Use of Phosphorous-31 Nuclear Magnetic Resonance Spectroscopy to DetermineSafe Timing of Chemotherapy after Hepatic Resection1
David A. Kooby, Kristen L. Zakian, Surya N. Challa, Cornelia Matei, Henrik Petrowsky, Hyok-Hee Yoo,Jason A. Koutcher, and Yuman Fong2
Departments of Surgery [D. A. K., S. N. C., H. P., Y. F.] and Radiology [K. L. Z., C. M., H- H. Y., J. A. K.], Memorial Sloan-Kettering Cancer Center, New York, New York 10021
ABSTRACT
Liver resection induces accelerated growth of residual hepatic micro-metastases. Adjuvant chemotherapy may improve outcome if adminis-tered early after resection but may prove lethal if initiated prior tocompletion of DNA synthesis in regenerating liver. This study investigatesphosphorus-31 nuclear magnetic resonance (31P-NMR) as a noninvasivetool for measuring energy changes reflective of hepatic DNA synthesis andfor predicting safe timing of chemotherapy after 70% hepatectomy. Toevaluate metabolic changes in regenerating liver, quantitative three-dimensional 31P-NMR was performed, using the technique of chemicalshift imaging at various time points after 70% hepatectomy in adult maleFischer rats. Animals receiving a course of 2*-deoxy-5-fluorouridine(FUDR; 100 mg/kg, i.p. four times per day3 5), initiated at the time ofoperation, were also evaluated to observe the effects of chemotherapy onliver regeneration. Forty-eight hours after resection, hepatic nucleosidetriphosphate (NTP), which reflects ATP content, fell 37% (P < 0.03) inanimals undergoing hepatectomy alone. By contrast, animals receivingFUDR after hepatectomy demonstrated a mitigated NTP response, with adrop of only 17% (P 5 not significant), suggesting that interruption ofDNA synthesis leads to a reduced consumption of ATP. Direct measuresof DNA synthesis and nuclear proliferation were correlated with NMRfindings. [3H]Thymidine incorporation and Ki67 immunohistochemistrywere performed on liver samples from rats undergoing 70% hepatectomywith and without FUDR. Both [ 3H]thymidine incorporation and Ki67expression were inhibited significantly at 48 h in animals receiving hep-atectomy and FUDR, compared with those not treated with FUDR. Todetermine whether NMR changes could be used to identify safe timing ofchemotherapy after hepatectomy, rats were treated with a 5-day course ofFUDR initiated either prior to or after NMR changes normalized. Animalstreated with FUDR at the point of NTP normalization (72 h) showedsignificantly improved survival over those that began treatment at oper-ation (75% versus17%; P 5 0.0005, log rank test). FUDR inhibits hepaticDNA synthesis and influences mortality if administered too early afterhepatectomy. Chemical shift imaging is a noninvasive tool that can iden-tify metabolic changes coinciding with DNA synthesis and nuclear prolif-eration after hepatectomy.31P-NMR may be useful for determining safetiming of chemotherapy after liver resection.
INTRODUCTION
In the United States,;50,000 patients will develop hepatic metas-tases from primary colorectal cancer each year (1). Currently, hepaticresection provides the only opportunity for cure in this group ofpatients; however, two-thirds of those who undergo resection developrecurrent disease (2). The most common site of recurrence is in theremaining liver (3), suggesting that the unresected portion oftenharbors undetectable microscopic foci of disease. Furthermore, sev-
eral animal models have demonstrated that partial hepatectomy actu-ally accelerates growth of residual microscopic disease (4–6).
The benefits of adjuvant chemotherapy for primary colorectal can-cer (7, 8) and its role after hepatic resection for metastatic disease (9)have been demonstrated. Most oncologists, however, are reluctant toinitiate therapy until 4 weeks after liver resection, because of concernsthat cytotoxic agents will interfere with the process of DNA synthesisin the regenerating liver. These concerns prevail, despite experimentalevidence obtained in animal studies, which demonstrate the peak ofhepatic DNA synthesis to occur before 72 h after 70% hepatectomy(10, 11).
