A critical role for heme synthesis and succinate in the regulation of pluripotent states transitions

Decision letter after peer review:

Thank you for submitting your article "A critical role for heme synthesis and succinate in the regulation of pluripotent states transitions" for consideration by eLife. Your article has been reviewed by 2 peer reviewers, and the evaluation has been overseen by a Reviewing Editor and Mone Zaidi as the Senior Editor. The reviewers have opted to remain anonymous.

The reviewers have discussed their reviews with one another, and the Reviewing Editor has drafted this to help you prepare a revised submission.

Essential revisions:

1) Are the levels of heme itself, or the mentioned 7 enzymes of the heme pathway regulated during differentiation?

2) Further experiments are needed to determine whether the lack of proper Tgf β and FGF-ERK signalling activation are cause or consequence of the observed differentiation defects. How do cells sense heme deficiency and does how heme deficiency result in inability to activate ERK/MAPK and TGF β signaling?

3) Can the authors attempt to titrate pathway inhibition to a similar level as observed in heme pathway deficient ESCs? Furthermore, can the differentiation defect be rescued upon overstimulation of FGF-ERK and TGF β to reach WT levels?

4) Analyse cell proliferation and viability after SA (or AA5) treatment. If there is an impact on cellular health, this needs to be reported and taken into careful consideration when interpreting results.

5) The statement made in lines 150-152 and shown as Suppl. Figure 2b should be further supported by proper quantification using replicate assays.

6) Expression levels in TS cells or in trophoblast tissue must be used as control to judge the effect size. The statement that SA treatment expands the lineage potential of ESCs needs to be supported by appropriate and statistically strong data beyond the increase in some marker genes (which are also expressed in embryonic tissues).

7) Please provide direct evidence for disruption of heme biosynthesis directly instructing cells to take on a 2CLC state. How strong is the effect of inhibiting heme synthesis and succinate in inducing the 2CLC population as the absolute number of Zscan4+ or MuERVL+ cell is < 3%? The authors could use other 2CLC inducers (such as acetate (Rodriguez-Terrones, 2020) and RA (Iturbide, A., 2021) as a control, for comparison. The functional assay of 2 cell-like cells after heme synthesis inhibition should be test in vivo by 8-cell or blastocyst injection.

8) Quantification and comparison of pan succinate levels in Figures4c and 5a would be required for interpretation.

9) What happens with positive enriched genes in the SA-treated condition? Upregulation plot should be added to Figure 2a and discussed.

10) In the section describing the upregulation of 2-cell-stage-related genes when SA is added, please add a scramble as control, treat 2iL cells with an inhibitor of a different metabolic pathway for 48h and analyze by rt-PCR a) the 2-cell-related genes, b) the differentiation towards trophoblast cells.

11) The authors provide a very detailed molecular analysis on the effect of heme synthesis inhibition on the pluripotency exit. However, functional characterization (such as teratoma formation) is important to tell whether these naive mESC after treatment with heme synthesis blocker have any defects on the pluripotency exit and differentiation.

12) Although it is clear that heme synthesis is important for the mouse pluripotency status transition, it is not clear whether this is a mechanism present in human as well. The authors presented bioinformatic data suggesting that this might be the case. Does heme synthesis blockade increase the emergence of human 2 cell-like cells?

13) A genetic approach would complement and strengthen the specificity of the conclusions obtained with chemicals. For example, the inhibition of heme synthesis by SA should also be confirmed by knockout or knockdown of proteins in this pathway like ALAD, PBGD et al.

Reviewer #1 (Recommendations for the authors):

The authors will find below a list of my concerns.

1. For readers of the stem cell field a more detailed introduction into the heme synthesis pathway needs to be included in the introduction,

2. The statement made in lines 150-152 and shown as Suppl. Figure 2b should be further supported by proper quantification using replicate assays.

3. Lines 207-209. The rationale and the conclusion behind the experiment shown in Suppl. Figure 4 must be better explained.

4. Line 98: the word 'prevented' is too strong for the observed effect. Oct6, Otx2 and Dnmt3b are clearly upregulated, merely less than in WT.

5. The authors could put their findings in context of data on the impact of modulating the intracellular αKG/succinate ratio on pluripotency maintenance (https://doi.org/10.1038/nature13981; https://doi.org/10.15252/embj.201899518).

6. Does BM also rescue the exit from pluripotency phenotype?

7. Quantification and comparison of pan succinate levels in Figures4c and 5a would be helpful and required for interpretation. As there is basically no staining detectable in untreated cells in 4a and some in 5a, it is unclear whether SA and AA5 effects are similar in size. It appears that AA5 treatment has a much weaker effect.

Reviewer #2 (Recommendations for the authors):

– In the first part, regarding the analysis of previously published results of CRISPR-Cas9 screens (lines 77-87): It would be ideal to use an alternative annotation tool for gene ontology analyses. DAVID is already outdated. If the finding is corroborated with an alternative method, it not only gives strength to the initial observations but allows finding new targets of interest.

– The second title, "Heme synthesis inhibition prevents the activation of key developmental signaling pathways," needs to be more specific. Because heme is a critical component of general cell homeostasis, it is evident that the inhibition will affect key signaling pathways, not only in development but in any stage of mammal's life.

In this same section, what happens with positive enriched genes in the SA-treated condition? It should be added the up-regulation plot side Figure 2a, and a correspondent discussion about it.

– In the section describing the upregulation of 2-cell-stage-related genes when SA is added, it is needed to add a scramble-inhibition as control. Treat 2iL cells with an inhibitor of a different metabolic pathway for 48h and analyze by rt-PCR (a) the 2-cell-related genes, (b) the differentiation towards trophoblast cells.

– What happens with the glycine levels after heme synthesis inhibition? Does glycine play a role in the induction of a 2-cell-like state? Ideally authors should include an experiment to test this role.

