“In science novelty emerges only with difficulty, manifested by resistance, against a background provided by expectation.”
Thomas S. Kuhn, The Structure of Scientific Revolutions
To kiss or not to kiss – that is the question!
I started my research as a demonstrator in the Medical School of a country that was on the wrong side of the Iron Curtain. This affected me not only personally, but also had an impact on my research attitude. Our group, under the tutelage of Jan Neuwirt, was interested in erythropoiesis and the burgeoning research on hematopoietic stem cells of that time. At one-point Jan (“Herr Dozent”; the designation carried over from the Austrian period) remarked to me that it would be important to clarify the pathogenesis of anemia of chronic disease (ACD, currently known as anemia of inflammation). For inexplicable reasons, Jan conceived the idea that ACD was caused by “disturbed” hemoglobinization. He responded to my question “how can I demonstrate it?” by saying something like “take 10 rabbits, divide them in the two groups; one control and the second one with sterile abscesses induced by turpentine oil. Following the development of anemia in the turpentine-treated rabbits, inject both groups with radioactive iron.” I performed this experiment using 59Fe purchased in the GDR; I observed a remarkable increase in 59Fe incorporation into erythrocytes in the experimental group. However, the turpentine-injected animals had reticulocytosis and hemolysis that likely explained my observations with 59Fe. Despite this failure “the iron has entered my soul”; I applied for, and was accepted into, the graduate program shortly before finishing my medical degree. In compliance with Jan’s advice, I began setting up in vitro experiments with reticulocytes obtained from control and turpentine-injected rabbits.
However, a call upon me from the Czechoslovak Socialist Army (ČSA) interrupted my work. We were trained to serve as medical officers at The Military Medical School for six weeks before being assigned to battalions. The lectures were delivered primarily by military physicians, but also by political commissars. One of them gave particularly illuminating talks. In one of his sermons, he warned us about the menaces of American imperialists among whom the most dangerous was the well-known “General Motors”. We had to march to breakfast as an organized troop; I quickly noticed that we walked along the doors with the inscription “Library”. It was a piece of cake to sneak into this room. It was like a gift from the heavens. The library was very well equipped with the international journals; I learned a lot. In one noteworthy instance, I found an earlier published Marcel Bessis’ article entitled “Sideroblastic Anaemia, Mitochondria and Erythroblastic Iron” (Br J Haematol. 11:49, 1965; the BJH Issue based on “A Symposium on Sideroblastic Anemia”). I learned about sideroblastic anemia during my studies, but never saw electron micrographs of “sideroblasts”, i.e., erythroblasts with iron-loaded mitochondria. At that time, I was telling our students that “iron is never free”; so, it was puzzling for me, what’s the chemical form in which iron accumulates in mitochondria. The answer came almost 40 years later with the seminal discovery of mitochondrial ferritin by Sonia Levi, Paolo Arosio and their colleagues (J Biol Chem. 276:24437, 2001). (The product of the so-called frascati [SLC25A37] gene got the name Mitoferrin-1 [Mfrn1]. This is unfortunate nomenclature, since for many people Mfrn1 appears identical to mitochondrial ferritin.)
After this brainwashing (for me enlightening) period, all my colleagues were assigned to battalions all over the country. I was allotted to the military resting centre in the mountains (Giant Mountains; Krkonoše in Czech) close to the Polish border. I have no explanation for these unbelievably lucky periods in my life. Will I somehow pay for them? I wore the uniform only one day, without any objections from the officers who spent their three-week holidays there with their families. I do not remember any serious medical condition except for the suspicion of appendicitis in a young girl. For me, it involved an arrangement of rapid transport to the nearby hospital.
I loved hiking and skiing, but I also worked hard on the review paper for my PhD comprehensive preliminary oral exam. I brought with me several books entitled “Eryhropoiesis” (Jacobson LO & Doyle M, eds,, Grune & Straton, 1962) based on a Conference on Erythropoietin; “The Red Blood Cell” (Bishop C. & Surgenor, DM, eds, Academic Press, 1964 [500 pages!]) and, in particular, “Iron Metabolism” (Gros P, ed., Springer-Verlag, 1964), based on the CIBA-sponsored conference in Aix-en-Provence in 1963. Jay Katz and Jim Jandl wrote a wonderful chapter on “The role of transferrin (Tf) in the transport of iron into the developing red cell”. I believe they were the first investigators to demonstrate that the half-life of Tf-bound iron was ~ 100 min, whereas the half-life of Tf was ~ 8 days; this was proof that Tf was not degraded during iron delivery to cells. I am wondering whether today’s budding young scientists would be able to design such a study. Everyone is familiar with the somewhat profound adage “we stand on the shoulders of giants”, but I am not sure whether people know what these giants did.