DNA synthesis in the regenerating human liver is difficult tomeasure, because methods used to obtain this information in animalmodels are invasive. Furthermore, medications and co-morbiditiessuch as cirrhosis and hepatitis may alter the timing and extent of theregenerative process (12, 13).31P-NMR3 spectroscopy has been usedto assess energy metabolism after partial hepatectomy in animalstudies, although previous studies have required laparotomy for place-ment of the surface coil (14, 15). Three-dimensional31P CSI is anoninvasive31P-NMR technique that can be used to measure levels ofhigh-energy phosphate compoundsin situ (16). This study investi-gates the application of31P-NMR using the CSI technique to measureenergy changes in rat livers after standard 70% hepatectomy. Itevaluates the effects of chemotherapy on liver regeneration through31P-NMR and correlates these changes with direct measurements ofhepatic DNA synthesis and nuclear proliferation. Finally, it examineswhether31P-NMR can be used as a marker of liver regeneration todetermine safe timing of chemotherapy after hepatic resection.
MATERIALS AND METHODS
Partial Hepatectomy
All animal work was performed under the guidelines approved by theMemorial Sloan-Kettering Institutional Animal Care and Use Committee.Adult male Fischer rats (Charles River, Wilmington, MA), weighing between280 and 350 g, were housed in pathogen-free quarters in the animal facility.The animals were maintained in a 12-h day/night cycle and provided access torat chow (PMI Mills, St. Louis, MO) and waterad libitum until the time ofoperation, at which point animals were pair-fed to the poorest eaters. Partial(70%) hepatectomy was performed similar to the method described previously(17). Briefly, animals were anesthetized with pentobarbital sodium (50 mg/kg;Wyeth Laboratories, Inc., Philadelphia, PA) by i.p. injection. Under sterileconditions, laparotomy was performed through a midline incision. Left andmedian hepatic lobes were identified, ligated with 3-0 silk suture (Ethicon,Inc., Somerville, NJ), and removed. Abdominal closure was performed in twolayers with 4-0 nylon suture (Ethicon, Inc.), and all animals received 3 ml offluid resuscitation (0.9% NaCl by i.p. injection) at the end of the procedure. Allsham-operated animals underwent similar operative exposure and gentle ma-nipulation of the left and median hepatic lobes to control for surgical stress. Alloperations were performed between the hours of 10 a.m. and noon to preventeffects of diurnal mitotic variation (18).
Received 11/8/99; accepted 5/17/00.The costs of publication of this article were defrayed in part by the payment of page
charges. This article must therefore be hereby markedadvertisementin accordance with18 U.S.C. Section 1734 solely to indicate this fact.
1 Supported by Training Grant T32 CA 09501 from the USPHS and in part byGrants RO1CA76416, RO1CA72632, and RO1CA61524 (to Y. F.) and Grant1R24CA83084–01 (to J. A. K.) from the NIH and Grant MBC-99366 (to Y. F.) from theAmerican Cancer Society.
2 To whom requests for reprints should be addressed, at Department of Surgery,Memorial Sloan-Kettering Cancer Center, 1275 York Avenue, New York, NY 10021.Phone: (212) 639-2016; Fax: (212) 639-4031; E-mail: [email protected].
3 The abbreviations used are:31P-NMR, phosphorous-31 nuclear magnetic resonance;5-FU, 5-fluorouracil; CSI, chemical shift imaging; FUDR, 29-deoxy-5-flurouridine; MRI,magnetic resonance imaging; NTP, nucleoside triphosphate; PC, phosphocholine; PE,phosphoethanolamine; PME, phosphomonoester; qd, four times per day.