– The following is optional, but it could be a considerable improvement to provide the bulk RNA seq comparing the inhibition of SUCNR1 vs. the activation induced by succinate, to see all other pathways that could crosstalk and describe better the mechanism of the succinate in pluripotent/totipotent states.

– The authors provide a very detailed molecular analysis on the effect of heme synthesis inhibition on the pluripotency exit. However, functional characterization (such as teratoma formation) is important to tell whether these naive mESC after treatment with heme synthesis blocker have any defects on the pluripotency exit and differentiation.

– Although it is clear that heme synthesis is important for the mouse pluripotency status transition, it is not clear whether this is a mechanism present in human as well. The authors presented bioinformatic data suggesting that this might be the case. It is important that the authors perform functional studies in the human cells to further substantiate this point.

– It is not clear how the cell senses heme deficiency and how heme deficiency results in inability to activate ERK/MAPK and TGF β signaling. Could the authors provide comments into this aspect?

– How strong the effect of inhibiting heme synthesis and succinate in inducing the 2CLC population are as the absolute number of Zscan4+ or MuERVL+ cell are quite low (< 3%)? I suggest that the authors include other 2CLC inducers such as acetate (Rodriguez-Terrones, 2020) and RA (Iturbide, A., 2021) as a control for comparison.

– The authors performed the 2CLC experiment with 2iL condition. However, a quick review of the literature indicates that many of these studies have been done in FBS/LIF system. Can the authors comment on the difference between 2iL condition and FBS/LIF in the study of 2CLC?

– The authors analyzed that heme synthesis is also important human pluripotent state transition. It will be interesting to see the scenario in human pluripotent stem cells, which will largely make the study difference, especially if the heme synthesis blockade will also increase the emergence of human 2 cell-like cells.

– The functional assay of 2 cell-like cells after heme synthesis inhibition should be tested in vivo by 8-cell or blastocyst injection

– The whole study of inhibition is chemical achieved. As a supplement, the authors should also test with genetic intervention instead of sole chemicals. For example, the inhibition of heme synthesis by SA should also be confirmed by knockout or knockdown of proteins in this pathway like ALAD, PBGD et al. Also similar with AA5.

– To make the finding here more convincing and significant, it will be good to confirm in embryos.

– The manuscript should include the metabolism part in the introduction instead of using large paragraphs in the main text to introduce these pathways and previous studies.

Essential revisions:

1) Are the levels of heme itself, or the mentioned 7 enzymes of the heme pathway regulated during differentiation?

We thank the reviewer to point that it is important to verify whether there could be a regulation of heme biosynthesis during the naïve-to-primed transition. To answer this, we explored the expression of the 8 genes encoding enzymes in the pathway, from this study and previously published ones, both in human and mouse and from in vivo blastocysts or in vitro models (Grow et al., 2015; Sperber et al., 2015; Di Stefano et al., 2018; Nakamura et al., 2016). While some of the enzymes show a slight change during the exit of the naïve state, none are consistent between the different studies and no significant trend for the pathway is observed. This has been added as Table 2.

2) Further experiments are needed to determine whether the lack of proper Tgf β and FGF-ERK signalling activation are cause or consequence of the observed differentiation defects. How do cells sense heme deficiency and does how heme deficiency result in inability to activate ERK/MAPK and TGF β signaling?

3) Can the authors attempt to titrate pathway inhibition to a similar level as observed in heme pathway deficient ESCs? Furthermore, can the differentiation defect be rescued upon overstimulation of FGF-ERK and TGF β to reach WT levels?

We thank the reviewers for their highly relevant comments. We have addressed these two questions simultaneously.

Heme deficiency is canonically sensed in mammals through the nuclear factor BACH1 (reviewed in (Zhang et al., 2018)) or the activation of the Integrated Stress Response (ISR) through the Heme Responsive Inhibitor (HRI) or EIF2AK1 (reviewed in (Pakos‐Zebrucka et al., 2016)).

BACH1 heterodimerized with MAF proteins (among others) binds to DNA and represses the transcription of target genes. Upon heme binding to BACH1, the complex dissociates and BACH1 is exported in the cytosol, thus relieving the repression of target genes. As seen in these immunofluorescence analyses of BACH1 subcellular localization (Figure 2 – Supplement Figure 1a), the abundance of nuclear BACH1 is comparable in 2iL, EPI, and EPI +SA conditions, indicating that the inhibition of heme synthesis does not affect BACH1 activity. On the opposite and as expected, the addition of hemin provokes the nuclear exclusion of BACH1. This data indicates that BACH1 is not responsible for the heme deficiency sensing in our model. This was added to the manuscript.

As another putative way of heme depletion to act on the transition from naïve to primed, we also investigated the role of the ISR. Indeed, various environmental and pathological conditions, including protein homeostasis (proteostasis) defects, nutrient deprivation, viral infection, and oxidative stress, activate the ISR to restore the balance by reprogramming gene expression, all converging to the phosphorylation and activation of the eukaryotic translation initiation factor eIF2. Heme depletion is known to be one of the stresses activating this pathway, allowing the phosphorylation and activation of the heme-regulated inhibitor (HRI) kinase, in turn phosphorylating EIF2α. To assess the activation of this pathway, we targeted two levels: the phosphorylation of EIF2α by western blot analysis and the global protein synthesis levels with a SUnSET assay (Schmidt et al., 2009) (Figure 2 – Supplement Figure 1b-c). The increase in EIF2α phosphorylation (Figure 2 – Supplement Figure 1b), a direct result of the activity of HRI in absence of heme, indicates the activation of a kinase upstream of the pathway. The global reduction in protein synthesis, observed with the levels of puromycin-labelled peptides (Figure 2 – Supplement Figure 1c), further confirms the activation of the ISR upon inhibition of heme synthesis.

