INVITED COMMENTARY Labeled Glucose Analogs in the Genomic Era

The leftmost structure is a conven- of down). These anomers of D-glucose In a preclinical research article in tional open-chain representation of the can be separated from one another by this issue of The Journal of Nuclear D-glucose molecule, showing the num- crystallization, but their fairly rapid in- Medicine, Bormans et al. (1) examine bering system of the 6 carbon atoms terconversion in biologic fluids means the uptake of 11C-labeled D-glucose an- and the stereochemistry of the hydro- that there are no important conse- alogs ␣-methyl-D-glucoside and ␤-meth- gen atoms and hydroxyl groups on quences of this phenomenon for D-glu- yl-D-glucoside in the mouse kidney. each atom. The aldehyde function is cose itself, even when enzymes and It appears that both of these com- shown at the top and designated as transporters recognize only 1 of the pounds are taken up efficiently from carbon atom 1 (C1). C6 carries a pri- anomers. However, the ␣-D-glucopyr- the glomerular filtrate by tubular ep- mary hydroxyl group, and the other 4 anose and ␤-D-glucopyranose ring ithelial cells. However, ␤-methyl-D- carbon atoms (C2ÐC5) each carry a structures can be “locked in” by for- glucoside is present in kidney at 30 secondary hydroxyl group. Because of mation of glucosides. Thus, when glu- min to a much greater extent than the chirality of the central carbon at- cose is reacted with an excess of meth- 2 ␣-methyl-D-glucoside, suggesting that oms, D-glucose is 1 of 16 (4 ) “aldo- anol under acidic conditions, a mix- glucose transport from the tubular ep- hexose” isomers with this general ture of methyl-␣-D-glucopyranoside ithelial cells to the blood is slower for structure. One of these, L-glucose, is (␣-methylglucoside) and ␤-methylglu- ␤-methyl-D-glucoside and that this the mirror image of D-glucose. Another coside is produced. Bormans et al. (1) constitutes a trapping mechanism that conventional representation of D-glu- used the very powerful radiomethylat- would allow PET imaging of the first cose is shown in the second structure, ing agent, 11C-methyl triflate, in tracer transport step of glucose reabsorption where the hydrogen atoms on C1ÐC5 quantities to produce methylglucosides in kidney. The article by Bormans et are not shown. In solution, D-glucose with high specific activity suitable for al. prompts discussion of several is- exists as an equilibrium between low nuclear medical use (third and fourth sues. To understand the significance of concentrations of the open-chain form structures from the left). These 2 com- this work it is helpful to review the and cyclized forms containing a pounds were then separated by chiral basic chemistry, biochemistry, and 6-member (“pyranose”) ring. (Forms high-performance liquid chromatogra- physiology of D-glucose. The struc- containing a 5-member ring are also phy so that their biologic behavior tures of glucose and of the labeled minor components of the equilibrium could be examined separately. compounds studied by Bormans et al. mixture but are not shown here.) For- are shown below: mation of the 6-member ring creates a HISTORICAL BACKGROUND It is almost 30 y since chemists at the Brookhaven National Laboratory first synthesized 2-deoxy-2-18F-fluoro- D-glucose (18F-FDG), which today sus- tains the field of clinical PET (2). FDG, like glucose, is transported into the brain and other tissues by facilita- tive (nonenergy linked) carrier mole- cules. Furthermore, FDG, like glucose, is a substrate for hexokinase isoforms and is converted to FDG-6-phosphate. However, FDG-6-phosphate is unlike glucose-6-phosphate in that, for funda- mental structural reasons, it cannot un- new asymmetric center at C1 and, dergo the next step in metabolism of Received Mar. 25, 2003; accepted Mar. 26, therefore, 2 distinct isomers (histori- glucose itself, which is conversion to 2003. cally termed “anomers”) referred to as For correspondence or reprints contact: S. John fructose-6-phosphate. Another prop- Gatley, PhD, Medical Department, Brookhaven ␣-D-glucopyranose (shown here) and erty that distinguishes FDG from glu- National Laboratory, Building 490, Upton, NY ␤-D-glucopyranose (where the hy- 11973. cose is that FDG is only poorly trans- E-mail: [email protected] droxyl group on C1 points up instead ported by the energy-linked carriers for