A total of 61 rats were used to observe the31P-NMR spectral changesassociated with partial hepatectomy and chemotherapy as summarized in Table1. Seventeen animals underwent 70% hepatectomy; 11 animals underwentsham-laparotomy and began a course of FUDR (100 mg/kg, i.p. qd3 5; RocheLaboratories, Nutley, NJ); and 16 rats underwent 70% hepatectomy andreceived FUDR therapy as described. At 48, 72, and 96 h after operation, fiveor six animals from each group were subjected to31P-NMR (at 96 h, fouranimals undergoing hepatectomy and FUDR were evaluated, and no sham-laparotomy FUDR controls were examined). Six additional animals subjectedto hepatectomy underwent31P-NMR at 120 h. To obtain baseline metabolitelevels, five animals undergoing sham-laparotomy and six nonoperated controlswere analyzed.
Localized 31P-NMR was used to monitor the individual and combinedeffects of chemotherapy (FUDR) and 70% hepatectomy on liver energymetabolism. Studies were performed on a 4.7 Tesla Bruker-CSI spectrometerwith a 33-cm horizontal bore as described previously (19). Animals wereanesthetized with isoflurane inhalation and placed prone with the abdomenpositioned over a 31-mm diameter, two-turn phosphorus surface coil. Forquantification, a 65-ml sphere containing methylene diphosphonic acid dilutedin water (50%), and HCl (50%) was located at the center of the coil. The entire31P platform was positioned inside a proton-tuned birdcage resonator, whichwas used to obtain T1-weighted, cross-sectional, anatomical scout images ofthe animal (TR5 500 ms;TE5 10 ms; averages5 4; slice thickness5 3 mm;slice separation5 1 mm; FOV 5 72 mm). Obtaining31P-NMR data by thetechnique of CSI permits noninvasive localization of spectra to the liver andcoregistration of the spectral grids with the MR images (16). Through thisrelationship,in vivo quantitative information can be obtainedin situ withoutperforming laparotomy for placement of a31P coil directly on the organ’ssurface.
The CSI pulse sequence consisted of a hard pulse, phase-encoding, andacquisition, using 10–18 averages with aTR of 0.5 s. The field of view waseither 72 or 64 mm, depending on the size of the liver, and an 83 8 3 8 matrixwas encoded. Total scan time ranged from 43 to 80 min. After corrections forsaturation, flip angle, and receive coil sensitivity, single voxel spectra (voxelsize of 512 or 729 mm3) were quantified by comparison to the methylenediphosphonic standard. Results obtained were compared with those gatheredfor sham-operated, pair-fed controls. Saturation factors were calculated basedupon metabolite T1 values measured in control animals.
31P-NMR data were processed using SAGE/IDL (GE, Milwaukee WI; RSI,Boulder, CO), and MRUI-AMARES software was used for time-domainfitting of resonances (20). Direct quantification of NTPs, Pi, PMEs, and pHlevels were obtained using these methods. NTP levels were based upon thepeak area of theb-NTP moiety, which is most representative of the triphos-phate pool (21). pH measurements were based on the chemical shift betweenPi anda-NTP.
Measurements of DNA Synthesis and Nuclear Proliferation
[3H]Thymidine Incorporation. [3H]Thymidine incorporation was used tomeasure the effects of FUDR therapy on DNA synthesis in regenerating rat
liver to determine whether31P-NMR metabolite measurements correlate di-rectly with nuclear events. Thirty-one adult male Fischer rats were used in thisportion of the study. Briefly, 24 rats underwent 70% hepatectomy, followed bydaily i.p. injections of 0.9% NaCl (n5 12) or FUDR (100 mg/kg;n 5 12),initiated at the time of operation. At each of four time points after resection(24, 48, 72, and 96 h) three NaCl-treated and three FUDR-treated animals eachreceived [3H]thymidine (NEN Life Science Products, Inc., Boston, MA; S.A.20 Ci/mmol), 0.5mCi/g body weight, in 1 ml of 0.9% NaCl via i.p. injection.Animals were sacrificed 1 h later by exsanguination, and liver tissue washarvested for standard genomic DNA extraction. [3H]Thymidine incorporationwas determined by scintillation counting on a Beckman LS6000IC (BeckmanInstruments, Fullerton, CA). The radioactivity of each sample was normalizedby DNA concentration, and amount of DNA synthesis was expressed ascpm/mg DNA. Additional rats undergoing sham-laparotomy with (n 5 4) andwithout (n 5 3) FUDR therapy served as baseline controls.