To further explore the possible implication of this ISR-HRI axis in the inhibition of native-to-primed transition induced by heme synthesis inhibition, we used a chemical activator of HRI (BTdCPU; (Chen et al., 2011)) to assess whether it could induce a similar effect as SA or not. Treatment of cells with 2µM BTdCPU, reduces protein synthesis to a similar extent that SA (Figure 2 – Supplement Figure 1d), but fails at preventing the naïve stage exit (Figure 2 – Supplement Figure 1e), indicating an HRI-independent mechanism. These results were added to the manuscript and are now part of Figure 2 – Supplement Figure 1.

In a final attempt to identify the mechanism behind the failure of MAPK and TGFβ activation following heme synthesis inhibition, we also searched the GO terms associated with heme binding that could regulate the pathways or receptor mentioned. The analysis did not reveal any putative regulators. However, following an interesting suggestion from the reviewers, we observed that an increase in Activin A or FGF2 concentration (2- or 3-fold increase) did not rescue the naïve exit defect (Figure 2 – Supplement Figure 1j), suggesting a strong intracellular mechanism. Since activin and FGF are the sole signals used to trigger the exit, an impairment of those would indeed prevent it, especially for TGFβ pathway, as shown in the supplementary figure 3i. Thus, instead of chemically block the TGFb pathway, we decided to investigate the temporal activation of SMAD3 phosphorylation during the exit of naïve stage, with or without SA (Figure 2b-c). Quantification of the phosphorylation kinetics of SMAD3 indicates a significant decrease starting after 24h post induction of the exit when treated with SA (Figure 2b-c). This correlates with the gene expression kinetics of Fgf5 and Tbx3 (Figure 2d), suggesting a key role of the SMAD3 pathway perturbation in the effect.

Mechanistically, it is known that heme synthesis inhibition raises the levels of succinyl-CoA and succinate, since 8 moles of succinyl-CoA and 8 moles of glycine are needed to produce 1 mole of heme, heme synthesis acting thus as a “succinyl-CoA sink” (Atamna 2004). In addition, succinate accumulation upon heme synthesis inhibition is reinforced by the drastically reduced abundance of succinate dehydrogenase (SDH) consuming succinate in the TCA that consumes succinate. Indeed, SDH is a heme-dependent complex known to be destabilized by the loss of heme nicely reviewed in (Kim et al., 2012), as confirmed in Figure 2e, showing the drastically reduced SDH abundance after heme synthesis inhibition. Together this indicates a possible accumulation of succinate in response to heme synthesis inhibition, an accumulation that could cause the transition defect. To test this hypothesis, we used butylmalonate as an inhibitor of the dicarboxylate carrier (DIC; SLC25A10) to prevent leakage of succinate from the mitochondrial matrix to the cytosol, and this nicely resumed the transition as shown by the gene expression profile of mESCs in the EpiLC + SA + BM condition (Figure 2f). Succinate accumulation and/or SDH ablation is known to increase protein lysine succinylation (Smestad et al., 2018), an effect that is observed upon SA treatment (Figure 2g) and that we mimicked by inhibiting Sirt7, a cytosolic and nuclear desuccinylase (Figure 2g). The ability of this molecule to mimic SA in the pathway inhibition (Figure 2h-i) and to prevent the exit of the naïve state (Figure 2f) points at succinylation events as the cause of the transition defects.

This new data is now inserted, commented and discussed in the manuscript as part of figure 2 or Figure 2 – Supplement Figure 1.

4) Analyse cell proliferation and viability after SA (or AA5) treatment. If there is an impact on cellular health, this needs to be reported and taken into careful consideration when interpreting results.

We used a live:dead staining kit (Invitrogen L3224, using calcein-AM and ethidium homodimer-1) to analyze the survival fraction of cells after two days in culture, following either SA or AA5 treatment. No significant change was observed (Author response image 1A). The cell cycle analysis was performed by flow cytometry with propidium iodide. Overall, SA treatment increases slightly the G2 cell population, while AA5 increases the S population (Author response image 1B). Since these two drugs induced similar effects in the acquisition of 2CLCs in the naïve population, it is thus unlikely that the effects of these molecules on cell viability and proliferation do affect the results obtained and presented.

Author response image 1

Download asset Open asset

5) The statement made in lines 150-152 and shown as Suppl. Figure 2b should be further supported by proper quantification using replicate assays.

6) Expression levels in TS cells or in trophoblast tissue must be used as control to judge the effect size. The statement that SA treatment expands the lineage potential of ESCs needs to be supported by appropriate and statistically strong data beyond the increase in some marker genes (which are also expressed in embryonic tissues).

Questions 5 and 6, both related to the increase in cell potential, are thus treated simultaneously. While further quantification of gene expression (Cdx2, Tead4, Gata3, Eomes, Hand1, Elf5, Krt18, Cited1, Tfap2c and Ets1) following SA or AA5 treatment still shows a trend toward an increase of differentiation efficiency (Author response image 2), we observed high variability and the increase in the expression of trophoblast-specific genes was not significant (n=6). While this still could be interesting, we hypothesize that since the number of 2CLCs is increased from 1 to 3 % upon SA or AA5 treatment, the effect on lineage potential will remain limited (see also the response to question 7). Because of the lack of significance, we removed the data and related discussion on the extended potential from the manuscript.

Author response image 2

Download asset Open asset

7) Please provide direct evidence for disruption of heme biosynthesis directly instructing cells to take on a 2CLC state. How strong is the effect of inhibiting heme synthesis and succinate in inducing the 2CLC population as the absolute number of Zscan4+ or MuERVL+ cell is < 3%? The authors could use other 2CLC inducers (such as acetate (Rodriguez-Terrones, 2020) and RA (Iturbide, A., 2021) as a control, for comparison. The functional assay of 2 cell-like cells after heme synthesis inhibition should be test in vivo by 8-cell or blastocyst injection.

