPO levels, serum iron, or TSat (Figure 4C). Weak correlations (ρ < 0.4) were observed with hemoglobin, IL-6, and the percentage of BSA affected, whereas moderate positive correlations with CRP and ferritin levels highlight the more relevant role of inflammation and iron status in regulating hepcidin levels (Figure 4C).

Analysis of hepcidin in patients categorized by their iron status revealed that the proportion of patients with elevated hepcidin was associated with iron stores (Figure 4D). Hepcidin was high in 6 of 44 (14%) patients with ID, 10 of 28 (36%) patients with preserved iron stores, and 5 of 5 (100%) patients at RIOL (ferritin of ≥500μg/L, including the patient on tocilizumab treatment). Among patients with elevated hepcidin, 17 of 21 (81%) had low TSat (≤15%) whereas most patients with ID (91%, 40/44) also had TSat of ≤15%, despite only 6 of 44 (14%) having high hepcidin (Figure 4E).

EPO response in patients with RDEB, as measured by logarithmic EPO, inversely correlated with hemoglobin levels, and linear regression analysis revealed a slope of −0.17 (Figure 5A). Standard diagnostic laboratory ranges indicated that EPO levels were unexpectedly within the normal range (NR) in 16 of 40 (40%) patients with anemia, including patients with moderate anemia, suggesting an insufficient response to anemia (Figure 5A-B). Conversely, patients with the lowest hemoglobin levels exhibited levels far above the hemoglobin −logEPO curve (Figure 5A). In contrast, in 10% of individuals who did not have anemia, EPO was below the NR (Figure 5B), although none of them had renal disease. In those with anemia, EPO levels above the NR were associated with lower hemoglobin; greater RDEB severity (reflected in EBDASI activity and total scores); and higher CRP, transferrin, and ferritin levels, which are likely indicative of inflammation (Figure 5C). However, no significant differences were observed in age, RPI (Figure 5C), or the levels of IL-6, IL-1β, IL-10, or TNF (not shown) between patients with anemia with EPO levels above the NR and those within it.

ERFE levels in the HC cohort (median, 1.2 ng/mL) were lower than those described in previous studies in adults²⁹ (median, 8 ng/mL), infants³⁰ (median, 7.4 ng/mL), and children³¹ (median, 38 ng/mL). These differences may reflect methodological variations across studies or population-specific characteristics.

ERFE was elevated in patients with RDEB compared with HCs, regardless of their anemia status. However, patients with anemia exhibited the highest ERFE levels (Figure 6A-B), demonstrating a moderate negative correlation with hemoglobin levels and a positive correlation with the percentage of BSA affected, CRP, and IL-6 levels (Figure 6C). ERFE and EPO levels were also positively associated (Figure 6C) and patients with elevated ERFE also showed higher EPO than those within HC cutoff (supplemental Figure 5A). Nevertheless, the rise of ERFE was not strictly linked to EPO response, with 14 of 38 (37%) patients without anemia having had elevated ERFE despite having EPO within the NR (Figure 6B). Additionally, although 29 of 36 (83%) patients with anemia had elevated ERFE (Figure 6B), no significant differences in ERFE levels were observed between patients with anemia with EPO levels above the NR and those within the NR (supplemental Figure 5B). Circulating iron and TSat showed an inverse correlation with ERFE (Figure 6C), indicating that patients with ID had higher ERFE. However, patients with anemia and patients without anemia with true or functional ID (TSat ≤ 15%) did not always exhibit increased ERFE levels (Figure 6D-E). Despite the established role of ERFE as an inhibitor of bone morphogenetic protein–dependent hepcidin secretion,¹⁹ 7 patients with anemia with elevated ERFE also had high hepcidin levels and ID (Figure 6C).

Given the complexity of determining and restoring iron stores in the context of a chronic inflammatory disease and anemia, further investigation was conducted to explore the differences among patients classified as: no anemia with preserved iron stores (N), IDWA, AI, or AI combined with IDA (Figure 7).

No significant age differences between groups were observed, although the youngest patient with IDWA was aged 1 year, whereas the youngest in the AI and AI + IDA groups were aged 9 and 4 years, respectively (Figure 7A). The severity of anemia, as assessed by hemoglobin levels, was similar across the anemic categories (Figure 7B). However, mean corpuscular hemoglobin was lower in both AI and AI + IDA groups than for those without anemia, whereas a reduction in mean corpuscular volume was observed only in those in the AI + IDA group, indicating a higher prevalence of microcytosis (Figure 7C-D).