1082 THE JOURNAL OF NUCLEAR MEDICINE ¥ Vol. 44 ¥ No. 7 ¥ July 2003 glucose that are found, for example, in that a multiplicity of transporters groups (e.g., iodoalkyl, iodovinyl, or the small intestine, which harvests di- would be required for precise control iodophenyl moieties), which has been etary glucose, and in the kidney, which of transport of glucose into and out of successfully used for receptor-binding recovers the glucose present in the glo- many different types of cells. radiopharmaceuticals, because the en- merular filtrate. Thus, FDG is excreted zyme cannot tolerate bulky substitu- in the urine, a circumstance that aids GLUCOSE ANALOGS THAT ents at C2. It is also likely that the the clearance of untrapped FDG from ARE TRANSPORTED AND facilitative glucose transport process tissues and so may improve its radio- PHOSPHORYLATED cannot handle large groups at C2, be- pharmaceutical characteristics (3). The behavior of radiolabeled glu- cause 2-deoxy-2-bromo-D-glucose ex- Except in extreme circumstances, such cose analogs toward hexokinase and hibits little uptake into brain (12,15). as deep hypoglycemia or the near-ab- the prospects for development of alter- sence of transporters, FDG studies are natives to FDG, particularly single LABELED GLUCOSE ANALOGS not informative about glucose trans- photon-emitting sugars, were dis- THAT ARE TRANSPORTED BUT port because the phosphorylation step cussed in an earlier review (5) and are NOT PHOSPHORYLATED is normally rate determining for tissue briefly reiterated and updated here. FDG and the other 2-substituted hex- glucose consumption. Development of Early workers (6) recognized that glu- oses that are good substrates for hexoki- labeled compounds that are specific cose analogs substituted at C1, C5, or nase are also good substrates for facilita- markers for glucose transport, there- C6 with halogens are either unstable or tive transport. Another positron-labeled fore, represents an opportunity in ra- are unreactive with hexokinase for fun- sugar, 3-O-11C-methyl-D-glucose, is diopharmaceutical research. damental structural reasons, whereas transported by both facilitative and FDG was developed before the ad- glucose substituted at C4 with a fluo- concentrative transporters but is phos- vent of modern molecular biologic rine atom is a very poor substrate for phorylated only very slowly (16). It techniques, which allow us, for exam- yeast hexokinase (7). Of the 8 D-aldo- has thus been used to estimate regional ple, to recognize families of biologi- hexoses, only D-glucose and its C2 glucose distribution volumes in brain, cally active proteins in DNA sequence epimer D-mannose are good substrates and its slower plasma clearance kinet- data, extract messenger RNA with par- for phosphorylation by hexokinases. ics due to tubular reabsorption is an tial complementarity to known genes, 2-Deoxy-D-glucose, 2-deoxy-2-fluoro- advantage for this purpose. Tritium- generate pure clones of genes coding D-mannose, and 2,2-difluoro-deoxy-D- labeled 3-O-methyl-D-glucose is com- for interesting proteins, and make mu- glucose are also good substrates and, monly used for similar purposes in lab- tant and chimeric proteins, cells, and in fact, are suitable in positron-labeled oratory experiments. Another glucose animals for mechanistic studies. Be- forms for measurement of local glu- derivative substituted at C3, 3-deoxy- fore the genomic