Aspartate transaminases are required for blood development

Narges Pourmandi, Greggory Myers, Arjun Jha, Kelsey Temprine, Amanda Sankar, Nupur K. Das, Ridwana Khan, Siva Kumar Natarajan, Cristina Castillo, Peter Sajjakulnukit, Noah S. Nelson, Matthew D. Perricone, Indrani Talukder, Aaron D. denDekker, Lin Lin, Dominik Awad, Wesley Huang, Lei Yu, Navdeep S. Chandel, Rami Khoriaty, Yatrik M. Shah, Costas A. Lyssiotis

Posted on: 26 November 2025

Preprint posted on 19 September 2025

Not aspartate production or redox control, but a secret third thing: Aspartate transaminases may be important for epigenetic regulation during erythropoiesis.

Selected by Hannah Pletcher

Categories: biochemistry, cell biology, developmental biology

A) Schematic of blood cell development from preprint. B) Schematic of functions of aspartate transaminases from preprint. — Figure Caption: A) Pourmandi et al. (2025) set out to define the metabolism of early erythropoiesis. The timeline of erythropoiesis is defined in Fig. 1A of the preprint. B) After examining the different potential functions of aspartate transaminases (GOT1/2), shown in the schematic from Fig. 2A of the preprint, they found that aspartate transaminases may be required to maintain α-ketoglutarate levels during blood development. Figures provided by Narges Pourmandi and reproduced with permission from the preprint authors.

What I liked/why this new work is important:

This new work further uncovers metabolic processes essential for early erythropoiesis, and as mentioned in the preprint, targeting metabolism is a promising therapeutic strategy for anemia. Further, this work proposes a novel mechanism for GOT1/2 in early erythropoiesis.

Personally, I really appreciate the rigor of this preprint—the number of in vivo and in vitro models used and the confirmation of results from mice with human systems really strengthens the impact of the paper. Moreover, I like that the authors showed the negative results that led them to their overall conclusion about the role of GOT1/2 in erythropoiesis. Their process of examining each of the potential functions of GOT1/2 in erythropoiesis resulted in a very satisfying and compelling story.

Background:

Red blood cells (RBCs) transport oxygen and carbon dioxide throughout the body. RBC depletion causes anemia, resulting in fatigue and weakness, poor health outcomes, and increased morbidity and mortality (Gardner et al., 2023). A common therapeutic strategy to combat anemia is RBC transfusion, but this can lead to further clinical consequences (Ozment & Turi, 2009). In 2020, the FDA approved a treatment to reduce the burden of transfusion for certain disorders with defects in RBC development (Cappellini et al., 2020; Fenaux et al., 2020; King & Khoriaty, 2025). However, there is still a lack of therapies addressing the underlying causes of anemia.

One recently explored therapeutic strategy is targeting metabolism. Erythropoiesis, or the production of RBCs, is metabolically intense, so this preprint sought to define the metabolism of early erythropoiesis (Pourmandi et al., 2025). The authors discovered a role for aspartate transaminases, GOT1 and GOT2, in early erythropoiesis. The team found that instead of producing aspartate or maintaining redox levels through the malate-aspartate shuttle (MAS), GOT1 and GOT2 may actually be important for regulating α-ketoglutarate (aKG) levels for epigenetic regulation during erythropoiesis.

Key findings:

GOT2 is necessary for erythropoiesis.

With the goal of identifying metabolic changes during early erythropoiesis, the author team conducted metabolomics on four different models of erythropoiesis, which all displayed an increase in aspartate levels. This led the team to hypothesize that aspartate levels are necessary for erythropoiesis. Aspartate is synthesized by GOT2 in the mitochondria, so to test their hypothesis that aspartate synthesis is necessary for erythropoiesis, the team knocked out (KO) Got2 in mice, both in the whole body and specifically in erythroid progenitors. Both Got2 KO models had lower RBC counts, small RBCs, and high spleen weights, indicating that Got2 KO results in anemia. Therefore, GOT2 is necessary for erythropoiesis. Staining of bone marrow cells with cell surface markers revealed that in the erythroid-specific Got2 KO model, there were increases in early erythroid precursors and decreases in terminally mature erythroid precursors, suggesting that erythroid differentiation is disrupted. They confirmed the results from their mouse models in human cell culture, which is important because mice and humans have well-studied differences in erythropoiesis (An et al., 2014).

