Reduces EPP-Related Liver Disease by Targeting Erythroid PPIX

Erythropoietic protoporphyria (EPP) (OMIM: 177000) is a cutaneous light sensitivity disorder that typically results from pathogenic variants in ferrochelatase (FECH), the last enzyme in heme biosynthesis. A decrease of more than approximately 65% of FECH activity causes accumulation of the photoactive pigment metal-free protoporphyrin IX (PPIX) in erythrocytes and secondarily in plasma and tissues (1). Approximately 95% of patients with EPP have a rare, pathogenic missense, nonsense, spliced, or deleted FECH allele in trans of a common intronic variant c.315-48T>C (rs2272783) that increases the use of an aberrant splice site, resulting in a premature stop codon (Figure 1A).

Generation of a K562 human erythroleukemic cell line with EPP. (A) Schematic of the human FECH locus showing the c.315-48T>C low expression splice variant (adapted from Mirmiran et al., ref. 20). (B) Aberrant FECH splicing in the K562-EPP cell line model of EPP. Lanes were run on the same gel but were noncontiguous, as indicated by the white line. (C) PPIX fluorescence in the control K562 (left) and K562-EPP (right) cell lines. FSC-H, forward scatter – height. (D) Quantification of PPIX levels in control K562 and K562-EPP cell lines. BLQ, below the limit of quantification. P values were calculated using an unpaired, 2-tailed t test with Welch’s correction. Data are presented as the mean ± SD.

X-linked protoporphyria (XLPP) (OMIM #300752) is largely clinically indistinguishable from EPP (2, 3). Patients with XLPP have C-terminal truncating pathogenic variants in the last exon of 5-aminolevulinic acid synthase 2 (ALAS2), the first and rate-limiting step in erythroid heme synthesis. ALAS2 condenses glycine with succinyl coenzyme A to produce 5-ALA (ALA), which is subsequently multimerized and modified to form PPIX. XLPP ALAS2 pathogenic variants lead to the loss of a negative regulatory domain, resulting in ALAS2 gain-of-function alleles that have increased catalytic activity, leading to ALA overproduction and PPIX accumulation. XLPP and EPP are collectively called protoporphyria, and less than 10% of patients with protoporphyria have XLPP (4, 5).

The prevalence of EPP and XLPP in Europe is reported to be approximately 1:100,000 individuals (6). An analysis of FECH variants in the UK Biobank has provided evidence that EPP prevalence is 1:17,000 (5, 7). Patients with protoporphyria experience life-long, painful phototoxicity from visible light exposure (1). Phototoxicity manifests as severe pain episodes that last for days and are unresponsive to analgesics, including narcotics. Even casual exposure to sunlight can precipitate pain. Moreover, due to light sensitivity, patients often have limited career and social opportunities that significantly affect their quality of life (8, 9).

In addition to photosensitivity, patients may also develop anemia, gallstones, and liver disease. Whereas 20%–30% of patients have chronic hepatic transaminitis, 1%–5% of patients will develop acute cholestatic liver failure that can be fatal or require liver transplantation (10, 11). A liver transplant may be followed by hematopoietic stem cell transplantation (HSCT) to prevent the recurrence of PPIX-related liver disease in the donor liver (12). Even though HSCT is curative, because of the up-front risk of death and long-term complications, such as graft versus host disease, HSCT is not used as first-line therapy.

Light sensitivity in protoporphyria extends into the visible spectrum, and sunscreens do not protect against phototoxic reactions. The only approved drug for phototoxicity prevention in adults is afamelanotide, a subcutaneous implant of a synthetic analog of human α-melanocyte–stimulating hormone, which increases melanin levels in the skin to provide protection against ultraviolet radiation in sunlight (13). Afamelanotide leads to an increase in pain-free light exposure and improvement in quality of life. It has also been hypothesized to improve liver function through an undefined mechanism (13, 14). Other drugs or approaches have been investigated (15, 16). Some, such as isoniazid, were unsuccessful (17, 18). ALAS2 inhibitors (19), an antisense oligonucleotide strategy targeting the c.315-48C allele (20, 21), and several other therapies are currently in development or clinical trials (22–25).

Heme production in the erythroblast is critically dependent on intramitochondrial glycine, which is a substrate for ALAS2. The very high demand for heme in the developing RBC requires extracellular glycine for adequate hemoglobinization. Deletion of the cell-surface glycine transporter (GLYT1), which is expressed primarily in the brain and erythroid precursors, results in decreased fetal liver erythroblast glycine uptake, reduced iron incorporation into heme, and microcytic anemia in newborn mice (26).

