FRONTLINES
`
`Iron overload:
`molecular clues to its cause
`
`TIBS 24 – MAY 1999
`
`At present two questions need to
`be solved. First, how do intestinal cells
`measure the levels of body iron stores
`and adjust iron absorption accordingly?
`Second, does mutated HFE perturb TfR
`function and thereby influence the con-
`trol of iron absorption? These two ques-
`tions are tightly linked. If crypt cells do
`not receive enough iron, because TfRs are
`inefficient in the absence of normal HFE,
`they might become iron deprived and
`adjust iron absorption accordingly, once
`they have matured to become villus cells.
`All cells sense and adjust cytoplasmic
`free-iron levels by using iron-regulatory
`proteins (IRPs: IRP-1 and IRP-2)17,18. These
`mRNA-binding proteins become active
`upon iron deprivation and interact with
`RNA-hairpin structures known as iron-
`responsive elements (IREs). Several cyto-
`plasmic mRNAs that encode proteins es-
`sential for iron metabolism contain IREs
`and are regulated by IRPs. The mRNAs
`of H and L ferritins, the iron-storage pro-
`teins, each have an IRE in their 59 un-
`translated region, and their translation
`is repressed by the binding of IRPs. By
`contrast, binding of IRP prevents the
`degradation of TfR mRNA, which has
`five IREs in its 39 untranslated region,
`and TfR translation is increased. Other
`as-yet-unknown target mRNAs involved
`in intestinal iron transport might also
`respond to this post-transcriptional
`regulatory system.
`Because high IRP activity indicates a
`low level of chelatable iron in cells, the
`observation that IRPs are more active in
`macrophages and cells of the duodenum
`in hemochromatosis patients than in
`normal individuals is highly relevant19,20.
`Hemochromatosis patients also have di-
`minished intestinal ferritin expression19,
`which supports the idea that, in these
`patients, cells of the duodenum and
`macrophages are iron deprived in spite of
`excessive levels of body iron. Diminished
`ferritin synthesis might cause inefficient
`trapping of iron on its way through vil-
`lus cells19. Alternatively, increased IRP
`activity or low levels of intracellular iron
`might influence the synthesis or func-
`tion of iron-transport proteins directly.
`However, it is not clear whether hemo-
`chromatosis affects apical iron trans-
`port into cells or basolateral transport
`out of cells. Two groups have postulated
`that the mRNA for the metal-transport
`protein DCT1 (also known as Nramp2) is
`a target for IRPs, because it contains a
`sequence that resembles an IRE21,22.
`DCT1 mRNA is synthesized at high lev-
`els in the intestine and induced upon
`iron deprivation21. However, regulation
`
`More than 5% of Caucasians carry a de-
`fective allele of the HFE gene, and one in
`300 individuals has both alleles mutated
`and is, therefore, genetically predisposed
`to the iron-overload disorder known as
`idiopathic hemochromatosis1. Normally,
`iron stores in the organism are controlled
`by tuning intestinal iron absorption to the
`body’s demand for iron. Such regulation
`is defective in hemochromatosis. Patients
`absorb two- to three-fold more iron than
`is normal from food, and the metal accu-
`mulates in the liver, the pancreas and
`heart muscle. The disorder manifests it-
`self mainly in men of >40 years of age and
`provokes liver cirrhosis, diabetes and/or
`heart failure.
`The recent isolation of the HFE gene by
`positional cloning2 has launched excit-
`ing mechanistic studies of the regulation
`of iron absorption and the cause of iron
`overload. Interestingly, the problem en-
`compasses basic cell-biology questions
`that initially were not suspected to be in-
`volved. The first surprise is that the de-
`fective HFE allele encodes a nonclassical
`HLA-type protein that has a missense mu-
`tation, Cys282fiTyr, in the a3 domain2.