Ki67 Immunohistochemistry. Ki67 immunohistochemistry was per-formed on liver specimens from 30 pair-fed Fischer rats at various time pointsafter 70% hepatectomy. Half the rats (n5 15) received a course of FUDR (100mg/kg i.p., qd3 5) initiated at operation, whereas the remaining animals(n 5 15) were treated with an equal volume of 0.9% NaCl, as control. At eachtime point (24, 48, 72, 96, and 120 h) after 70% hepatectomy, three rats fromeach treatment group were sacrificed, and their livers were harvested and fixedin 4% paraformaldehyde (Sigma Chemical, St. Louis, MO). Tissues weresubsequently transferred to 70% ethanol, embedded in paraffin blocks, andstored at 4°C until staining. Paraffin blocks were sectioned and boiled in 0.01M citric acid (pH 6.0), in a microwave oven for 10 min to unmask tissueantigens. Sections were incubated at 4°C overnight with mouse monoclonalanti-mouse Ki67 antibody (NCL-Ki67-MM1; Novocastra Laboratories, New-castle upon Tyne, United Kingdom) diluted 1:200 in PBS. Specimens wereincubated with biotinylated antimouse secondary antibody (Histomouse-SpKit; Zymed Laboratories, San Francisco, CA), followed by peroxidase-conju-gated streptavidin (Zymed) and finally, a chromagen (AEC)-substrate mixture(Zymed). Ki67-positive cells were quantified by computer-assisted imageanalysis of specimens from three animals/group and three slides/animal, usingImage-Pro Plus software (Media Cybernetics, Silver Spring, MD). Resultswere determined as the percentage of positive (red/brown) nuclei at3200.Livers from three additional naı̈ve animals served as baseline controls.
Survival Study
This study was performed to determine whether hepatic energy changesidentified by31P-NMR might be useful to guide administration of chemother-apy after partial hepatectomy. Sixty adult male Fischer rats were divided intofive groups of 12 animals. Thirty-six rats underwent 70% hepatectomy andreceived a course of FUDR (100 mg/kg i.p., qd3 5) beginning at operation(n 5 12, “immediate”), 72 h after 70% hepatectomy (n5 12, “early”), or 168 hafter operation (n5 12, “late”). The 72-h time point was selected to exploit theperiod closest to that when31P-NMR energy changes first recovered. Twelverats underwent 70% hepatectomy and were treated with a 5-day course of anequal volume of 0.9% NaCl initiated at operation to control for injectiontrauma. Twelve sham-operated animals received a 5-day course of FUDR
Table 1 In vivo comparison of phosphorous metabolites in adult Fischer rat livers as measured using31P-NMR by CSI
96 h hepatectomy 6 6.246 0.11 5.876 1.33 0.956 0.22 5.556 1.87 7.336 0.0296 h sham-laparotomy1 FUDR NAb NA NA NA NA NA96 h hepatectomy1 FUDR 4 5.886 0.64 6.726 2.49 1.136 0.36 5.816 1.71 7.396 0.04
120 h hepatectomy 6 6.246 0.87 5.106 0.88 0.836 0.17 6.766 1.21 7.386 0.02a P , 0.05, as compared with sham-laparotomy pair-fed controls.b NA, this group was not evaluated.