era, structureÐactiv- cose metabolic rates (8–10). Radioio- 3-fluoro-D-glucose, is also well trans- ity relationships for glucose analogs dinated 2-deoxy-2-iodo-D-glucose is ported (17) but is poorly phosphorylated for accumulation and metabolism in chemically unstable (11,12) and, there- (8). However, 3-deoxy-3-18F-fluoro-D- various cells and tissues had been de- fore, unsuitable for radiopharmaceuti- glucose appears to be less suitable for termined, and this information together cal use. McCarter et al. (13) improved PET experiments than 3-O-11C-meth- with physiologic data from tissue prep- the stability of 2-iodoÐsubstituted glu- yl-D-glucose because several meta- arations provided evidence for multi- cose and mannose by replacing the hy- bolic products are also formed from ple transport systems for glucose. For drogen on C2 by a fluorine atom. How- 3-deoxy-3-fluoro-D-glucose, at least in example, facilitative glucose uptake in ever, neither 2-deoxy-2-fluoro-2-iodo-D- heart (18). The structural requirements muscle is sensitive to insulin, whereas glucose nor 2-deoxy-2-fluoro-2-iodo-D- for interaction of carbohydrates with glucose uptake in brain or erythrocytes mannose possessed detectable activity facilitative carriers that were estab- has little or no sensitivity. Clinical ob- with hexokinase from yeast or bovine lished many years ago on the basis of servations supported the involvement heart, and 2-deoxy-2-123I-iodo-D-man- studies in erythrocytes (19,20) suggest of at least 2 gene products in energy- nose was found, in mouse biodistribu- that 3-deoxy-3-iodo-D-glucose or any linked glucose transport, because dis- tion experiments, to lack the properties analog with a larger group at C3 will tinct syndromes for familial glucose of a marker for glucose metabolism (10). not be well transported. This expecta- malabsorption and renal glycosuria Glucose derivatives modified at C2 tion is supported by recent studies had been recognized (4). The existence with bulky substituents—for example, (21). of several transport systems for glu- 2-O-methyl-D-glucose—are very poor The C6 position is also tolerant in cose was not surprising in view of the substrates for hexokinase, and 2-deoxy- terms of facilitative transport of glu- central importance of glucose in met- 2-chloro-D-glucose is only a moderately cose analogs, and 6-deoxy-D-glucose, abolic pathways. The need to maintain good substrate (6,14). Thus, it does not 6-deoxy-6-fluoro-D-glucose, and 6-de- plasma glucose levels within tight lim- appear possible to adopt the strategy oxy-6-iodo-D-glucose all appear to be its and the tissue-specific requirements of preparing glucose derivatives sub- transported. The data of Koumanov et for handling glucose logically meant stituted with other iodine-containing al. (22) are consistent with facilitated

LABELED GLUCOSE ANALOGS IN THE GENOMIC ERA ¥ Gatley 1083 transport of 6-deoxy-6-123I-iodo-D-glu- use of the more energetically expen- The complete picture of sugar trans- cose into the heart and brain. Although sive SGLT1 for the bulk of the glucose port in some tissues, perhaps most tis- these authors concluded that this radio- transport. There is evidence for at least sues, can be complex. Thus, recent im- iodinated sugar was promising for clin- one more member of the SGLT family munohistochemical studies in rodents ical studies, its uptake in brain and in kidney (25). have demonstrated the presence of at heart was lower than the uptake of The other family of glucose trans- least 6 members of the GLUT family labeled 3-O-methyl-D-glucose. porters is composed of facilitative in the brain. GLUT1 is found in endo- transporters whose prototype is termed thelial cells and controls entry of glu- CURRENT UNDERSTANDING OF GLUT1 and which is capable only of cose into the brain, whereas GLUT3 GLUCOSE TRANSPORT carrying glucose down a gradient. The and GLUT4 are expressed in a