Although patients with anemia had higher EBDASI scores than those without anemia, the combination of IDA and AI was not associated with greater disease severity compared with isolated AI. The medians of the percentage of BSA affected and CRP levels were higher in patients with AI than in those with AI + IDA but these differences lost significance after Dunn correction (Figure 7E-G). No differences in IL-6 levels were observed between the groups (Figure 7H), nor in the other cytokines analyzed (data not shown).

Differences in ferritin were according to iron status classification criteria whereas serum iron levels were similar in AI and AI + IDA (Figure 7I-J). Iron restriction (TSat ≤ 15%) was detected in all patients with anemia except for 1 in the RIOL category. Low TSat was also observed in a significant proportion of patients without anemia, either associated with absolute ID (74% of patients with IDWA) or indicating FID (26% of patients in the N group; Figure 7K). In addition, although transferrin levels are generally reduced by inflammation, they were significantly lower in patients with AI than those with AI + IDA, suggesting a possible compensatory effect because of the coexistence of ID¹⁵ (Figure 7L).

Hepcidin levels (Figure 7M) were highest in patients with AI, with all but one (marked in orange) having values above the reference cutoff (92%, 11/12). Among them, 10 of 11 (91%) had low TSat, indicating that hepcidin was likely causing FID in these patients.³² In fact, the only patient with normoferremia had received IVFe 27 days prior. Nonsignificant differences were observed between other groups. Interestingly, only 1 patient with IDWA had elevated hepcidin levels whereas among those without anemia and preserved iron stores, 5 of 23 (22%) had elevated hepcidin levels. Active cutaneous infection or colonization of wounds with pathogenic bacteria might results in increased ferritin and hepcidin independently of iron status. However, although active cutaneous infection was more prevalent in those with anemia, no significant differences were detected between AI and AI + IDA groups (supplemental Figure 2B).

EPO subresponders to anemia (EPO at NR) were observed at similar proportions among patients with AI and those with AI + IDA, but EPO below NR was restricted to the N group (Figure 7N). Furthermore, patient groups with anemia exhibited the highest levels and a similar degree of increase in ERFE levels (AI = 82%, 9/11; AI + IDA = 83%, 20/24; Figure 7O).

This study represents, to our knowledge, the first analysis of the etiopathogenic and regulatory factors of anemia within a representative cohort of patients with RDEB. The data reported here shed light on the heterogeneity of this condition and accentuate the need for personalized evaluation of systemic modulators to determine the most effective therapeutic approach.

Our findings confirm that blood loss, ID, and inflammation play key roles in the development of anemia in patients with RDEB.^2,4,6 Hematologic parameters observed in patients with anemia along with the strong negative correlation between hemoglobin, the percentage of affected skin, and the levels of CRP underscore this relationship. Affected BSA of ≥20% and levels of CRP of ≥15mg/L emerge as potential risk factors for anemia in this genodermatosis.

Interpreting ferritin levels in the context of inflammation requires caution. In patients with severe inflammatory states, wounds colonized by pathogenic microbiota, or active cutaneous infections, ferritin levels can be significantly elevated, leading to an overestimation of iron stores. Based on World Health Organization criteria, our results showed a slightly higher prevalence of IDA (66%) than that previously described in a pediatric RDEB cohort (52%) that used the soluble transferrin receptor to log10-transformed ferritin ratio.² This difference may be because of clinical and demographic variations between the cohorts.

FID without anemia and IDWA were prevalent among the RDEB cohort. However, the low percentage of these patients taking oral iron supplementation suggests that these conditions may be underdiagnosed. Limited iron availability, even without anemia, can lead to a wide array of symptons.³³ Furthermore, the erythropoietic response to EPO is reduced, and terminal erythropoiesis is hampered,^24,34 increasing the risk of anemia.