Aspartate regulation is not responsible for maintaining normal erythropoiesis.

GOT1 is the cytosolic counterpart of GOT2, which canonically operates in the opposite direction, breaking down aspartate. Given the opposing roles of GOT1/2, the authors hypothesized that deleting GOT1 in erythroid progenitors would increase aspartate and have the opposite effect as deleting GOT2. Surprisingly, KO of Got1 resulted in similar anemic phenotypes as KO of Got2 in erythroid progenitors, contrary to their hypothesis. Since GOT1/2 can operate in the reverse of their canonical directionality, the authors tested for compensation amongst these isoforms by generating an inducible double KO (DKO) of Got1 and Got2 in erythroid progenitors. These DKO mice exhibited similar anemic phenotypes as described above. Because the aspartate levels differed across the three mouse models of Got KO, but they resulted in similar anemia, the team suggested that it is not regulation of aspartate by GOT1/2 that maintains normal erythropoiesis.

Cellular redox regulation is not responsible for maintaining normal erythropoiesis.

If GOT1/2 aren’t necessary for regulating aspartate levels for erythropoiesis, what is their function? GOT1/2 are part of the malate-aspartate shuttle (MAS), where they work with malate dehydrogenase 1/2 (MDH1/2) to transport electrons for the reducing equivalent NADH into the mitochondria because NADH cannot cross mitochondrial membranes on its own. The MAS is therefore important for regulating the cellular redox state. Disrupting the MAS by KO of Got1/Got2 could cause reductive stress and downstream metabolic dysfunction. To determine the role of GOT1/2 in regulating the redox state through the MAS, the preprint authors generated an inducible KO of another MAS protein, Mdh1, in erythroid progenitors. These KO mice did not exhibit anemia or the anemic phenotypes shown by the Got KO mice. Therefore, the anemia resulting from Got KO is not due to the role of GOT1/2 in regulating the cellular redox state through the MAS. They validated this result with an orthogonal system in which an exogenous NADH oxidase, bacterial LbNOX, alleviates the reductive stress imposed by MAS inhibition. Even though the expression of LbNOX alleviated the reductive stress in erythroid-specific Got2 KO mice, the anemia remained.

GOT1/2 may be important for epigenetic regulation in erythropoiesis.

Finally, since Got1/2 KO are not disrupting erythropoiesis because of their function in aspartate regulation and cellular redox control, the authors hypothesized that changes in the other product of the canonical GOT2 reaction, aKG, are responsible for disrupting erythropoiesis. aKG is an important epigenetic regulator, as it is a cofactor for demethylating TET enzymes. The authors observed decreased histone methylation in the GOT1/2 KO human cells. This data suggests repression of genes in GOT1/2 deficient erythroid precursors.

Overall, the preprint authors identified GOT1/2 as required enzymes for early erythropoiesis. They investigated the mechanism of GOT1/2 in this process, showing that rather than being important for the regulation of aspartate or cellular redox state, these enzymes may be necessary for epigenetic regulation through aKG levels. This study motivates future work characterizing these epigenetic changes and further exploration of metabolism as a therapeutic target for anemia.

Future directions & questions for the authors:

I really appreciated the characterization of different erythroid precursors via staining that allowed you to identify the block in erythropoiesis when you knock out Got1/2 in erythroid progenitors. Can you explain why the Pre-CFU-E and CFU-E are displayed as percentage of live cells for the Got2 cKO mouse model, but the remaining progenitor counts are displayed as total cells normalized to 2 hind limbs (Fig. 2d, 2h, 2l)?
I was intrigued by the pattern of Annexin V marker across differentiation in the sgGOT1 and sgGOT2 CD34+ HSPCs, in which it increases on days 7 and 14 but shows no change from the control on day 11 (Fig. 4d). Why do you think the apoptosis levels fluctuate across differentiation? Is this pattern typical across differentiation?
Do you plan to follow up on your proposed model for GOT1/2 regulation of aKG metabolism and downstream epigenetics during early erythropoiesis? If so, what questions are you interested in exploring?
Do you think KO of the glutaminases (encoded by Gls and Gls2), being upstream of GOT1/2 in the production of aKG, would have a similar effect on early erythropoiesis given their similar expression pattern across erythropoiesis (Lyu et al., 2024)?
I find it very interesting that your work supports GOT1/2 as both important for erythropoiesis, but previous work showed opposing roles for these enzymes in hematopoietic stem cell function (Qi et al., 2021). How do you think these two studies fit together?