`This defect prevents its association with
`b
`2-microglobulin and, as shown for other
`MHC class I proteins, abolishes transport
`of HFE to the cell surface3,4. Given that
`hemochromatosis is a recessive trait, de-
`ficient HFE surface expression appears
`to cause, directly or indirectly, excessive
`iron absorption by the intestine. Indeed,
`knockout mice that lack either HFE or
`b
`2-microglobulin show a similar iron-
`overload phenotype to that observed
`in hemochromatosis patients5–7.
`How can a missing HFE surface mol-
`ecule influence iron absorption? The
`problem is more complex than one
`would expect. HFE is found intracellu-
`larly in perinuclear regions8,9, presum-
`ably the site of its biosynthesis in the
`rough endoplasmic reticulum, as well as
`in peripheral vesicular structures9 and at
`the cell surface8,9. As in the cases of other
`HLA molecules, expression is restricted
`to the basolateral surface in polarized
`epithelial cells8. HFE might therefore con-
`trol the rate of transport of iron out of
`intestinal villus cells at their basolateral
`surface. However, immunoprecipitation
`164
`
`experiments have not revealed a direct
`interaction between HFE and known iron
`transporters. Instead, the second sur-
`prise is that HFE associates with the
`transferrin receptor (TfR)10,11. HFE binds
`to TfR soon after biosynthesis in the
`rough endoplasmic reticulum and is then
`brought to the cell surface in a complex9.
`This finding agrees with the known local-
`ization of TfR to the basolateral surface
`and to endosomal vesicles.
`Can the association between HFE and
`TfR explain iron overload? The answer
`at first appears to be ‘no’: the classical
`pathway for TfR recycling captures iron-
`loaded transferrin from serum, which,
`after receptor-mediated endocytosis, re-
`leases iron into early endosomes12. Thus,
`TfRs probably facilitate cellular iron up-
`take rather than facilitate its release
`(Fig. 1). Should we rethink the classical
`pathway and postulate that it can func-
`tion in the opposite direction in villus
`cells? Probably not: it seems very un-
`likely that iron could rebind to apotrans-
`ferrin in the low-pH environment of en-
`dosomes, and dissociation of iron-loaded
`transferrin from its receptor would be
`highly inefficient at the neutral pH of the
`basolateral cell surface13,14. Moreover,
`little evidence for TfR expression in dif-
`ferentiated intestinal villus cells exists.
`Immunolocalization reveals that both
`HFE and TfR reside mainly in the intesti-
`nal crypt cells8,15,16. The interference of
`mutated HFE with intestinal iron metab-
`olism might, therefore, take place in the
`crypt cells rather than in differentiated
`villus cells. Given that crypt cells are
`the immature precursors of villus cells, a
`dysregulation in crypts might ultimately
`influence iron absorption in villi.
`Crypt cells are ideally suited to meas-
`uring the levels of body iron stores: they
`depend on iron from serum transferrin
`for growth. Lack of iron could alter the
`intracellular programs of gene control
`that vary iron absorption from the diet
`according to the demands of the body.
`Iron absorption in normal individuals
`depends on body iron stores: iron ab-
`sorption increases up to threefold dur-
`ing iron deficiency or pregnancy. This
`control appears to be perturbed in
`hemochromatosis.
`
`0968 – 0004/99/$ – See front matter © 1999, Elsevier Science. All rights reserved. PII: S0968-0004(99)01386-9
`
`PGR2020-00009
`Pharmacosmos A/S v. American Regent, Inc.
`Petitioner Ex. 1053 - Page 1
`
`

`

`TIBS 24 – MAY 1999
`
`of DCT1 by IRP has not been confirmed
`experimentally. Mutations in the DCT1
`gene of microcytic anemia (mk) mice
`and Belgrade (b) rats indicate that DCT1
`plays a role in transport of iron from the
`endosome to the cytoplasm in erythroid
`cells22,23. Whether, and if so how, DCT1
`is also responsible for intestinal iron
`absorption remains to be studied.