The updated figure 5c now includes RA as a positive control as presented in Iturbide et al. 2021 and as suggested by the reviewers. Interestingly, we observe that the gene expression levels of the 2CLC markers after RA treatment increase to a similar extent than with AA5, but not SA.

These authors showed that the RA-induced 2CLC increase is due to activation of the RARγ since inhibition of RXR (with the HX531 inhibitor) does not rescue the effect while a RAR inhibitor (AGN193109) does. Very interestingly, the AA5-induced 2CLC signature seems to depend on both receptors as HX531 (RXR inhibitor) and AGN193109 (pan-RAR inhibitor) do inhibit the AA5-induced 2CLC gene expression signature.

Since genetic ablation of heme synthesis enzymes is embryonic lethal around the implantation time (Lindberg et al., 1996; Bensidhoum et al., 1998; Phillips et al., 2001; Conway et al., 2017; Magness et al., 2002), we expect this experiment to bear toxicity for the embryo, especially since accumulation of heme synthesis intermediate is known for its toxicity (Lämsä et al., 2012; Handschin et al., 2005). Furthermore, injection of the 2CLC in early embryos is rarely used, including in Iturbide, 2021 or Rodriguez-Terrones, 2020.

8) Quantification and comparison of pan succinate levels in Figures4c and 5a would be required for interpretation.

We thank the reviewers for pointing at this issue. We performed a quantification of the pan succinyllysine residue levels by flow cytometry and the results are now shown in addition to the more qualitative confocal micrographs in figures 4 and 5. These results better support our claims and conclusion.

9) What happens with positive enriched genes in the SA-treated condition? Upregulation plot should be added to Figure 2a and discussed.

The upregulation plot has been added to figure 2a. A majority of the Gene ontology (GO) terms found upregulated in EpiLC SA vs EpiLC includes cytochrome-associated proteins for xenobiotic detoxification or oxidative phosphorylation. This probably corresponds to a compensatory mechanism since the absence of heme reduces the abundance of proteins encoded by to those genes, as previously shown (Lämsä et al., 2012; Vinchi et al., 2014).

10) In the section describing the upregulation of 2-cell-stage-related genes when SA is added, please add a scramble as control, treat 2iL cells with an inhibitor of a different metabolic pathway for 48h and analyze by rt-PCR a) the 2-cell-related genes, b) the differentiation towards trophoblast cells.

To validate the metabolic effect of SA or AA5 on the 2CLC population, we thank the reviewers for suggesting using alternative metabolism inhibitors. We used two different metabolic inhibitors: Etomoxir as an inhibitor of Carnitine Palmitoyl Transferase (CPT-1), thus blocking fatty acid oxidation, and 6-aminonicotinamide (6-AN), an antimetabolite blocking the pentose phosphate pathway (PPP) metabolism. These two molecules fail to induce a 2CLC signature (Author response image 3).

Author response image 3

Download asset Open asset

11) The authors provide a very detailed molecular analysis on the effect of heme synthesis inhibition on the pluripotency exit. However, functional characterization (such as teratoma formation) is important to tell whether these naive mESC after treatment with heme synthesis blocker have any defects on the pluripotency exit and differentiation.

As explained in the Discussion section, the fact that heme synthesis knock-out is embryonic lethal around the implantation stage in mouse (Lindberg et al., 1996; Bensidhoum et al., 1998; Phillips et al., 2001; Conway et al., 2017; Magness et al., 2002) strongly suggests an implantation defect, hence, pluripotency exit. This is further supported by the loss of Oct4 in ALAD KO mESCs after removal of hemin (Supp. Figure 2b). Furthermore, teratoma formation assay or embryoid body formation assays after a temporary blockade (2 days) of heme synthesis through SA would quickly be attenuated as heme synthesis would resume after injection, especially since these assays extends for weeks.

12) Although it is clear that heme synthesis is important for the mouse pluripotency status transition, it is not clear whether this is a mechanism present in human as well. The authors presented bioinformatic data suggesting that this might be the case. Does heme synthesis blockade increase the emergence of human 2 cell-like cells?

To answer this question, we used Elf1 hESC cells (Ware et al., 2014) grown in RSeT media to mimic the naïve conditions, and triggered the exit toward the primed stage for 4 days in TeSR media (both commercially available, StemCell technologies), with or without 0.5 mM of SA. We observe a significant prevention of the loss of naïve markers, as in mESCs, while not preventing the increase in a primed signature. This has been updated and commented in the manuscript, with an updated Figure 1 – Supplement Figure 1c.

In addition, naïve Elf1 hESCs maintained either in RSeT medium or in 2iL-I-F (GSK3 and MEK inhibitors, LIF, IGF1 and FGF2) media (Sperber et al., 2015) treated with 0.5 mM of SA for 2 days, show a significant upregulation of markers associated with the 8-cell stage (the human homolog of the 2CLC), corresponding to the zygote genome activation event (Taubenschmid-Stowers et al., 2022), suggesting a similar mechanism (Figure 3g, Author response image 4).

Author response image 4

Download asset Open asset

13) A genetic approach would complement and strengthen the specificity of the conclusions obtained with chemicals. For example, the inhibition of heme synthesis by SA should also be confirmed by knockout or knockdown of proteins in this pathway like ALAD, PBGD et al.

We thank the reviewer for this important comment. Using the CRISPR-Cas9 technology, we generated an ALAD KO E14-mESC line using CRISPR/Cas9 (Figure 1 – Supplement Figure 1a). To maintain a healthy population, this line is grown in a medium supplemented with hemin and removal of this supplement allows to reveal the heme synthesis inhibition. As shown in the updated figure 3 (Figure 3d), removal of hemin in the culture medium for 2 days induces a strong increase in 2CLC marker expression.