region- Since the first report of the cloning GLUT molecules have amino acid se- ally specific manner in neurons (29). of a mammalian glucose transporter in quences consistent with 12 transmem- Glia contain GLUT1, but with a dis- 1985 (23), molecular biologic tech- brane domains. GLUT1 is responsible tinctive pattern of posttranslational niques have helped to confirm the ex- for glucose transport into erythrocytes glycosylation (30). Poppe et al. (31) istence of 2 very different families of and through the bloodÐbrain barrier reported messenger RNA for SGLT1 mammalian glucose transporters, known and is upregulated in many tumors. A in brain. To make things even more as SGLTs and GLUTs (24). One family syndrome resulting from GLUT1 defi- complicated, a very recent article has comprises molecules that are sodium ciency has been recognized (26) and suggested that, in addition to the facil- glucose cotransporters (SGLTs) that can involve spasticity, developmental itative GLUT isoforms and concentra- are believed to have 14 transmembrane delay, and microcephaly. A recent ar- tive Naϩ-linked glucose transporters, 18 domains. These couple the movement ticle (27) describes the use of F-FDG brain contains a novel concentrative of sodium down a gradient with the PET in diagnosis of this syndrome. Hϩ-linked glucose transporter (32). pumping of glucose up a gradient. The GLUT2 is found in tissues that include As mentioned above, the reabsorp- high Naϩ outside gradient used to the liver and pancreatic ␤-cells. Defi- tion of glucose is completed by its drive concentrative glucose transport is ciency of GLUT2 leads to a form of passage through the basolateral mem- maintained by the sodium, potassium glycogen storage disease known as branes of tubule cells via the GLUT2 adenosine triphosphatase in the basolat- FanconiÐBickel syndrome (28) that in- carrier. It appears that methyl-D-glu- eral membrane. The prototype, SGLT1, volves hepatic glycogen accumulation cosides accumulated by epithelial cells is responsible for transport of glucose and a distinctive tubular nephropathy. can also be reabsorbed into the blood, out of the small intestine across the These disorders reflect roles of GLUT2 because the fraction of the injected 11C brush border membrane of intestinal in transporting newly synthesized glu- appearing in urine is Ͻ5% of that ap- epithelial cells, but is thought to nor- cose out of hepatocytes and in trans- mally play a relatively minor role in porting glucose through the basolateral pearing in urine after administration of reabsorption of glucose in the kidney. membranes of tubular and intestinal methyl-L-glucosides (Table 2 in Bor- SGLT1 moves 2 sodium ions into ep- epithelial cells. Its affinity for glucose mans et al. (1)). Thus, the methyl-D- ithelial cells per glucose molecule and is lower than that of GLUT1, but it has glucosides seem to be substrates for so can pump glucose against a high a broader substrate specificity and also GLUT2. However, there is some con- concentration gradient. A related trans- transports other sugars such as fructose fusion on this point in the literature. It porter, SGLT2, is found in the first (S1 and glucosamine. GLUT3 has a high is generally accepted that methyl-D- and S2) part of the proximal kidney tu- affinity for glucose and is the major glucosides are not substrates for bule, where it pumps glucose from blood form found in neurons. GLUT4 is the GLUT1 (33), and the small degree of filtrate through the apical membrane of insulin-sensitive transporter found in uptake by brain and erythrocytes re- epithelial cells. SGLT2 is believed to muscle tissue. GLUT1ÐGLUT4 are ported by Bormans et al. supports this exhibit a Naϩ-to-glucose ratio of 1:1 collectively termed class I facilitative view. It is possible that species differ- and pumps glucose from a relatively glucose transporters. At least 8 other ences in GLUTs