References:

An, X., Schulz, V. P., Li, J., Wu, K., Liu, J., Xue, F., Hu, J., Mohandas, N., & Gallagher, P. G. (2014). Global transcriptome analyses of human and murine terminal erythroid differentiation. Blood, 123(22), 3466–3477. https://doi.org/10.1182/blood-2014-01-548305

Cappellini, M. D., Viprakasit, V., Taher, A. T., Georgiev, P., Kuo, K. H. M., Coates, T., Voskaridou, E., Liew, H.-K., Pazgal-Kobrowski, I., Forni, G. L., Perrotta, S., Khelif, A., Lal, A., Kattamis, A., Vlachaki, E., Origa, R., Aydinok, Y., Bejaoui, M., Ho, P. J., … Piga, A. (2020). A Phase 3 Trial of Luspatercept in Patients with Transfusion-Dependent β-Thalassemia. New England Journal of Medicine, 382(13), 1219–1231. https://doi.org/10.1056/NEJMoa1910182

Fenaux, P., Platzbecker, U., Mufti, G. J., Garcia-Manero, G., Buckstein, R., Santini, V., Díez-Campelo, M., Finelli, C., Cazzola, M., Ilhan, O., Sekeres, M. A., Falantes, J. F., Arrizabalaga, B., Salvi, F., Giai, V., Vyas, P., Bowen, D., Selleslag, D., DeZern, A. E., … List, A. F. (2020). Luspatercept in Patients with Lower-Risk Myelodysplastic Syndromes. New England Journal of Medicine, 382(2), 140–151. https://doi.org/10.1056/NEJMoa1908892

Gardner, W. M., Razo, C., McHugh, T. A., Hagins, H., Vilchis-Tella, V. M., Hennessy, C., Taylor, H. J., Perumal, N., Fuller, K., Cercy, K. M., Zoeckler, L. Z., Chen, C. S., Lim, S. S., Aali, A., Abate, K. H., Abd-Elsalam, S., Abdurehman, A. M., Abebe, G., Abidi, H., … Kassebaum, N. J. (2023). Prevalence, years lived with disability, and trends in anaemia burden by severity and cause, 1990–2021: Findings from the Global Burden of Disease Study 2021. The Lancet Haematology, 10(9), e713–e734. https://doi.org/10.1016/S2352-3026(23)00160-6

King, R. A., & Khoriaty, R. (2025). Hereditary disorders of ineffective erythropoiesis. Blood Cells, Molecules, and Diseases, 111, 102910. https://doi.org/10.1016/j.bcmd.2025.102910

Lyu, J., Gu, Z., Zhang, Y., Vu, H. S., Lechauve, C., Cai, F., Cao, H., Keith, J., Brancaleoni, V., Granata, F., Motta, I., Cappellini, M. D., Huang, L. J.-S., DeBerardinis, R. J., Weiss, M. J., Ni, M., & Xu, J. (2024). A glutamine metabolic switch supports erythropoiesis. Science, 386(6723), eadh9215. https://doi.org/10.1126/science.adh9215

Ozment, C. P., & Turi, J. L. (2009). Iron overload following red blood cell transfusion and its impact on disease severity. Biochimica et Biophysica Acta (BBA) – General Subjects, 1790(7), 694–701. https://doi.org/10.1016/j.bbagen.2008.09.010