Bitopertin is an investigational oral, potent, and selective noncompetitive inhibitor of GLYT1 and has favorable pharmacokinetic properties, including an approximately 40-hour half-life in humans (27–29). The drug was initially developed for treatment of the negative symptoms (e.g., apathy and social withdrawal) of schizophrenia. Although early studies of bitopertin suggested efficacy in schizophrenia, this was not confirmed in later phase III trials (30–33). Nonetheless, bitopertin has exhibited an acceptable safety profile in clinical trials that have altogether enrolled over 4,000 participants.

In healthy volunteers, bitopertin results in a dose-dependent, but clinically negligible, reduction in hemoglobin (Hb) and results in erythrocytes that are hypochromic and microcytic (30, 33–35). Preclinical work in rats demonstrated that chronic dosing resulted in a microcytic, hypochromic, regenerative anemia with siderocytes (36). Treatment of patients with another structurally unrelated GLYT1 inhibitor also induces mild anemia, suggesting that the effect on erythropoiesis is GLYT1 dependent (37). The drug has been shown to decrease erythroid PPIX and heme levels in human healthy and β-thalassemic erythroid precursors (38). We hypothesized that bitopertin might effectively reduce erythroid metal-free PPIX accumulation in protoporphyric cells, potentially resulting in downstream clinical benefits such as decreased photosensitivity and liver disease. During the conduct of the studies reported here, Halloy and collaborators tested several glycine uptake inhibitors for their effect on heme synthesis and demonstrated that high concentrations of bitopertin (1 μM) reduced PPIX accumulation (30%–40%) in erythroblasts derived from CD34+ hematopoietic stem cells (HSCs) from a single patient with EPP (39). Bitopertin has now completed phase II clinical trials to evaluate its efficacy in erythropoietic protoporphyria (NCT05308472 and ACTRN12622000799752).

Here, we examined the effects of bitopertin in EPP and XLPP. We established that bitopertin has an EC50 of 32 nM in a cellular model of EPP generated by knocking down FECH in erythroblasts. We also confirmed that bitopertin reduced PPIX accumulation in erythroblasts derived from adult HSCs obtained from 3 unrelated patients with EPP, at a final concentration of as low as 10 nM. In mouse models of EPP and XLPP, oral administration of bitopertin reduced erythrocyte PPIX accumulation in a dose-dependent manner and rapidly acted to limit PPIX accumulation in the reticulocytes, which contained disproportionate amounts of PPIX compared with mature RBCs. Treatment with 200 parts per million (ppm) bitopertin for 8 and 32 weeks was associated with improvement of some liver disease parameters, significantly reducing liver fibrosis and ductular reaction in the EPP model. Altogether, these results support the ongoing development of bitopertin as an agent for the treatment of EPP and XLPP.

Bitopertin reduces PPIX accumulation in EPP cell models and in EPP patient–derived erythroblasts

Effect of bitopertin on K562-EPP cells. To confirm that bitopertin can reduce PPIX accumulation in EPP, we created a K562 erythroid cell line with an EPP genotype (K562-EPP) by introducing the common hypomorphic variant c.315-48T>C in trans of a null allele (p.Thr81fs8*). As expected, the c.315-48T>C allele resulted in an aberrant, longer FECH mRNA (Figure 1B). The PPIX concentration in undifferentiated K562-EPP cells was more than 10,000-fold higher than in the parental cell line (405.7 ± 20.6 pmol/106 cells vs < 0.2 pmol/106 cells; Figure 1, C and D). In K562-EPP cells, bitopertin inhibited the production of ALA and PPIX in a dose-dependent manner, with EC50 values of 3.2 nM and 3.4 nM, respectively (Figure 2, A and B), and had little effect on heme synthesis (Figure 2C).

In vitro effects of bitopertin in the K562-EPP cell line and FECH-depleted human erythroblasts. Data showing (A) 5-aminolevulinic acid (5-ALA), (B) PPIX, and (C) heme production in response to varying concentrations of bitopertin in K562-EPP cells. (D) PPIX accumulation in human umbilical vein HSCs treated with control or FECH shRNAs and differentiated into erythroblasts. P values were calculated using an unpaired, 2-tailed t test with Welch’s correction. (E) Dose response of PPIX production to varying concentrations of bitopertin in human umbilical vein HSCs treated with FECH shRNAs differentiated into erythroblasts. Data are presented as the mean ± SD.