We also performed the transition from naïve to primed in these ALAD KO mESCs. Genetic ablation of ALAD also prevents mESCs to properly exit the naïve state as the naïve marker expression is maintained. On the other hand, primed markers, except Dnmt3b, are upregulated (Figure 1 – Supplement Figure 1b). While this is in apparent contrast to the chemical inhibition of ALAD, the complete loss of ALAD seems to perturb the pluripotency state as seen by the complete loss of Oct4 expression, thus impairing the proper transition. This loss of Oct4 transcript could be due to the complete loss of heme, impairing the formation of G-quadruplexes, known to promote Oct4 gene expression (Renčiuk et al., 2017; Gray et al., 2019).

References

Bensidhoum M, Larou M, Lemeur M, Dierich A, Costet P, Raymond S, Daniel JY, De Verneuil H and Ged C (1998) The disruption of mouse uroporphyrinogen III synthase (uros) gene is fully lethal. Transgenics 2: 275–280

Chen T, Ozel D, Qiao Y, Harbinski F, Chen L, Denoyelle S, He X, Zvereva N, Supko JG, Chorev M, et al. (2011) Chemical genetics identify eIF2α kinase heme-regulated inhibitor as an anticancer target. Nat Chem Biol 7: 610–616

Conway AJ, Brown FC, Fullinfaw RO, Kile BT, Jane SM and Curtis DJ (2017) A mouse model of hereditary coproporphyria identified in an ENU mutagenesis screen. DMM Dis Model Mech 10: 1005–1013

Gray LT, Puig Lombardi E, Verga D, Nicolas A, Teulade-Fichou MP, Londoño-Vallejo A and Maizels N (2019) G-quadruplexes Sequester Free Heme in Living Cells. Cell Chem Biol 26: 1681-1691.e5

Grow EJ, Flynn RA, Chavez SL, Bayless NL, Wossidlo M, Wesche DJ, Martin L, Ware CB, Blish CA, Chang HY, et al. (2015) Intrinsic retroviral reactivation in human preimplantation embryos and pluripotent cells. Nature 522: 221

Handschin C, Lin J, Rhee J, Peyer AK, Chin S, Wu PH, Meyer UA and Spiegelman BM (2005) Nutritional regulation of hepatic heme biosynthesis and porphyria through PGC-1alpha. Cell 122: 505–515

Kim HJ, Khalimonchuk O, Smith PM and Winge DR (2012) Structure, function, and assembly of heme centers in mitochondrial respiratory complexes. Biochim Biophys Acta 1823: 1604–1616

Lämsä V, Levonen AL, Sormunen R, Yamamoto M and Hakkola J (2012) Heme and heme biosynthesis intermediates induce heme oxygenase-1 and cytochrome P450 2A5, enzymes with putative sequential roles in heme and bilirubin metabolism: different requirement for transcription factor nuclear factor erythroid- derived 2-like 2. Toxicol Sci 130: 132–144

Lindberg RLP, Porcher C, Grandchamp B, Ledermann B, Bürki K, Brandner S, Aguzzi A and Meyer UA (1996) Porphobilinogen deaminase deficiency in mice causes a neuropathy resembling that of human hepatic porphyria. Nat Genet 12: 195–199

Magness ST, Maeda N and Brenner DA (2002) An exon 10 deletion in the mouse ferrochelatase gene has a dominant-negative effect and causes mild protoporphyria. Blood 100: 1470–1477

Nakamura T, Okamoto I, Sasaki K, Yabuta Y, Iwatani C, Tsuchiya H, Seita Y, Nakamura S, Yamamoto T and Saitou M (2016) A developmental coordinate of pluripotency among mice, monkeys and humans. Nature 537: 57–62

Pakos‐Zebrucka K, Koryga I, Mnich K, Ljujic M, Samali A and Gorman AM (2016) The integrated stress response. EMBO Rep 17: 1374–1395

Phillips JD, Jackson LK, Bunting M, Franklin MR, Thomas KR, Levy JE, Andrews NC and Kushner JP (2001) A mouse model of familial porphyria cutanea tarda. Proc Natl Acad Sci 98: 259–264

Renčiuk D, Ryneš J, Kejnovská I, Foldynová-Trantírková S, Andäng M, Trantírek L and Vorlíčková M (2017) G-quadruplex formation in the Oct4 promoter positively regulates Oct4 expression. Biochim Biophys acta Gene Regul Mech 1860: 175–183

Schmidt EK, Clavarino G, Ceppi M and Pierre P (2009) SUnSET, a nonradioactive method to monitor protein synthesis. Nat Methods 6: 275–277

Sperber H, Mathieu J, Wang Y, Ferreccio A, Hesson J, Xu Z, Fischer KA, Devi A, Detraux D, Gu H, et al. (2015) The metabolome regulates the epigenetic landscape during naive-to-primed human embryonic stem cell transition. Nat Cell Biol 17: 1523–1535

Di Stefano B, Ueda M, Sabri S, Brumbaugh J, Huebner AJ, Sahakyan A, Clement K, Clowers KJ, Erickson AR, Shioda K, et al. (2018) Reduced MEK inhibition preserves genomic stability in naive human embryonic stem cells. Nat Methods 15: 732–740

Taubenschmid-Stowers J, Rostovskaya M, Santos F, Ljung S, Argelaguet R, Krueger F, Nichols J and Reik W (2022) 8C-like cells capture the human zygotic genome activation program in vitro. Cell Stem Cell 29: 449-459.e6

Vinchi F, Ingoglia G, Chiabrando D, Mercurio S, Turco E, Silengo L, Altruda F and Tolosano E (2014) Heme exporter FLVCR1a regulates heme synthesis and degradation and controls activity of cytochromes P450. Gastroenterology 146: 1325–1338

Ware CB, Nelson AM, Mecham B, Hesson J, Zhou W, Jonlin EC, Jimenez-Caliani AJ, Deng X, Cavanaugh C, Cook S, et al. (2014) Derivation of naive human embryonic stem cells. Proc Natl Acad Sci 111: 4484–4489

Zhang X, Guo J, Wei X, Niu C, Jia M, Li Q and Meng D (2018) Bach1: Function, Regulation, and Involvement in Disease. Oxid Med Cell Longev 2018

Reviewer #1 (Recommendations for the authors):

The authors will find below a list of my concerns.