As in other conditions, it can be hypothesized that hepcidin is the primary driver of sustained FID/ID in patients with RDEB. However, our data reveal that hepcidin was elevated in only 25% of the patients, mainly those with higher levels of ferritin (linked to FID) or those who had received second-line treatments for iron restoration. These patients represent most of those with isolated AI, suggesting they may have derived from those with combined AI + IDA who received these treatments, restoring their iron stores. Nevertheless, as the underlying inflammation remained unresolved, AI persisted, and the activation of hepcidin secretion induced FID, leading to an inadequate bone marrow response. Supporting this notion, the minimum age among patients with isolated AI was higher than those with combined AI + IDA. A recent retrospective study reported that parenteral iron can increase hemoglobin levels in adult patients with RDEB, although responses were heterogeneous and limited, possibly because of the detrimental effect of ongoing inflammation.³⁵ 

The highest levels of ferritin, classified as RIOL, were associated with the highest levels of hepcidin. Despite ferritin’s dual role during inflammation, hyperferritinemia is associated with high mortality³⁶ and has been proposed as a prognostic marker for early mortality in RDEB.² Consistent with these previous findings,² we observed that all patients with the highest ferritin levels had received IVFe and/or blood transfusions. This highlights the need to reevaluate the potential iatrogenic effects of these treatments in RDEB, especially when administered without concomitant therapies that reduce their frequency. Such treatments may inadvertently worsen the patient’s condition.^37,38 

Oral iron supplementation is the first-line treatment for ID in RDEB but is frequently insufficient.⁶ Factors such as concomitant celiac disease, proton pump inhibitors, histamine-2 receptor antagonists prescription, TNF signaling, and/or the IL-4 receptor macrophage–dependent reduction of transferrin in the duodenum could be hindering iron absorption independently of hepcidin.^8,39-41 Additionally, the side effects of long-term use of oral iron, beyond constipation, should be considered. Unabsorbed iron alters gut microbiota, and cosupplementation of ferrous salts with vitamin C exacerbates oxidative stress in the gastrointestinal tract, compromising its epithelial lining.⁴² 

The coexistence of AI and IDA presents a significant therapeutic challenge.^8,15 Without adequate control of inflammation, attempts to restore iron stores may inadvertently reactivate hepcidin secretion (leading to FID), whereas the impaired response to EPO and the negative impact of inflammatory cytokines on erythropoiesis persist. Identifying potential targets to develop new therapeutic strategies is essential for advancing the management of anemia in RDEB. One area of focus is targeting cytokines involved in the inflammatory process. IL-6, significantly elevated in this genodermatosis,^25,43,44 is the main inducer of hepcidin-dependent ID¹⁷ and shows the highest correlation level with other cytokines.²⁵ Our cohort included an adult patient with severe RDEB, with inflammatory renal disease who was treated with tocilizumab. After blocking IL-6 signaling, anemia was corrected despite 70% of BSA being affected. Notably, hepcidin levels were raised, associated with RIOL and FID. Because of the risks of infection, however, IL-6 receptor blockade or high doses of broad-spectrum anti-inflammatory drugs are not the first choice of treatment in RDEB. As an alternative, we, and others, have proposed the IL-4/IL-13 signaling pathway as a therapeutical target for RDEB.^25,44 A small group of patients with RDEB showed some improvement with short-term dupilumab treatment, although no data on anemia status are available.^45,46 Similar findings were observed with JAK-STAT inhibitors, baricitinib, or upadacitinib.⁴⁷ Interestingly, although these drugs may increase anemia because of their effects on EPO signaling, momelotinib, which inhibits both JAK1/2 and ACVR1 (a key mediator in IL-6–induced hepcidin production), has shown potential for ameliorating AI.^48,49 Nevertheless, increasing iron bioavailability could raise the risk of infection and sepsis in RDEB.⁷ Therefore, careful monitoring will be required.

This study is, to our knowledge, the first to analyze EPO levels in a cohort of patients with RDEB. Although exploratory studies have proposed the coadministration of erythropoiesis stimulating agents (ESA) with IVFe infusion in RDEB,^6,50,51 the efficacy and possible side effects of ESA has not yet been systematically described in this population. EPO is known to have protective effects beyond erythropoiesis,^18,52 but our data showed 2 distinct scenarios: moderate and severe cases of anemia with endogenous EPO was elevated beyond the expected range for their hemoglobin levels, and others in whom EPO remains within NR despite anemia. In the former group, ID and inflammatory signals (including high mobility group box 1 protein), could be inhibiting EPO signaling, whereas in the latter group, EPO response to anemia might be impaired.^8,53 These observations suggest that ESA requirements or responses could differ among patients with RDEB. Recently, transferrin receptor inactivation has been proposed as an alternative treatment for anemia in chronic kidney disease, because it enhances EPO responsiveness of erythroid cells while adjusting iron availability by reducing hepcidin levels.⁵⁴ Such strategies could hold promise for patients with RDEB, in which balancing EPO responsiveness and iron availability remains a key therapeutic challenge.