FECH knockdown in human erythroblasts. We expanded human umbilical cord HSCs and differentiated them into erythroblasts for 16 days after infection with a lentiviral vector encoding either a control shRNA or FECH shRNA. This resulted in an approximately 80% decrease in FECH mRNA and protein expression levels and in an approximately 10-fold increase in PPIX fluorescence (Supplemental Figure 1, B and C, and Figure 2D; supplemental material available online with this article; https://doi.org/10.1172/JCI181875DS1). In this model, the PPIX inhibitory EC50 of bitopertin was 32 nM (Figure 2E). Over a range of concentrations from 0.45 nM to 1 μM, we found that bitopertin had no effect on cell viability or on the fraction of CD235+CD71+ erythroblasts (Supplemental Figure 1D).

EPP patient–derived erythroblasts. We isolated CD34+ HSCs from the peripheral blood (PB) of 3 unrelated patients with EPP and differentiated each sample into erythroblasts in the presence of vehicle (DMSO) or bitopertin either in the range of the EC50 of 10 nM, as previously used in thalassemic erythroblasts (38), or 50 nM (target dose) bitopertin added before the initial accumulation of PPIX (Figure 3, A–C), or 10 nM bitopertin added after the initial accumulation of PPIX (Figure 3, B and C). As a limited number of cells were obtained, only the 10 nM concentration added at the beginning of differentiation was evaluated in all 3 patients. Compared with controls, each bitopertin treatment scheme resulted in a marked reduction in PPIX accumulation, while it had only a limited dose- and time-dependent effect on morphological differentiation.

In vitro effects of bitopertin in erythroblasts differentiated from EPP patients. (A–C) Time course of PPIX accumulation and morphologic analyses in independent differentiation experiments. Erythroblasts were differentiated using PB from 3 unrelated EPP donors (EPP patients 1, 2, and 3) and exposed to DMSO or 10 nM or 50 nM bitopertin in DMSO beginning before or after PPIX accumulation. Horizontal dotted lines indicate the MFI of the PB at the collection of each donor sample.

Bitopertin inhibits glycine uptake in mouse and rat reticulocytes

To quantitate the extent to which bitopertin can inhibit the uptake of glycine in vivo, we dosed male C57BL/6J mice with 0.3–30 mg/kg bitopertin by oral gavage for 3 days and female Wistar rats with 0.1–3 mg/kg bitopertin by oral gavage for 7 days (Figure 4A). PB contains reticulocytes and RBCs. In rodents, more so than in humans, reticulocytes retain their ability to synthesize heme as demonstrated by the expression of key enzymes in the pathway (Supplemental Figure 2) (40). The plasma concentration of bitopertin and the ability of PB cells to take up 3H-glycine were determined ex vivo. The EC50 of the effect of bitopertin on glycine uptake was 1.2 nM and 2.0 nM in mice (Figure 4B) and rats (Figure 4C), with a maximal inhibition of 55% and 65%, respectively. In the rat, the effect of bitopertin was not different in the animals treated for either 1 day or 7 days, indicating that there was no tachyphylaxis to the inhibitory effect. Overall, these analyses indicate that bitopertin could inhibit RBC glycine uptake at doses that were readily achieved with oral administration.

In vitro glycine uptake of rodent PB exposed to bitopertin in vivo. (A) Experimental schema. (B) Male mice were exposed to daily oral gavage of 0, 0.3, 1, 3, 10, or 30 mg/kg bitopertin once per day for 3 days. On day 3, the concentration of plasma bitopertin and the percentage inhibition of 3H-glycine uptake were determined 1, 4, 8, or 12 hours after dosing. Schematic was created with BioRender.com (Schmidt, P., 2025; https://BioRender.com/oefao6p). The relationship between the inhibition of glycine uptake and drug concentration is displayed as 12-hour average inhibition and 12-hour average free plasma concentration of bitopertin. (C) Female rats were treated by gavage with 0, 0.1, 0.3, 1, and 3 mg/kg bitopertin once per day for 7 days. On days 0 (black) and 7 (red), the concentration of plasma bitopertin and the percentage inhibition of 3H-glycine uptake were determined 0.5, 2, 4, 8, or 12 hours after dosing. The relationship between the inhibition of glycine uptake and drug and concentration is displayed as 12-hour average inhibition and 12-hour average free plasma concentration of bitopertin.

Bitopertin treatment in mouse models of XLPP and EPP