During my stay in the mountains, my friend and colleague Jaroslav (a.k.a., Jeff) Prchal sent me copies of two articles by Irving M. London and co-workers. Based on their studies, the authors concluded that “the negative feedback control of heme synthesis by heme affords a mechanism for the regulation of porphyrin synthesis (hemin significantly inhibited he the utilization of glycine-2-14C for heme synthesis in rabbit reticulocytes, whereas the effect on the utilization of ALA-4-14C was variable and less significant; BBRC.18:243, 1965). Coupled with the stimulatory effect of hemin on the synthesis of globin (BBRC.18:236, 1965), this control mechanism may participate in the coordination of the synthesis of heme and of globin”. I was captivated by this elegant control mechanism that maintains the balance between the synthesis of heme and globin. London’s observations significantly affected my own research program.
After finishing my duty of defending socialism by serving in the ČSA, I went to see Neuwirt to tell him that there was no evidence that ACD was caused by “disturbed” hemoglobinization, and that this disorder was represented by a mild-to-moderate normochromic/normocytic anemia characterized by decreased serum iron and total iron-binding capacity with adequate iron stores (Semin Hematol. 4:351, 1966). I said to him that I would like to build my research on London’s experiments from the point of view of cellular iron homeostasis. Jan answered, “Just do what needs to be done”. Since then, I was free to do what I wanted to do, except for the past five years or so. Currently, granting agencies want me to perform primarily applied medical research; it’s a tragedy. The physicist, Wilhelm Röntgen, discovered X-rays, changing medicine forever.
My counter-intuitive hypothesis postulated that the inhibition of ALA synthase in reticulocytes by heme (BBRC.18:243, 1965) would cause the accumulation 59Fe derived from 59Fe-Tf and that biochemical analysis of the 59Fe-containing entities in cell lysates would reveal intermediates in 59Fe transfer from 59Fe-Tf to “heme synthetase” (i.e., ferrochelatase; Blood. 14:486, 1959). As expected, hemin inhibited 59Fe incorporation into heme, but in contrast to my prediction, it also inhibited 59Fe uptake by reticulocytes. When I showed these results to Jan, he told me that I merely confirmed London. I responded, “I am not so sure, wait a few days – patience brings roses”. I realized that I had to use another inhibitor of ALA synthase and chose isonicotinic acid hydrazide (INH). INH is an antagonist of pyridoxine which is essential for glycine and succinyl-CoA to form ALA (J Biol Chem. 226:18, 1957). I observed that INH inhibited heme synthesis, but enhanced 59Fe uptake by reticulocytes by 15 to 20%. Based on many studies that followed (e.g., J Biol Chem. 260:14717, 1985), my co-workers and I unequivocally demonstrated that the feedback inhibition of heme synthesis in erythroid cells involves the inhibition of cellular iron acquisition from Tf. Collectively, our studies revealed that in erythroid cells, iron acquisition, rather than ALA production, is the rate-limiting step in heme synthesis (rev. in Blood. 89:1, 1997). We also explained the discrepancy between London’s and our conclusions by showing that heme inhibited glycine uptake by reticulocytes (Blood. 65:850, 1985). My hypothesis was not counter-intuitive after all, since most people would predict that the inhibition of heme synthesis should cause a decrease in cellular iron uptake.
As I alluded to earlier, I was just a timid boy born on the wrong side of the Iron Curtain. All my research was curiosity-driven and I never expected that I would publish in international journals. Nevertheless, I felt that my above-mentioned results generated important new knowledge and should be published in a renowned journal. I discussed this with Jan, whose reaction was lukewarm. When I mentioned Blood, he exclaimed to me, “you are completely crazy”. I eventually convinced him, and we submitted our manuscript to Blood (Dr. William Dameshek was Editor-in-Chief at the time of our submissions). It took about one year of struggles; our manuscript was accepted for publication January 20, 1969 (Blood. 33:690, 1969).