`This brings us back to the second un-
`solved question: how does mutated HFE
`cause activation of IRP in duodenal en-
`terocytes of hemochromatosis patients?
`I postulate that the mutation or loss of
`HFE diminishes TfR function in duode-
`nal crypt cells and impairs uptake of iron
`from serum transferrin (Fig. 1). When HFE
`does not associate with TfR, this could
`perturb a step in the TfR cycle: the bind-
`ing of transferrin to its receptor, the
`clustering of receptor–ligand complexes
`and formation of coated vesicles for
`endocytosis, the release of iron in endo-
`somes, receptor recycling, or release of
`apotransferrin at the cell surface (Fig. 1).
`Thus far, transfection studies carried
`out in normal cells show that HFE and
`TfR associate soon after synthesis and
`colocalize in the perinuclear region, on
`the cell surface and in endosomes9. In-
`terestingly, the interaction between TfR
`and HFE is weakened at pH 6 in vitro:
`this suggests that HFE dissociates in
`endosomes (whose pH is 6) and thereby
`facilitates release of iron from transfer-
`rin or recycling of TfR (Ref. 24). In addi-
`tion, wild-type HFE, when overexpressed
`in 293T or HeLa cells, lowers the affinity
`of surface TfR for transferrin9,11, whereas
`the Cys282fiTyr mutant of HFE has no
`effect11. This latter finding is unsurpris-
`ing: mutant HFE neither binds to TfR
`nor reaches the surface11. A diminished
`TfR–transferrin affinity in the presence
`of excess wild-type HFE is unexpected
`and might appear to contradict the
`hypothesis developed here. However,
`we cannot extrapolate from the overex-
`pression experiments to the situation in
`cells that lack HFE. Therefore, it remains
`to be seen whether the results of HFE-
`overexpression studies are relevant to
`hemochromatosis. We must measure all
`parameters of TfR function in cells that
`lack wild-type HFE or in cells that ex-
`press the HFE mutant, and compare the
`results with those obtained in normal
`cells. We might then identify a specific
`step in iron uptake that is perturbed
`when HFE does not assist TfR.
`In summary, we are still searching for
`the role of HFE in the regulation of duo-
`denal iron absorption. The discovery that
`it interacts with TfR suggests strongly
`
`FRONTLINES
`
`(a) HFE
`
`+/+ crypts
`
`(b) HFE
`+/+ villi
`
`(c) HFE —/— crypts
`
`(d) HFE —/— villi
`
`?
`
`?
`
`?
`
`Transferrin receptor
`
`Transferrin
`
`Apotransferrin
`
`Wild-type HFE
`
`Mutated HFE
`
`Fe3+ or Fe2+
`
`IRPs
`
`Apical iron (cid:0)
`transporter
`
`Basolateral iron (cid:0)
`transporter
`
`Endosome
`
`Gut
`
`Serum
`
`Figure 1
`A model for the regulation of intestinal iron absorption. (a) Crypts of normal (HFE1/1)
`individuals. (b) Villi of normal individuals. (c) Crypts of hemochromatosis patients (HFE2/2).
`(d) Villi of hemochromatosis patients. I postulate that, in crypt cells, HFE assists uptake of
`iron from serum by the transferrin receptor. The availability of iron from transferrin influ-
`ences iron-regulatory protein (IRP) activity and ferritin levels. Crypt cells differentiate to
`form villus cells and adjust the rate of iron absorption from the intestinal lumen with re-
`spect to the availability of serum iron that was sensed initially. When HFE is mutated or
`missing, an unknown step in transferrin-iron uptake is perturbed: crypt cells become iron
`deprived and, once differentiated, increase iron absorption accordingly.
`
`that HFE is required for the function of
`these receptors. Even if we assume that
`dysfunction of the HFE–TfR complex in
`intestinal cells perturbs the regulation of
`iron absorption, we still must determine
`whether the regulation occurs through
`IRPs or other iron-regulatory circuits. Fi-
`nally, we must identify the iron-transport
`
`proteins whose synthesis is controlled by
`this mechanism and might be regulated
`incorrectly in hemochromatosis.