3801
31P-NMR STUDY OF LIVER REGENERATION AND CHEMOTHERAPY
initiated at operation to control for the effects of chemotherapy alone. Allinjections were administered i.p., and all animals were pair-fed to pooresteaters. Survival was recorded until 40 days postoperatively.
Statistics
Comparisons of31P-NMR metabolite levels, [3H]thymidine incorporation,and Ki67 expression were performed using Student’st test (two-tailed). Sig-nificance was defined by aP , 0.05 (two-sided). Survival analyses wereperformed by the Kaplan-Meier method (22). Log-rank testing was used todetermine significance of survival curves (23, 24).
RESULTS31P-NMR Spectroscopy
31P-NMR was performed on 61 adult male Fischer rats to identifychanges in hepatic high-energy phosphate levels associated with liverregeneration. Additionally, animals undergoing 70% hepatectomy andsimultaneous administration of FUDR were evaluated by31P-NMR todetect alterations in liver metabolites induced by this agent and furthercharacterize and validate concerns regarding administration of chem-otherapy during active DNA synthesis. FUDR was selected for thisstudy because this fluorinated pyrimidine is known to interfere withDNA synthesis, and it is the drug of choice for therapy of gastroin-testinal liver metastases.
Fig. 1 illustrates results obtained by31P-NMR. This is a represent-ative T1-weighted, cross-sectional MRI through the liver of a Fischerrat lying in the prone position (spinal cord is seen near the top centerof the image). Superimposed on the MRI is the corresponding31P-NMR spectral grid. Each CSI voxel measures 93 9 3 9 mm for avolume of 729 mm3. The highlighted voxel represents a spectrumfrom liver parenchyma only. The various peaks correspond to phos-phorous metabolite levels present in the region of interest.
Our analysis included absolute quantitation of levels of NTPs, Pi,PMEs, and pH at various time points after partial hepatectomy inanimals both with and without simultaneous administration of FUDRtherapy. Table 1 shows summary data for this set of experiments. All
results represent the average of five or six studies performed indifferent animals and expressed in mM/l 6 SD. The quantitationprocedure did not correct for point spread function (25). In phantomstudies, we have estimated that the point spread function results in a30% overestimation of metabolite concentrations.
A significant drop in hepatic NTP (37%;P , 0.05) was observed48 h after 70% hepatectomy, when compared with sham-laparotomypair-fed controls (Fig. 2,A andB). NTP levels recovered by 72 h andnormalized completely by 96 h. Pi levels did not change significantly;however, the hepatic Pi:NTP ratio increased in hepatectomized ani-mals at the 48-h time point (1.376 0.26; P , 0.05). Animals thatunderwent 70% hepatectomy and simultaneous FUDR therapy dem-onstrated a mitigated NTP response at 48 h compared with hepatec-tomized animals that did not receive chemotherapy (Figs. 2D and 3).NTP levels in this group were depleted only 17% (P5 not signifi-cant). Both Pi levels and Pi:NTP ratios did not change significantly inthis group either.
Compared with sham-laparotomy controls, PME metabolites levelsdid not change significantly in animals that underwent 70% hepa-tectomy alone or in those that had sham-laparotomy and simultaneousFUDR therapy. By contrast, significantly elevated PME levels wereobserved in the group of animals that had 70% hepatectomy andsimultaneous FUDR therapy at 48 h after operation (Table 1). Nosignificant pH fluctuations were observed in any of the groups eval-uated.
Measurements of DNA Synthesis and Nuclear Proliferation
[3H]Thymidine Incorporation. [3H]Thymidine incorporation wasmeasured in liver samples from animals treated both with and withoutsimultaneous FUDR therapy, at various time points after 70% hepa-tectomy. Liver samples from both sham-operated and sham-operatedFUDR-treated animals demonstrated minimal [3H]thymidine incorpo-ration. Peak uptake in hepatectomized animals occurred 48 h afteroperation (Fig. 4). Tissue from hepatectomized animals that did notreceive FUDR therapy demonstrated 41% greater [3H]thymidine in-
Fig. 1. Phosphorus-31 chemical shift spectral grid overlyingcorresponding T1-weighted axial image through the liver of aFischer rat. The animal is lying prone, with its spine situated nearthe top center of the image. The voxel of interest (highlighted byawhite box) is centered within liver tissue, and the various peaksrepresent averaged concentrations of phosphorous metabolitespresent in that voxel.