1. For readers of the stem cell field a more detailed introduction into the heme synthesis pathway needs to be included in the introduction.

A description of the heme synthesis pathway is now included in the introduction section of the revised version, line 65-67. The following has been added: “This pathway, starting in mitochondria, uses succinyl-CoA and glycine as starting material. It then proceeds to successive cytosolic reactions before ending by the formation of the heme molecule in the mitochondrial matrix.”

2. The statement made in lines 150-152 and shown as Suppl. Figure 2b should be further supported by proper quantification using replicate assays.

While further quantification of gene expression (Cdx2, Tead4, Gata3, Eomes, Hand1, Elf5, Krt18, Cited1, Tfap2c and Ets1) following SA or AA5 treatment to validate the increase in lineage potential still shows an increase (Author response image 3), we observed high variability and the increase in the expression of trophoblast-specific genes was not significant (n=6). While this still could be interesting, we hypothesize that since the number of 2CLCs is increased from 1 to 3 % upon SA or AA5 treatment, the effect on lineage potential will remain limited (see also the response to the Editor’s question 7). Because of the lack of significance, we removed the data and related discussion on the extended potential from the revised manuscript.

3. Lines 207-209. The rationale and the conclusion behind the experiment shown in Suppl. Figure 4 must be better explained.

The lines 251-254 of the manuscript have been amended to clarify the conclusion of Suppl. Figure 4 (now Figure 5 – Supplement Figure 1 in the revised manuscript). The following is now stated: “Of these three classes, we ruled out a role for PHDs, responsible for HIF1α degradation, as the abundance of this transcription factor was not increased by SA nor by AA5 treatment, demonstrating a lack of stabilization following a putative PHD inhibition (Figure 5 – Supplement Figure 1).”

4. Line 98: the word 'prevented' is too strong for the observed effect. Oct6, Otx2 and Dnmt3b are clearly upregulated, merely less than in WT.

We updated the line 98 to more accurately phrase the observed effect.

5. The authors could put their findings in context of data on the impact of modulating the intracellular αKG/succinate ratio on pluripotency maintenance (https://doi.org/10.1038/nature13981; https://doi.org/10.15252/embj.201899518).

We amended the end of the Discussion section to include the missing reference and commented further on the difference between these studies and ours.

6. Does BM also rescue the exit from pluripotency phenotype?

We thank reviewer 1 for this suggestion. Indeed, addition of BM to SA during the naïve state exit rescues the observed phenotype, revealing succinate as a mechanistic actor. This is now shown in the updated figure 2 (Figure 2f).

7. Quantification and comparison of pan succinate levels in Figures4c and 5a would be helpful and required for interpretation. As there is basically no staining detectable in untreated cells in 4a and some in 5a, it is unclear whether SA and AA5 effects are similar in size. It appears that AA5 treatment has a much weaker effect.

In order to further back our claims, we quantified the pan-succinyllysine residue abundance through flow cytometry. These quantifications were added to figures 2g, 4c and 5a. The quantification strengthens the results observed by immunofluorescence.

Reviewer #2 (Recommendations for the authors):

– In the first part, regarding the analysis of previously published results of CRISPR-Cas9 screens (lines 77-87): It would be ideal to use an alternative annotation tool for gene ontology analyses. DAVID is already outdated. If the finding is corroborated with an alternative method, it not only gives strength to the initial observations but allows finding new targets of interest.

We compared the results generated by DAVID with other functional annotation tools. We ran over-representation analysis using the R package ClusterProfiler (Yu et al., 2012), which assess enrichment of terms by calculating p-values using the hypergeometric distribution method described in Boyle et al., 2004. The analysis was performed using GO databases (including biological process, molecular function and subcellular localization) as well as hallmark (h) (Liberzon et al., 2015) and canonical pathways (CP, part of curated database, C) databases of Molecular Signature Database (MSigDB, please see https://www.gsea-msigdb.org/). The results show significant enrichment of terms related to heme metabolism (“heme biosynthetic process”, “porphyrin-containing compound biosynthetic process”, “protoporphyrinogen IX metabolic process”) for both the human and mouse CRISPR-Cas9 screen list (Author response Tables 1 and 2) confirming the results generated by DAVID. We maintain the DAVID results in the main figure as the DAVID tool has been continuously updated and improved from initially reported publications (Jiao et al., 2012; Sherman et al., 2007, 2022). As of today, the last update of the DAVID knowledgebase (v2023q1) was released in April 2023 and is cited in more than 60 000 scientific publications highlighting its robustness.

– The second title, "Heme synthesis inhibition prevents the activation of key developmental signaling pathways," needs to be more specific. Because heme is a critical component of general cell homeostasis, it is evident that the inhibition will affect key signaling pathways, not only in development but in any stage of mammal's life.

In this same section, what happens with positive enriched genes in the SA-treated condition? It should be added the up-regulation plot side Figure 2a, and a correspondent discussion about it.

We thank reviewer 2 for these comments. We accordingly corrected the section 2 as a whole, changing its title and the related figure 2a. The new secondary title is Heme synthesis inhibition prevents the activation of key signaling pathways associated with implantation. The upregulated terms have been also commented in the main text section (lines 132-134).

– In the section describing the upregulation of 2-cell-stage-related genes when SA is added, it is needed to add a scramble-inhibition as control. Treat 2iL cells with an inhibitor of a different metabolic pathway for 48h and analyze by rt-PCR (a) the 2-cell-related genes, (b) the differentiation towards trophoblast cells.