or tissue-specific high glucose concentration—initially, members of the GLUT family have posttranslational modification (34) the plasma concentration of about 5 been identified (24). They include could account in part for this confu- mmol/L. It is believed to be responsible, GLUT5, which is responsible for sion. Because ␤-11C-methylglucoside in euglycemic individuals, for the bulk transport of fructose at the apical mem- was retained by kidney to a much of glucose reabsorption. Further along brane of small intestinal epithelial greater extent than ␣-11C-methylglu- the tubule, in the S3 domain, SGLT1 cells. Thus, the absorption of dietary coside (kidney-to-blood ratio of about further reduces the glucose concentra- fructose relies entirely on facilitative 8 compared with about 2 at 30 min, tion to the near-zero level found in the transport, unlike the situation for glu- and radioactivity remaining in the kid- fully formed urine. This picture of se- cose, but both sugars use different neys of 31% and 8% injected dose, quential operation of SGLT2 and transporters at the apical and basolat- respectively), the suggestion of Bor- SGLT1 is logical because it avoids the eral poles of the epithelial cells. mans et al. that ␤-11C-methylglucoside

1084 THE JOURNAL OF NUCLEAR MEDICINE ¥ Vol. 44 ¥ No. 7 ¥ July 2003 is a much poorer substrate for GLUT2 a transported analog, especially if be less competition with glucose and than the ␣-glucoside is compelling. GLUT1 selective, might be a useful the possibility of more stable uptake. A The demonstration by imaging plate tumor marker. Issues that would have disadvantage is that a radioligand autoradiography that, after intravenous to be resolved include competition be- would be sensitive to the concentration administration of a mixture of ␣- and tween radiotracer and glucose for up- of transporter, rather than flux through ␤-11C-methyl-D-glucosides, 11Cislo- take in target tissue and uptake into the transporter, as with a transported calized in the kidney cortex and outer erythrocytes that would lower tissue- analog, and might be insensitive to fac- medulla is consistent with retention in to-background ratios, at least for extra- tors that regulate transporter flux. Nev- tubular cells. Judging by the tissue dis- cerebral tissues. A general problem ertheless, subtype-specific high-affin- tribution data, this represents almost with radiotracers that are substrates for ity radioligands capable of binding in exclusively ␤-11C-methyl-D-glucoside, facilitative transporters is the absence vivo to glucose carriers might have though it would be useful if this were of any trapping mechanism. According wide applicability. confirmed by autoradiography after ad- to Bormans et al. (1), trapping of ministration of pure ␤-11C-methyl-D- ␤-methyl-D-glucoside in the kidney is CONCLUSION glucoside rather than a mixture of ano- provided by the fact that there is a mers. From a technical point of view, greater inward energy-linked flux into Future studies with ␣- and ␤-meth- this is an interesting and challenging tubule cells than the outward facilita- yl-D-glucosides should include experi- example of ex vivo autoradiography tive flux. ments with cells transfected with using 11C. Similar experiments do not Despite this problem, some work GLUT2 to seek support for the hypoth- appear to have been conducted using has continued in this area. Brunet- esis that the ␤-compound accumulates longer-lived isotopes, though the auto- Desruet et al. (40) synthesized 2 ra- in kidney because it is a poor substrate radiographic measurement of uptake of dioiodinated acetal derivatives of for efflux via this GLUT isoform. ␣-14C-methyl-D-glucoside in explants glucose, but these were not trans- Also, if it is true that most of the glu- of fetal and newborn mouse kidney ported by either GLUT1 or GLUT4. cose in glomerular filtrate is reab- was recently