The rise of ERFE in patients with RDEB with anemia and those without anemia raises important questions. Elevated ERFE levels were associated with low TSat, suggesting that ERFE signaling was insufficient to restore iron availability or normalize hemoglobin levels. ERFE elevation may indicate ineffective erythropoiesis¹⁹ in RDEB, driven by ID and inflammation, as reflected by the low RPI observed in patients with anemia. Interestingly, ERFE levels did not consistently correlate with EPO elevation, leaving the mechanism behind sustained high ERFE levels, and its apparent ineffectiveness in mobilizing iron or enhancing the response to OIS in RDEB, unresolved. ERFE, also known as myonectin, has positive and negative effects beyond iron metabolism regulation.¹⁹ It is also a myokine that can ameliorate skeletal muscle dysfunction and modulate lipid metabolism, although the mechanism underlying such effects is unknown.^55,56 These additional roles of ERFE could have implications for RDEB pathology and warrant further investigation.

RDEB exemplifies a chronic, complex inflammatory disease in which multisystemic comorbidities stemming from congenital skin damage contribute to AI and IDA. Despite the limitations inherent to cross-sectional observational research, our study underscores the intricate interplay between iron metabolism and inflammation in RDEB. These insights may apply to other chronic inflammatory diseases in which anemia is a common comorbidity, guiding more effective, individualized therapies. Future longitudinal studies in diverse cohorts are needed to refine treatment strategies that address both anemia and cutaneous symptoms, ultimately improving overall patient outcomes.

The authors express profound gratitude to the patients and their families for their participation in this study. The authors further extend their deepest appreciation to the patient advocacy groups Dystrophic Epidermolysis Bullosa Research Association (DEBRA)-España and Berritxuak for their support and collaboration.

This study was funded by the Hospital Universitario La Paz Dermatology Service; by Foundation for Biomedical Research of La Paz University Hospital (FIBHULP) (EC_5215; I4V-MC-JAIP); and projects granted by the Ministry of Science and Innovation of Spain PID2020-119792RB-I00, PID2020-1230680B-100, and PID2021-123068OB-I00/AEI/10.13039/501100011033/FEDER; Unión Eurpea (UE); and by the Carlos III Health Institute (ISCIII), Spain, cofinanced by the European Union (NextGenerationEU): La Red Española de Terapias Avanzadas (RICORS-TERAV) RD21/0017/0033 and RD21/0017/0010. M.J.E. was the recipient of a contract funded by DEBRA Austria, cofunded by DEBRA Sweden and EB-LOPPET, and supported by EB Research Network (León-1).

Contribution: M.J.E. and R.S. conceptualized and supervised the study and were responsible for project administration; L.Q.-C., R.M., M.J.E., and R.S. were responsible for the study methodology; L.Q.-C., R.M., N.B., M.G.C., and A.B. were responsible for validation; L.Q.-C., P.Z., M.d.R., A.V., M.J.E., and R.S. were responsible for formal analysis; L.Q.-C., R.M., I.P.-C., N.B., P.A., E.M.-M., M.G.C., A.B., E.J., J.V., M.C.d.A., and R.d.L. performed the investigation; L.Q.-C., R.M., I.P.-C., N.B., A.B., and R.d.L. provided resources; L.Q.-C., M.G.C., A.B., N.B., P.Z., M.J.E., and R.S. were responsible for data curation; L.Q.-C., M.J.E., and R.S. were responsible for data visualization and preparing the original manuscript draft; R.d.L., M.d.R., A.V., and M.J.E. acquired funding; and all authors reviewed and edited the manuscript.

Conflict-of-interest disclosure: The authors declare no competing financial interests.

Correspondence: Rosa Sacedón, Department of Cell Biology, Faculty of Medicine, Complutense University of Madrid, Pl de Ramón y Cajal, Madrid 28040, Spain; email: [email protected]; and María J. Escámez, Departamento de Bioingeniería, Universidad Carlos III de Madrid, Avda de la Universidad 30, Madrid 28911, Spain; email: [email protected].