In order to validate the metabolic effect of heme synthesis inhibition we inhibited two other unrelated metabolic pathways: the pentose phosphate pathway with 6-aminonicotinamide and the fatty acid oxidation with Etoxomir. We do not observe any effect of the inhibition of these pathways on the upregulation of the 2CLC program (Figure 1 - Supplement Figure 1c).

– What happens with the glycine levels after heme synthesis inhibition? Does glycine play a role in the induction of a 2-cell-like state? Ideally authors should include an experiment to test this role.

To investigate the putative role of a glycine accumulation on the acquisition of a 2CLC gene signature, we tested the supplementation in 10 mM of glycine of the 2iL naïve medium. As shown in Author response image 5, no gene expression increase is observed.

Author response image 5

Download asset Open asset

– The following is optional, but it could be a considerable improvement to provide the bulk RNA seq comparing the inhibition of SUCNR1 vs. the activation induced by succinate, to see all other pathways that could crosstalk and describe better the mechanism of the succinate in pluripotent/totipotent states.

While comparing the effect we observe on the 2CLC gene signature through heme synthesis or SDH inhibition to the previously published RA induction we noticed that inhibition of RAR and RXR rescued the increase in gene expression (Figure 5c). While this does not pinpoint the exact mechanism, these observations bring light to an intermediate involved.

– The authors provide a very detailed molecular analysis on the effect of heme synthesis inhibition on the pluripotency exit. However, functional characterization (such as teratoma formation) is important to tell whether these naive mESC after treatment with heme synthesis blocker have any defects on the pluripotency exit and differentiation.

As explained in the Discussion section, the fact that heme synthesis knock-out is embryonic lethal around the implantation stage in mouse (Lindberg et al., 1996; Bensidhoum et al., 1998; Phillips et al., 2001; Conway et al., 2017; Magness et al., 2002) strongly suggests an implantation defect, hence, pluripotency exit. This is further supported by the loss of Oct4 in ALAD KO mESCs after removal of hemin (Figure 1 – Supplement Figure 1b). Furthermore, teratoma formation assay or embryoid body formation assays after a temporary blockade (2 days) of heme synthesis through SA would quickly be attenuated as heme synthesis would resume after injection, especially since these assays extends for weeks.

– Although it is clear that heme synthesis is important for the mouse pluripotency status transition, it is not clear whether this is a mechanism present in human as well. The authors presented bioinformatic data suggesting that this might be the case. It is important that the authors perform functional studies in the human cells to further substantiate this point.

To answer this question, we used Elf1 hESC cells (Ware et al., 2014) grown in RSeT media to mimic the naïve conditions, and triggered the exit toward the primed stage for 4 days in TeSR media (both commercially available, StemCell technologies), with or without 0.5 mM of SA. We observe a significant prevention of the loss of naïve markers, as in mESCs, while not preventing the increase in a primed signature (Figure 1 – Supplement Figure 1c).

– It is not clear how the cell senses heme deficiency and how heme deficiency results in inability to activate ERK/MAPK and TGF β signaling. Could the authors provide comments into this aspect?

We thank the reviewers for their highly relevant comments. We have addressed these two questions simultaneously.

Heme deficiency is canonically sensed in mammals through the nuclear factor BACH1 (reviewed in (Zhang et al., 2018)) or the activation of the Integrated Stress Response (ISR) through the Heme Responsive Inhibitor (HRI) or EIF2AK1 (reviewed in (Pakos‐Zebrucka et al., 2016)).

BACH1 heterodimerized with MAF proteins (among others) binds to DNA and represses the transcription of target genes. Upon heme binding to BACH1, the complex dissociates and BACH1 is exported in the cytosol, thus relieving the repression of target genes. As seen in immunofluorescence analyses of BACH1 subcellular localization (Figure 2 – Supplement Figure 1a), the abundance of nuclear BACH1 is comparable in 2iL, EpiLC, and EpiLC+SA conditions, indicating that the inhibition of heme synthesis does not affect BACH1 activity. On the opposite and as expected, the addition of hemin provokes the nuclear exclusion of BACH1. This data indicates that BACH1 is not responsible for the heme deficiency sensing in our model.

As another putative way of heme depletion to act on the transition from naïve to primed, we also investigated the role of the ISR. Indeed, various environmental and pathological conditions, including protein homeostasis (proteostasis) defects, nutrient deprivation, viral infection, and oxidative stress, activate the ISR in order to restores balance by reprogramming gene expression, all converging to the phosphorylation and activation of the eukaryotic translation initiation factor eIF2. Heme depletion is known to be one of the stresses activating this pathway, allowing the phosphorylation and activation of the heme-regulated inhibitor (HRI) kinase, in turn phosphorylating EIF2α. To assess the activation of this pathway, we targeted two levels: The phosphorylation of EIF2α by western blot analysis and the global protein synthesis levels with a SUnSET assay (Schmidt et al., 2009). The increase in EIF2α phosphorylation (Figure 2 – Supplement Figure 1b), a direct result of the activity of HRI in absence of heme, indicates the activation of a kinase upstream of the pathway. The global reduction in protein synthesis, observed with the levels of puromycin-labelled peptides (Figure 2 – Supplement Figure 1c), further confirm the activation of the ISR upon inhibition of heme synthesis.

To further explore the possible implication of this ISR-HRI axis in the inhibition of native-to-primed transition induced by heme synthesis inhibition, we used a chemical activator of HRI (BTdCPU; (Chen et al., 2011)) to assess whether it could induce a similar effect as SA or not. Treatment with 2 µM BTdCPU, reduces protein synthesis to a similar extent that SA (Figure 2 – Supplement Figure 1d), but fails at preventing the naïve stage exit (Figure 2 – Supplement Figure 1e), indicating an HRI-independent mechanism. These results were added to the manuscript in lines 119-126.