`
`Acknowledgements
`I thank Claude Bron, Klaus Schümann
`and Eric Menotti for critical reading of
`and suggestions for the manuscript.
`
`165
`
`PGR2020-00009
`Pharmacosmos A/S v. American Regent, Inc.
`Petitioner Ex. 1053 - Page 2
`
`

`

`OBITUARY
`
`TIBS 24 – MAY 1999
`
`References
`1 Powell, J. L., Jazwinska, E. and Halliday, J. W.
`(1994) in Iron metabolism in Health and Disease
`(Brock, J. H. et al.,eds), pp. 227–270,
`W. B. Saunders
`2 Feder, J. N. et al. (1996) Nat. Genet.13, 399–408
`3 Waheed, A. et al. (1997) Proc. Natl. Acad. Sci.
`U. S. A.94, 12384–12389
`4 Feder, J. N. et al. (1997) J. Biol. Chem.272,
`14025–14028
`5 Santos, M. et al. (1996) J. Exp. Med.184,
`1975–1985
`6 Rothenberg, B. E. and Voland, J. R. (1996) Proc.
`Natl. Acad. Sci. U. S. A.93, 1529–1534
`7 Zhou, X. Y. et al. (1998) Proc. Natl. Acad. Sci.
`U. S. A.95, 2492–2497
`8 Parkkila, S. et al. (1997) Proc. Natl. Acad. Sci.
`U. S. A.94, 2534–2539
`9 Gross, C. N., Irrinki, A., Feder, J. N. and Enns, C. A.
`(1998) J. Biol. Chem.273, 22068–22074
`
`10 Parkkila, S. et al. (1997) Proc. Natl. Acad. Sci.
`U. S. A.94, 13198–13202
`11 Feder, J. N. et al. (1998) Proc. Natl. Acad. Sci.
`U. S. A.95, 1472–1477
`12 Kühn, L. C., Schulman, H. M. and Ponka, P.
`(1990) in Iron Transport and Storage (Ponka, P.,
`Schulman, H. M. and Woodworth, R. C., eds),
`pp. 149–191, CRC Press
`13 Morgan, E. H. (1983) Biochim. Biophys. Acta
`762, 498–502
`14 Dautry-Varsat, A., Ciechanover, A. and Lodish,
`H. F. (1983) Proc. Natl. Acad. Sci. U. S. A.80,
`2258–2262
`15 Banerjee, D., Flanagan, P. R., Cluett, J. and
`Valberg, L. S. (1986) Gastroenterology91,
`861–869
`16 Anderson, G. J., Powell, L. W. and
`Halliday, J. W. (1990) Gastroenterology98,
`576–585
`17 Klausner, R. D., Rouault, T. A. and Harford, J. B.
`
`(1993) Cell72, 19–28
`18 Hentze, M. W. and Kühn, L. C. (1996) Proc.
`Natl. Acad. Sci. U. S. A.93, 8175–8182
`19 Pietrangelo, A. et al. (1995) Gastroenterology
`108, 208–217
`20 Cairo, G. et al. (1997) Blood89, 2546–2553
`21 Gunshin, H. et al. (1997) Nature388, 482–488
`22 Fleming, M. D. et al. (1997) Nat. Genet.16,
`383–386
`23 Fleming, M. D. et al. (1998) Proc. Natl. Acad.
`Sci. U. S. A.95, 1148–1153
`24 Lebron, J. A. et al. (1998) Cell93, 111–123
`
`LUKAS C. KÜHN
`Swiss Institute for Experimental Cancer
`Research, CH-1066 Epalinges s/Lausanne,
`Switzerland.