reported (35). Xu et al. (41) have developed glu- sorbed through SGLT2, then experi- Henry et al. (36) showed that methyl cose analogs related to the antibiotic ments should be conducted in oocytes 6-deoxy-6-iodo-D-glucoside is trans- nojirimycin that contain a 6-member expressing this carrier rather than ported into rat brain, erythrocytes, and ring with a nitrogen atom instead of SGLT1, although it is recognized that neonatal cardiomyocytes (though to a an oxygen atom. These compounds this is technically more difficult (45). lesser extent than the 3-O-methylglu- bind to enzymes involved in glyco- In addition, although Bormans et al. cose), implying that this iodine-substi- sylation reactions and are potential (1) report the biodistribution of 11C- tuted methyl glucoside analog is a sub- tumor-imaging agents but were not methylglucosides in the mouse, their strate for both GLUT1 and GLUT4. transported into brain. studies of the transport through SGLT1 Although it would seem unlikely, it were done with the human cloned appears that the change in structure at HIGH-AFFINITY transporter. Though there is significant C6 compensates for the change in NONTRANSPORTED homology between mouse and human structure at C1, in terms of ability to be RADIOLIGANDS FOR genes (35), it is at least possible that ␣- GLUCOSE TRANSPORTERS transported by GLUT isoforms. It is and ␤-methyl-D-glucosides exhibit dif- interesting that these workers did not It was suggested over a decade ago ferent relative kinetics with mouse and find a similar difference in kidney up- that high-affinity ligands could be de- human SGLT1. Another question is take between ␣- and ␤-methyl-6-de- veloped for GLUT isoforms (42,43). the possible role of facilitative trans- ϩ oxy-6-iodo-D-glucoside. Experience with neurotransmitter re- porters other than GLUT2 and Na - ceptor and transporter radioligands linked transporters other than SGLT1 PROSPECTS FOR NEW suggests that dissociation constants in and SGLT2. Several secondary ques- RADIOPHARMACEUTICALS FOR the nanomolar range would be re- tions should also be addressed, such as IMAGING SUGAR TRANSPORT quired. Several compounds are known whether alteration of blood glucose PET and SPECT markers for the to interact with glucose transporters concentrations would affect the trans- process of glucose transport might be with much higher affinities than glu- port of ␤-methyl-D-glucosides from useful in early diagnosis of neurologic cose, including phloretin, phlorizin, the tubular cells. Nevertheless, Bor- diseases where transport is compro- forskolin analogs, and cytochalasin B mans et al. should be congratulated for mised, such as Alzheimer’s disease (20,42). These compounds and others, an imaginative study using a readily (37) or Huntington’s disease (38). Ad- such as 2,3-dioxoindole (44), might be prepared positron-emitting tracer that ditionally, because GLUT1 is well ex- leading compounds for development of raises new questions and has potential pressed by many tumors in tissues a more potent inhibitor. Potential ad- clinical applicability. Their article il- where a lower expression of a different vantages of this approach over using lustrates the use of sophisticated tech- isoform is normally predominant (39), reversibly transported substrates would niques, including chiral separation,