Pourmandi, N., Myers, G., Jha, A., Temprine, K., Sankar, A., Das, N., Khan, R., Natarajan, S., Castillo, C., Sajjakulnukit, P., Nelson, N., Perricone, M., Talukder, I., denDekker, A., Lin, L., Awad, D., Huang, W., Yu, L., Chandel, N., Khoriaty, R., Shah, Y., & Lyssiotis, C. (2025). Aspartate transaminases are required for blood development. BioRxiv: The Preprint Server for Biology, 2025.09.16.675844. https://doi.org/10.1101/2025.09.16.675844

Qi, L., Martin-Sandoval, M. S., Merchant, S., Gu, W., Eckhardt, M., Mathews, T. P., Zhao, Z., Agathocleous, M., & Morrison, S. J. (2021). Aspartate availability limits hematopoietic stem cell function during hematopoietic regeneration. Cell Stem Cell, 28(11), 1982-1999.e8. https://doi.org/10.1016/j.stem.2021.07.011

Tags: anemia, aspartate, aspartate transaminase, blood development, epigenetic regulation, erythropoiesis, metabolism, metabolomics

doi: https://doi.org/10.1242/prelights.42278

Read preprint

(No Ratings Yet)

Author's response

Narges Pourmandi shared

Future directions & questions for the authors:

I really appreciated the characterization of different erythroid precursors via staining that allowed you to identify the block in erythropoiesis when you knock out Got1/2 in erythroid progenitors. Can you explain why the Pre-CFU-E and CFU-E are displayed as percentage of live cells for the Got2 cKO mouse model, but the remaining progenitor counts are displayed as total cells normalized to 2 hind limbs (Fig. 2d, 2h, 2l)?

This is simply due to experimental workflow. To obtain the total cells normalized to 2 hind limbs, you multiply the percentage of (Pre-)CFU-E cells in live cells by the total cells in the hind limbs, which is determined by a manual hemocytometer count. I prefer this count because it factors in that there are different total cells collected from each mouse’s hind limbs and consequently provides a more accurate number of cells. For example, to use simple numbers, 50% (Pre-)CFU-E cells in live cells in 2 million total cells is 1 million total cells, whereas 50% of 10 million total cells is 5 million.

The Got2 cKO mouse model’s bone marrow analysis was done multiple times by different graduate students, and the result was decreased percentage of live cells each time. However, at this point in the project, the total cell count in the hind limbs was not completed in these experiments. In following mouse models, this count was done for greater accuracy. The percentages showed similar trends as the total cell numbers, and I could have reported these for consistency among all models, but I chose to report this data as normalized total cell numbers given the availability of the total cell counts and greater accuracy.

I was intrigued by the pattern of Annexin V marker across differentiation in the sgGOT1 and sgGOT2 CD34+ HSPCs, in which it increases on days 7 and 14 but shows no change from the control on day 11 (Fig. 4d). Why do you think the apoptosis levels fluctuate across differentiation? Is this pattern typical across differentiation?

In the 3-phase differentiation system, the primary human CD34+ cells are undergoing early erythropoiesis during days 0 – 7, roughly. I believe that GOT1/2 are important particularly in early erythropoiesis and this is captured by the fact that the KO cells are undergoing increased apoptosis at day 7. I showed one human donor in the preprint, but this result was recaptured in other donors.

The CD34+ cells are undergoing terminal erythroid differentiation (TED) from days 7 – 18, roughly. While it needs to be explored, my hypothesis is that there may be generally less reliance on the GOTs during TED. Like you mentioned, there was increased apoptosis at day 14 in the preprint, but there was no change at both days 11 and 14 in a subsequent donor. I would thus need to analyze more donors to see what the greater pattern is. For now, I am unsure the reason behind the increased apoptosis at day 14. It may be that this donor has more reliance on the GOTs during TED. I briefly explored this by knocking out GOT1 and GOT2 in a human erythroid line called HUDEP-2, which is an immortalized line in the proerythroblast cell stage that can differentiate through TED.¹ The GOT1 KO cells could not proliferate well, but the GOT2 KO cells could proliferate. Thus, it is possible that GOT1 may have a variable role during TED depending on the human donor, but further studies of GOT1/2 during TED would need to validate this. This variability in human donors was also captured in our bulk RNA sequencing experiment, and we included it as a limitation in the paper.