3802
31P-NMR STUDY OF LIVER REGENERATION AND CHEMOTHERAPY
corporation at 48 h than did tissue from animals that did receivechemotherapy (85versus50 cpm/mg DNA; P , 0.05). An overallreduction in the process of DNA synthesis was observed at all mea-sured time points when FUDR was given. These results suggest thatFUDR significantly inhibits DNA synthesis in regenerating liver.
Ki67 Immunohistochemistry. Ki67 immunohistochemical stain-ing of paraffin-fixed liver tissue was performed to measure changes inhepatocellular nuclear proliferation after partial hepatectomy in ani-mals treated both with and without simultaneous FUDR therapy. Ki67antigen expression was negative in nonhepatectomized controls and inliver sections obtained 24 h after hepatic resection. Expression peaked48 h after partial hepatectomy, fell at 72 h, and returned to baseline by96 h (Fig. 5). Significantly lower expression was observed in FUDRtreated animal livers at 48 h compared with sections from animals thatdid not receive chemotherapy (46%;P , 0.05). These data further
suggest that FUDR significantly inhibits the regenerative process inthe posthepatectomy liver at the nuclear level. Furthermore, the timingof antigen expression correlates with results obtained by31P-NMRand [3H]thymidine incorporation.
Survival Study
This survival study was undertaken to evaluate31P-NMR as a toolfor identifying metabolic changes in regenerating liver useful fortiming administration of chemotherapy after partial hepatectomy.Hepatectomized animals were treated with “immediate FUDR”(n 5 12), “immediate saline” (n5 12), “early FUDR” (n5 12) or“late FUDR” (n 5 12). Twelve animals underwent sham-laparotomyand “immediate FUDR” therapy as additional controls. The hepatec-tomized group that received “immediate FUDR” demonstrated thehighest mortality, with a cumulative survival of only 17% by day 12posthepatectomy. In this group, FUDR therapy was initiated prior tothe peak of NTP depletion, as measured using31P-NMR, and DNAsynthesis and nuclear proliferation, as measured by [3H]thymidineincorporation and Ki67 antigen immunostaining, respectively. The“early FUDR” hepatectomy group showed significantly improvedsurvival (75%) over the “immediate FUDR” group (Fig. 6;P 5 0.0005). This group began FUDR treatment 72 h after hepatec-tomy, just after the recovery of NTP, as observed by31P-NMR.Excellent survival was observed in the sham-laparotomy FUDR-treated animals (100%), the hepatectomy “early saline” animals(83%), and the hepatectomy “late FUDR” group (92%). All animalsthat survived more than a week beyond completion of chemotherapygained weight appropriately and survived long term.
DISCUSSION
Hepatic resection is potentially curative for patients with primaryhepatocellular cancer (26) and isolated liver metastases from colorectalmalignancies (2). Many of these patients, however, will develop hepaticrecurrence despite appropriate patient selection and sound operative tech-nique because of undetected microscopic disease already present at thetime of resection. Furthermore, after hepatectomy these micrometastasesare subject to the surge of growth factors associated with the normalprocess of liver regeneration and may be stimulated to grow at anaccelerated rate (4, 5, 27). Recent clinical evidence demonstrates thatadministration of adjuvant chemotherapy can improve results over sur-
Fig. 2. 31P-NMR spectra obtained from the livers of animals 48 h after operation.A,sham-laparotomy;B, 70% hepatectomy;C, sham-laparotomy and FUDR;D, 70% hepa-tectomy and FUDR.g, a, b-NTP, 5 g, a, andb moieties of nucleoside triphosphates.NTP content is best represented byb-NTP peak. Note drop inb-NTP peak inB (arrow),which is not seen inD.