LABELED GLUCOSE ANALOGS IN THE GENOMIC ERA ¥ Gatley 1085 autoradiography with 11C, and electro- of regional metabolic measurement in vivo via 28. Santer R, Steinmann B, Schaub J. Fanconi-Bickel metabolic trapping: 11C-2-deoxy-D-glucose and ha- syndrome: a congenital defect of facilitative glu- physiologic measurements of a specific logenated deoxyglucose derivatives. J Labelled cose transport. Curr Mol Med. 2002;2:213Ð227. transporter artificially expressed in oo- Compds Radiopharm. 1979;16:7Ð9. 29. Choeiri C, Staines W, Messier C. Immunohisto- cytes, to explain tissue distribution 12. Kloster G, Laufer P, Wutz W, Stocklin G. 75,77Br- chemical localization and quantification of glucose and 123I-analogues of D-glucose as potential tracers transporters in the mouse brain. Neuroscience. data. for glucose utilisation in heart and brain. Eur 2002;111:19Ð34. J Nucl Med. 1983;8:237Ð241. 30. Vannucci SJ, Maher F, Simpson IA. Glucose trans- ACKNOWLEDGMENT 13. McCarter JD, Adam MJ, Withers SG. Syntheses, porter proteins in brain: delivery of glucose to radiolabeling, and kinetic evaluation of 2-deoxy-2- neurons and glia. Glia. 1997;21:2Ð21. This work was conducted at the fluoro-2-iodo-D-hexoses for medical imaging. Car- 31. Poppe R, Karbach U, Gambaryan S, et al. Expres- Brookhaven National Laboratory un- bohydr Res. 1995;266:273Ð277. sion of the Naϩ-D-glucose cotransporter SGLT1 in 14. Sols A, DelaFuente G, Villar-Palasi C, Asenio C. neurons. J Neurochem. 1997;69:84Ð94. der contract DE-AC02-98CH10886 Substrate specificity and some other properties of 32. Shimokawa N, Okada J, Haglund K, Dikic I, Koi- with the U.S. Department of Energy baker’s yeast hexokinase. Biochim Biophys Acta. buchi N, Miura M. Past-A, a novel proton-associ- and supported by its Office of Health 1958;30:92Ð101. ated sugar transporter, regulates glucose homeosta- 15. Zhou YG, Shiue CY, Wolf AP. Syntheses of ra- sis in the brain. J Neurosci. 2002;22:9160Ð9165. and Environmental Research. diobrominated 2-deoxy-2-bromo-D-glucose and 33. Burant CF, Sivitz WI, Fukumoto H, et al. Mam- 2-deoxy-2-bromo-D-mannose for the measurement malian glucose transporters: structure and molecu- S. John Gatley, PhD of cerebral glucose metabolism in vivo. J Labelled lar regulation. Recent Prog Horm Res. 1991;47: Compds Radiopharm. 1982;19:1623Ð1625. 349Ð387. Brookhaven National Laboratory 16. Gatley SJ, Holden JE, Halama JR, DeGrado TR, 34. Ahmed N, Berridge MV. N-Glycosylation of glu- Upton, New York Bernstein DR, Ng CK. Phosphorylation of glucose cose transporter-1 (Glut-1) is associated with in- analog 3-O-methyl-D-glucose by rat heart. Bio- creased transporter affinity for glucose in human REFERENCES chem Biophys Res Commun. 1984;119:1008Ð1014. leukemic cells. Leuk Res. 1999;23:395Ð401. 17. Taylor NF, Romaschin A, Smith D. Metabolic and 35. Yang Q, Tian Y, Wada J, et al. Expression char- 1. Bormans GM, Van Oosterwyck G, de Groot TJ, et transport studies with deoxyfluoro-monosacchar- acteristics and relevance of sodium glucose co- al. Synthesis and biologic evaluation of 11C-meth- ides. In: Filler R, ed. Biochemistry Involving Car- transporter-1 in mammalian renal tubulogenesis. yl-D-glucoside, a tracer of the sodium-dependent bon-Fluorine Bonds. Washington, DC: American Am J Physiol Renal Physiol. 2000;279:F765ÐF777. glucose transporters. J Nucl Med. 2003;44:1075- Chemical Society; 1976:99Ð116. 36. Henry C, Koumanov F, Ghezzi C, et al. [123I]-6- 1081. 18. Halama JR, Gatley SJ, DeGrado TR, Bernstein DR, Deoxy-6-iodo-D-glucose (6DIG): a potential tracer 2. Gallagher BM, Ansari A, Atkins H, et al. Radio- Ng CK, Holden JE. Validation of 3-deoxy-3- of glucose transport. Nucl Med Biol. 1997;24:527Ð pharmaceuticals. XXVII. 18F-Labeled 2-deoxy-2- fluoro-D-glucose as a well transported, poorly 534. fluoro-D-glucose as a radiopharmaceutical for mea- phosphorylated analog of glucose in rat heart and 37. Simpson IA, Chundu KR, Davies-Hill TM, Honer suring regional myocardial glucose metabolism in brain. Am J Physiol. 1984;247:H756ÐH759. WG, Davies P. Decreased concentrations of vivo: tissue distribution and imaging studies in 19. Barnett JEG, Holman GD, Munday KA. Structural GLUT1 and GLUT3 glucose transporters in the animals. J Nucl Med. 1977;10:990Ð996. requirements for binding to the sugar-transport sys- brain of patients with Alzheimer’s disease. Ann 3. Moran JK, Lee HB, Blaufox MD. Optimization of tem of the human erythrocytes. Biochem J. 1973; Neurol. 