I became convinced that INH would be a very useful tool in my quest for intermediates in iron transport between Tf and hemoglobin. Based on a study conducted with invaluable help from Jitka Borová, we concluded that “…iron-transferrin complexes enter the reticulocyte cytoplasm, and most of released iron is taken up by mitochondria for heme synthesis. When protoporphyrin IX is not available, iron accumulates inside the mitochondria” (BBA. 320:143, 1973). I vividly remember a dream, in which I asked Alan Jacobs why iron accumulates in mitochondria when heme synthesis is inhibited. He responded “it’s obvious, iron knows where to go!” It is pertinent to mention in this context we later demonstrated that succinylacetone (SA, an inhibitor of ALA dehydratase) blocked 59Fe incorporation into hemoglobin, but contrary to control reticulocytes, 59Fe accumulated not only in mitochondria, but also in ferritin. However, 59Fe in ferritin represented only a very small fraction (less than 4%) of the total cellular 59Fe in SA-treated reticulocytes. When SA- and 59Fe-Tf-pretreated reticulocytes were washed and then treated with protoporphyrin IX (PPIX), the 59Fe in hemoglobin increased while there was virtually no change in the 59Fe content of ferritin (BBA. 720:96, 1982). This experiment also indicated that iron taken up by erythroid mitochondria can leave the organelle only after the iron is inserted into PPIX.
These experiments were crucial in formulating what is now known as the “kiss-and-run” hypothesis. This concept did not start on the spur-of-the moment, but has coalesced for over forty years.
In erythroid cells, more than 90% of iron enters mitochondria where ferrochelatase, the enzyme that inserts Fe2+ into PPIX, resides on the matrix side of the inner mitochondrial membrane. Strong evidence for a specific targeting of iron toward mitochondria came not only from the above-mentioned experiments with INH and succinylacetone, but also hereditary sideroblastic anemia caused by defects in heme synthesis (e.g., Am. Soc. Hematol. Educ. Prog., 2011 pp. 525-531).
Based on these considerations, we have formulated a hypothesis that in erythroid cells a transient mitochondrion-endosome interaction is involved in iron translocation to its destination. We have collected the following experimental evidence to support this hypothesis: 1) in reticulocytes, whose membrane TfRs are decorated with 59Fe-Tf, ~ 40% of cell-associated 59Fe can be detected in heme only 1 min after triggering internalization of the Tf-TfR complexes; 2) iron, delivered to mitochondria via the Tf-TfR pathway, is unavailable to cytoplasmic chelators that are unable to cross organellar membranes; 3) endosomal movement is required for iron delivery to mitochondria; 4) “free” cytoplasmic iron is not efficiently used for heme biosynthesis; and 5) the endosome-mitochondrion interaction increases the levels of mitochondrial iron (Blood. 105:368, 2005; Blood. 110: 125, 2007).
More recently Hamdi et al. (BBA-MCR. 1863:2859, 2016), provided unequivocal evidence that the highly efficient transport of iron toward ferrochelatase in erythroid cells requires a direct interaction between Tf-endosomes and mitochondria (the “kiss-and-run” mechanism). Using a novel method (flow sub-cytometry), we analyzed lysates of reticulocytes after labeling these organelles with different fluorophores. We have identified a double-labeled population representing endosomes interacting with mitochondria, as demonstrated by confocal microscopy. Moreover, we demonstrated that the endosome-mitochondrion association is reversible, since a “chase” with unlabeled holo-Tf caused a time-dependent decrease in the size of the double-labeled population; the dissociation of endosomes from mitochondria does not occur in the absence of holo-Tf. Additionally, mutated recombinant holo-Tf, that cannot release iron, significantly decreased the uptake of 59Fe by reticulocytes and diminished 59Fe incorporation into heme. This suggests that endosomes, which are unable to provide iron to mitochondria, cause a “traffic jam” leading to decreased endocytosis of holo-Tf.
Importantly, the substrate for the endosomal transporter, DMT1, is Fe2+, the redox form of iron which is also the substrate for ferrochelatase. These facts make our hypothesis quite attractive, since the “chaperone”-like function of endosomes may be one of the mechanisms that keeps the concentrations of reactive Fe2+ at extremely low levels in the oxygen-rich cytosol of erythroblasts and reticulocytes, preventing ferrous iron’s participation in the dangerous Fenton reaction.