– How strong the effect of inhibiting heme synthesis and succinate in inducing the 2CLC population are as the absolute number of Zscan4+ or MuERVL+ cell are quite low (< 3%) ?. I suggest that the authors include other 2CLC inducers such as acetate(Rodriguez-Terrones, 2020) and RA (Iturbide, A., 2021) as a control for comparison.

The updated figure 5 now includes RA as a positive control as presented in Iturbide et al., 2021 and as suggested by the reviewers. Interestingly, we observe that the gene expression levels of the 2CLC markers after RA treatment increase to a similar extent than with AA5, but not SA.

These authors showed that the RA-induced 2CLC increase is due to activation of the RARγ since inhibition of RXR (with the HX531 inhibitor) does not rescue the effect while a RAR inhibitor (AGN193109) does. Very interestingly, the AA5-induced 2CLC signature seems to depend on both receptors as HX531 (RXR inhibitor) and AGN193109 (pan-RAR inhibitor) do inhibit the AA5-induced 2CLC gene expression signature.

Since genetic ablation of heme synthesis enzymes is embryonic lethal around the implantation time (Lindberg et al., 1996; Bensidhoum et al., 1998; Phillips et al., 2001; Conway et al., 2017; Magness et al., 2002), we expect this experiment to bear toxicity for the embryo, especially since accumulation of heme synthesis intermediate is known for its toxicity (Lämsä et al., 2012; Handschin et al., 2005). Furthermore, injection of the 2CLC in early embryos is rarely used, including in Iturbide (2021) or Rodriguez-Terrones (2020).

– The authors performed the 2CLC experiment with 2iL condition. However, a quick review of the literature indicates that many of these studies have been done in FBS/LIF system. Can the authors comment on the difference between 2iL condition and FBS/LIF in the study of 2CLC?

We thank the reviewer to point at this difference. We also observe an increase in the expression of 2CLC marker genes in response to AA5, followed by a reduction when the FBS+LIF basal media was supplemented with BM (Author response image 6). This finding highlights a conserved mechanism with the 2iLIF basal media.

Author response image 6

Download asset Open asset

– The authors analyzed that heme synthesis is also important human pluripotent state transition. It will be interesting to see the scenario in human pluripotent stem cells, which will largely make the study difference, especially if the heme synthesis blockade will also increase the emergence of human 2 cell-like cells.

To answer this question, we used Elf1 hESC cells (Ware et al., 2014) grown in RSeT media to mimic the naïve conditions, and triggered the exit toward the primed stage for 4 days in TeSR media (both commercially available, StemCell technologies), with or without 0.5 mM of SA. We observe a significant prevention of the loss of naïve markers, as in mESCs, while not preventing the increase of a primed signature (Figure 1 – Supplement Figure 1c).

In addition, naïve Elf1 hESCs maintained either in RSeT medium or in 2iL-I-F (GSK3 and MEK inhibitors, LIF, IGF1 and FGF2) media (Sperber et al., 2015) treated with 0.5 mM of SA for 2 days, show a significant upregulation of markers associated with the 8-cell stage (the human homolog of the 2CLC), corresponding to the zygote genome activation event (Taubenschmid-Stowers et al., 2022), highlighting a similar mechanism (Author response image 4; Figure 3g).

– The whole study of inhibition is chemical achieved. As a supplement, the authors should also test with genetic intervention instead of sole chemicals. For example, the inhibition of heme synthesis by SA should also be confirmed by knockout or knockdown of proteins in this pathway like ALAD, PBGD et al. Also similar with AA5.

We thank the reviewer for this important comment. Using the CRISPR-Cas9 technology we generated an ALAD KO E14-mESC line using CRISPR/Cas9 (Figure 1 – Supplement Figure 1a). To maintain a healthy population, this line was grown in a medium supplemented with hemin and removal of this supplement allowed to reveal the heme synthesis inhibition. As shown in the updated figure 3 (Figure 3e), removal of hemin in the culture medium for 2 days induces a strong increase in 2CLC marker expression.

We also performed the transition from naïve to primed in these ALAD KO mESCs. Genetic ablation of ALAD also prevents mESCs to properly exit the naïve state as the naïve marker expression is maintained. On the other hand, primed markers, except Dnmt3b, are upregulated (Figure 1 – Supplement Figure 1b). While this is in apparent contrast to the chemical inhibition of ALAD, the complete loss of ALAD seems to perturb the pluripotency state as seen by the complete loss of Oct4 expression, thus impairing the proper transition. This loss of Oct4 transcript could be due to the complete loss of heme, impairing the formation of G-quadruplexes, known to promote Oct4 gene expression (Renčiuk et al., 2017; Gray et al., 2019).

– The functional assay of 2 cell-like cells after heme synthesis inhibition should be tested in vivo by 8-cell or blastocyst injection

– To make the finding here more convincing and significant, it will be good to confirm in embryos.

As explained in the Discussion section, the fact that heme synthesis knock-out is embryonic lethal around the implantation stage in mouse (Lindberg et al., 1996; Bensidhoum et al., 1998; Phillips et al., 2001; Conway et al., 2017; Magness et al., 2002) strongly suggests an implantation defect, hence, pluripotency exit. This is further supported by the loss of Oct4 in ALAD KO mESCs after removal of hemin (Supp. Figure 1A). While we expect either reduction in survival or development arrest by the treated embryos (heme synthesis or SDH inhibition), obtaining embryos for the experiments was challenging and thus was not performed.

– The manuscript should include the metabolism part in the introduction instead of using large paragraphs in the main text to introduce these pathways and previous studies.

We updated the introduction section in lines 65-67. However, the bulk of the heme synthesis/succinate flow remained in the result section to better guide the reader.