1994;35:546Ð551. urinary FDG excretion during PET imaging. J Nucl 131:211Ð221. 38. Gamberino WC, Brennan WA. Glucose transporter Med. 1999;40:1352Ð1357. 20. Betz AL, Drewes LR, Gilboe DD. Inhibition of isoform expression in Huntington’s disease. J Neu- 4. Elsas LJ, Rosenberg LE. Familial renal glycosuria: glucose transport into brain by phlorizin, phloretin rochem. 1994;63:1392Ð1397. a genetic reappraisal of hexose transport by kidney and glucose analogues. Biochim Biophys Acta. 39. Smith TA. Facilitative glucose transporter expres- and intestine. J Clin Invest. 1969;48:1845Ð1854. 1975;406:505Ð515. sion in human cancer tissue. Br J Biomed Sci. 5. Gatley SJ. Iodine-123-labeled glucose analogs: 21. Koumanov F, Henry C, Ghezzi C, et al. Biological 1999;56:285Ð292. prospects for a single-photon-emitting analog of studies of analogues of glucose iodinated in posi- 40. Brunet-Desruet MD, Ghezzi C, Morin C, Comet M, fluorine-18-labeled deoxyglucose. Nucl Med Biol. tions 1, 2, or 3. Nucl Med Biol. 1996;23:53Ð60. Fagret D. Biological evaluation of two iodine-123- 1995;22:829Ð835. 22. Koumanov F, Henry C, Ghezzi C, et al. Biological labeled D-glucose acetals prepared as glucose 6. Sols A, Crane RK. Substrate specificity of brain evaluation of two anomeric glucose analogues io- transporter radioligands. Nucl Med Biol. 1998;25: hexokinase. J Biol Chem. 1954;210:581Ð595. dinated in position 6. Nucl Med Biol. 1997;24:519Ð 473Ð480. 7. Bessell EM, Foster AB, Westwood JH. The use of 525. 41. Xu Y, Choi SR, Kung MP, Kung HF. Synthesis of deoxyfluoro-D-glucopyranoses and related com- 23. Mueckler M, Caruso C, Baldwin SA, et al. Se- radioiodinated 1-deoxy-nojirimycin derivatives: pounds in a study of yeast hexokinase specificity. quence and structure of a human glucose trans- novel glucose analogs. Nucl Med Biol. 1999;26: Biochem J. 1972;126:199Ð204. porter. Science. 1985;229:941Ð945. 833Ð839. 8. Bessell EM, Thomas P. The deoxyfluoro-D-gluco- 24. Wood IS, Trayhurn P. Glucose transporters (GLUT 42. Robbins JD, Appel NM, Laurenza A, Simpson IA, pyranose 6-phosphates and their effect on yeast and SGLT): expanded families of sugar transport DeSouza EB, Seamon KB. Differential identifica- glucose phosphate isomerase. Biochem J. 1972; proteins. Br J Nutr. 2003;89:3Ð9. tion and localization of adenylyl cyclase and glu- 130:20PÐ21P. 25. Diez-Sampedro A, Eskandari S, Wright EM, cose transporter in brain using iodinated derivatives 9. Braun AR, Carson RE, Adams HR, Finn RD, Fran- Hirayama BA. Naϩ-to-sugar stoichiometry of of forskolin. Brain Res. 1992;581:148Ð152. cis BE, Herscovitch P. A kinetic comparison of SGLT3. Am J Physiol Renal Physiol. 2001;282: 43. Mertens J, Terriere D, Boumon R. The radiosyn- [18F]2-fluoro-2-deoxyglucose and [18F]2-fluoro-2- F278ÐF282. thesis of a new potential glucose-transport protein deoxymannose using positron emission tomogra- 26. Klepper J, Voit T. Facilitated glucose transporter tracer: 3-[I-125]iodophloretin. J Labelled Compds phy. Nucl Med Biol. 1994;21:857Ð863. protein type 1 (GLUT1) deficiency syndrome: im- Radiopharm. 1994;34:1011Ð1021. 10. Matte G, Adam M, Lyster D. Biological evaluation paired glucose transport into brain—a review. Eur 44. Mousavi-Gargari ML, Bansal RC, Singh K, Mah- of 2-fluoro-2-[123I]iodo-mannose (FIM): biological J Pediatr. 2002;161:295Ð304. mood A. Inhibition of glucose transport in human evaluation of FIM. Nucl Med Biol. 2001;28:679Ð 27. Pascual JM, Van Heertum RL, Wang D, Engelstad erythrocytes by 2,3-dioxoindole (Isatin). Experien- 682. K, De Vivo DC. Imaging the metabolic footprint of tia. 1994;50:833Ð836. 11. Fowler JS, Lade RE, MacGregor RR, Shiue C, Glut1 deficiency on the brain. Ann Neurol. 2002; 45. Wright EM. Renal Naϩ-glucose cotransporters. Wan C-N, Wolf AP. Agents for the armamentarium 52:458Ð464. Am J Physiol Renal Physiol. 2001;280:F10ÐF18.

1086 THE JOURNAL OF NUCLEAR MEDICINE ¥ Vol. 44 ¥ No. 7 ¥ July 2003