Despite the overwhelming evidence for the kiss-and-run mechanism, the role of endosomes in distributing intracellular iron is accepted without enthusiasm (Front. Pharmacol. 5:173, 2014), misinterpreted (Front. Pharmacol. 5:45, 2014) or simply ignored. There is only one report (J Cell Biol. 214:831, 2016) confirming the interaction of endosomes with mitochondria in epithelial cells, whose iron dynamics is infinitesimal compared to hemoglobin-synthesizing cells. There is no study that would reveal a flaw in our theory.
Some investigators accept the possibility that “iron-laden endosomes directly contact and transfer iron to mitochondria”, but emphasise that this mechanism may occur only in reticulocytes. However, this conclusion conflicts with the rate of hemoglobin synthesis in reticulocytes compared to erythroblasts. In our experience, reticulocytes take up roughly 10 pmol Fe/106 cells/h from diferric-Tf. Based on this value, it takes approximately 200 h (or 8.3 days) for iron to accumulate in total erythrocyte hemoglobin. This interval is slightly longer than the average erythroid cell maturation time (~5–6 days) but, since iron uptake by reticulocytes is probably somewhat slower than in bone marrow erythroblasts, the agreement is remarkably close. (Ponka P, Sheftel AD. Erythroid Iron Metabolism. In: Iron Physiology and Pathophysiology in Humans, Anderson GJ &McLaren G, eds, Springer Science, Chapter 10, pp191-209, 2012). Hence, the “tsunami” in iron flux occurs at all stages of erythroid maturation. Importantly, no conclusions about actual rates of radioactive iron movement can be made based on CPMs and autoradiographies. Along the same lines, fluorescence techniques are not appropriate for the evaluation of intracellular iron trafficking. Additionally, it would be highly surprising, if a new regulatory mechanism involved in iron homeostasis started at the reticulocyte stage.
We can witness a surge of long-buried concepts. One of the proposed intermediates in the intracellular iron path is ferritin. A strong argument against this claim was collected and reported in Blood (89: 2611, 1997). It was also suggested that the autophagic turnover of ferritin via the nuclear receptor coactivator 4 (NCOA4) is a critical process for regulating intracellular iron bioavailability. The iron flux from the endosome to ferritin and from ferritin to the lysosome is allegedly essential for efficient hemoglobinization. However, this concept cannot withstand critical scrutiny. Reticulocytes, which transport Tf-borne Fe into PPIX with astonishing efficiency (please see above), do not contain lysosomes (Rapoport SM. The Reticulocyte, Boca Raton, FL: CRC, 1986; Ciechanover A. Cell Death Differ. 12:1178, 2005). Importantly, Darshan et al. (Hepatology. 50:852, 2009) demonstrated that the conditional deletion of the ferritin heavy chain in adult mice did not cause any decrease in hematocrit or hemoglobin levels, strongly indicating that ferritin is not involved in hemoglobinization. Additionally, another study (Cell Rep.14:411, 2016) has provided evidence that NCOA4 knockout is not lethal and NCOA4-null mice have only a mild microcytic hypochromic anemia.
Many times throughout my scientific career, I recalled the Aesop’s fable as retold by philosopher Isaiah Berlin: “A fox knows many things, but a hedgehog one important thing”. I always felt that I would have preferred being the hedgehog. As time flies, I don’t exclude the possibility that I won’t be either. However, without trying to be boastful, “All great truths start as blasphemies” (Bernard Shaw).
In any case, it would be premature to write the epitaph: “…My coals are spent, my iron’s gone, My nails are drove, my work is done…”. (Blacksmith’s epitaph in Nettlebed churchyard, commemorating William Strange, d. 1746). It is inevitable that one day this epitaph will be written. I hope my disciples will accept my baton and carry it further.
I wish to express my sincere gratitude to Alex Sheftel and Amel Hamdi for their valuable advice and insightful comments on this essay. I also greatly appreciate their devotion to the kiss-and-run concept and their mammoth contributions in its development. An-Sheng Zhang, Tariq Roshan and Daniel Vyoral are among other enthusiasts, whose findings have provided strong experimental support for this concept.
P.S. I understand that some of our colleagues are reluctant to embrace the kiss-and-run model and have discovered an illuminating story that may explain why. This came from reading about the experience of British explorer William Winwood Reade. He described falling in love with the beautiful daughter of an African king. After pursuing her for many months, he dared to steal a kiss. Unfortunately, things didn’t go so well. The girl, having never encountered this before, screamed before running away in tears. Only later did Reade find out that this princess had interpreted his kiss as an intention to eat her.