Do you plan to follow up on your proposed model for GOT1/2 regulation of aKG metabolism and downstream epigenetics during early erythropoiesis? If so, what questions are you interested in exploring?

Absolutely. We are still optimizing accurate detection of aKG in limited number of cells at the early differentiation timepoints, but I anticipate aKG being down in the GOT KO models. One key experiment currently being optimized is to supplement the human CD34+ cells with cell-permeable aKG and assess whether (a) the decreased proliferation and (b) the changes in histone markers resolve with supplementation. In addition, we want to understand what epigenetic changes are happening: less aKG should result in more DNA methylation, and we are interested in looking at DNA methylation marks.

Do you think KO of the glutaminases (encoded by Gls and Gls2), being upstream of GOT1/2 in the production of aKG, would have a similar effect on early erythropoiesis given their similar expression pattern across erythropoiesis (Lyu et al., 2024)?

I think they would, but through different mechanisms. A step above the glutaminases is the glutamine transporter, ASCT2. Oburoglu et al. showed that HSCs require ASCT2 to commit to the erythroid lineage.² They showed that this is because the cells require glutamine for nucleotide biosynthesis. Within this work, they utilized DON, a competitive inhibitor of glutaminase and de novo purine synthesis, and showed that this inhibited erythropoiesis. Importantly, they showed that this inhibition could not be rescued by aKG supplementation, but it could be rescued by nucleoside supplementation. Thus, inhibiting GLS or knocking out the GOTs both lead to inefficient erythropoiesis, but likely through different mechanisms.

I find it very interesting that your work supports GOT1/2 as both important for erythropoiesis, but previous work showed opposing roles for these enzymes in hematopoietic stem cell function (Qi et al., 2021). How do you think these two studies fit together?

It surprised and interested me as well. However, differentiation of hematopoietic stem cells into erythroid cells is a process that requires many different cell stages. While I expected the GOTs to have similar roles in the beginning of this process within HSCs and in the middle of the process within erythroid progenitor cells, it makes sense that they would have different functions due to the different needs of these cells. In HSCs, the GOTs are important for aspartate balance, but in erythroid progenitors, we are exploring how they are more important for aKG balance and epigenetics as these cells are preparing to enter terminal erythroid differentiation.

References:

Kurita, R. et al. Establishment of immortalized human erythroid progenitor cell lines able to produce enucleated red blood cells. PLoS ONE8, e59890–e59890 (2013).
Oburoglu, L. et al. Glucose and Glutamine Metabolism Regulate Human Hematopoietic Stem Cell Lineage Specification. Cell Stem Cell 15, 169–184 (2014).

Have your say Cancel reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.

Sign up to customise the site to your preferences and to receive alerts

Also in the biochemistry category:

Vitamin B12 supports skeletal muscle oxidative phosphorylation capacity in male mice

Luisa F. Castillo, Katarina E. Heyden, Abigail R. Williamson, et al.

Selected by 08 December 2025

Yan Aveiro dos Reis, Marcus Oliveira

Discussion

HAK-actin, U-ExM-compatible probe to image the actin cytoskeleton

Olivier Mercey, Luc Reymond, Florent Lemaître, et al.

Selected by 17 November 2025

Kanishka Parashar

Beyond venomous fangs: Uloboridae spiders have lost their venom but not their toxicity

Xiaojing Peng, Ludwig Dersch, Josephine Dresler, et al.

Selected by 13 November 2025

Daniel Fernando Reyes Enríquez, Marcus Oliveira

Also in the cell biology category:

Dosage-sensitive RBFOX2 autoregulation promotes cardiomyocyte differentiation by maturing the transcriptome

Mengmeng Huang, Feria A. Ladha, Yunxia Wang, et al.

Selected by 06 December 2025

Theodora Stougiannou

Discussion

ROCK2 inhibition has a dual role in reducing ECM remodelling and cell growth, while impairing migration and invasion

Daniel A. Reed, Anna E. Howell, Nadia Kuepper, et al.

Selected by 27 November 2025

Sharvari Pitke