`Email: lukas.kuehn@isrec.unil.ch
`
`Lars Ernster (1920–1998)
`
`(SMPs) are capable of coupling aerobic
`oxidation of succinate to energy-linked
`reduction of NADP1 by NADH. These
`experiments were proof that energy de-
`rived from respiration can be conserved
`and used without the intervention of the
`high-energy phosphate
`intermediates
`first proposed by Slater. This research
`was published in 1963 – two years after
`the formulation of the chemiosmotic
`hypothesis. However, it was more than
`ten years later that Lars, together with
`Boyer, Mitchell, Racker and Slater1, pub-
`lished a consensus review of the
`chemiosmotic hypothesis
`that ex-
`plained the essential features of the
`mechanism of energy transduction.
`Lars made several other pioneering
`scientific advances that had great im-
`pact on the field2,3 and deserve mention.
`He and his collaboractors introduced
`rotenone as a specific electron-transfer
`inhibitor for NADH oxidase; rotenone later
`became the prototype for respiratory-
`chain inhibitors that lack side effects.
`They isolated and purified DT diaphorase,
`an enzyme that can use NADH or NADPH
`as a substrate. This led to the proposal
`of a pathway for the oxidation of nicotina-
`mide adenine nucleotides outside the
`mitochondrial inner membrane. In addi-
`tion, Lars and his colleagues introduced
`the use of well-characterized SMPs,
`which retain the ability to carry out oxi-
`dative phosphorylation but are free from
`the barriers to substrates that cause dif-
`ficulties in many types of experiments
`on intact mitochondria. They proposed,
`in 1965, that the inner mitochondrial
`membrane is turned inside out during
`preparation of SMPs – that is, the in-
`accessible inside of the membrane be-
`comes exposed to the medium. This
`provided a unique system for studying
`the topology of the mitochondrial inner
`
`Lars Ernster (Fig. 1) was born in Budapest
`in 1920 but left his mother country in 1946
`for Stockholm. In 1956, he received his
`PhD from the University of Stockholm,
`where he spent his entire scientific career
`and authored ~ 600 publications. Initially,
`he was a research scientist at the Wenner
`Gren Institute for Experimental Biology
`(where he remained from 1946 until 1967);
`he then became Head of Physiological
`Chemistry. In 1967, he was appointed
`Professor and Chairman of the Depart-
`ment of Biochemistry and, in 1986, he
`became Professor Emeritus. Lars never
`retired, continuing to pursue his research
`to the last day of his life, as active as ever.
`He died at his home on 4 November 1998.
`Lars’s main area of interest was the
`physiology and biochemistry of mito-
`chondria, although his contributions to
`the field were diverse. During his early
`scientific career, he collaborated with
`his mentor, Olle Lindberg, at the Wenner
`Gren Institute. They studied mitochon-
`drial physiology, focusing on the effects
`of calcium uptake on and the protective
`role of manganese in mitochondrial func-
`tion. They also reported the inhibitory
`effect of amytal on site I electron trans-
`fer and, in 1954, published a major re-
`view that substantially advanced our
`knowledge of this organelle.
`By the end of the 1950s, hypotheses
`of oxidative phosphorylation had come
`under scrutiny. A number of young for-
`eign scientists came to Stockholm and
`worked under Lars’s guidance. Stockholm
`became one of the major international
`centres for research into mitochondria.
`166
`
`the discovery of oxidative
`Since
`phosphorylation, it had been generally
`accepted that phosphate was involved
`in the initial reactions of energy trans-
`duction. Lars and his collaborators,
`however, discovered that oligomycin
`does not inhibit energy-linked reduction
`of NAD1 by succinate when energy is
`supplied through aerobic oxidation of
`succinate. Furthermore, non-phosphory-
`lating
`submitochondrial
`particles
`
`Figure 1
`
`Lars Ernster.
`
`0968 – 0004/99/$ – See front matter © 1999, Elsevier Science. All rights reserved. PII: S0968-0004(99)01389-4
`
`PGR2020-00009
`Pharmacosmos A/S v. American Regent, Inc.
`Petitioner Ex. 1053 - Page 3
`
`