Fig. 3. Mean NTP levels as determined at various time points after 70% hepatectomyin adult Fischer rats by31P-NMR. f, data for hepatectomized animals;M, data forhepatectomized animals that received a course of FUDR, initiated at the time of operation.Bars,SD. p, P , 0.05.
Fig. 4. Mean [3H]thymidine incorporation into hepatic DNA after 70% hepatectomy inadult Fischer rats.f, data for hepatectomized animals;M, data for hepatectomizedanimals that received a course of FUDR, initiated at the time of operation.Bars, SEM.Differences between hepatectomy groups and hepatectomy-FUDR groups are significantat all time points after operation (p,P , 0.05).
3803
31P-NMR STUDY OF LIVER REGENERATION AND CHEMOTHERAPY
gical resection alone (9), and preclinical evidence suggests that earlieradministration of chemotherapy after hepatectomy may enhance effectsof these agents (28, 29). Currently, however, routine practice is to wait 4weeks after resection before initiating this treatment for fear of detrimen-tally altering the process of hepatic DNA synthesis that is essential forsurvival after liver resection. A noninvasive surrogate marker of hepaticDNA synthesis would potentially benefit these patients by permittingearlier initiation of chemotherapy.
Numerous liver-directed applications for NMR spectroscopy arebeing explored as more sophisticated systems are developed. NMR isbeing used experimentally to evaluate hepatic function in disease (30,31) and after transplantation (32), trace therapeutic metabolites (33–35), and evaluate response to therapy for various disease processes(36). Several studies have used NMR to evaluate spectral changesduring the process of liver regeneration. Early efforts used protonNMR to measure lipid changes as determined through T1 and T2
relaxation times (37–40). Further insight was provided by studiesusing31P-NMR, which permitsin vivo, whole-organ relative-quanti-fication of phospholipid and phosphoenergetic alterations. Campbellet al. (15) documented a drop in hepatic NTP levels with a concom-itant rise in Pi:NTP 48 h after 70% hepatectomy in rats. They reasonedthat in the absence of necrosis, these changes reflect ATP hydrolysisassociated with the energy-requiring process of hepatic DNA synthe-sis. Another study by Farghaliet al. (14) compared31P-NMR spectralchangesin vivo in regenerating livers of Sprague Dawley rats within
vitro spectra obtained from perchloric acid extracts of correspondingliver tissue. They reported a relative correlation between ATP levelsmeasured by both methods, suggesting that31P-NMR provides accu-rate biochemical quantification of phosphate compound metabolismin this system.
The current study adds to the previous work by correlating31P-NMR spectra in regenerating liver to direct measures of hepatic DNAsynthesis and nuclear proliferation. It also investigates the effects ofchemotherapy on the process of DNA synthesis and31P metabolism,when administered at the time of resection. Most importantly, thepresent study documents such spectral changes to be potentially usefulin directing administration of adjuvant chemotherapy after partialhepatectomy.
Prior to implementation of CSI, it was usually the practice toexpose the liver for direct placement of the surface coil to avoidcontamination of spectra by signals from surrounding tissue (14, 15,41, 42). Therefore, previous31P-NMR studies of liver regenerationwere actually invasive. Furthermore, hepatic31P-NMR changes havebeen associated with sham-laparotomy alone (40); thus, the process ofoperating on the animal to place the surface coil on the liver remnantpotentially alters results. In the current study, spectra are obtainedwithout performing a second operation. The31P coil remains externalto the animal, and these data are superimposed on a T1-weightedproton MR image in three dimensions, as described previously (16);thus, the results are not influenced by an additional laparotomy.
In most NMR studies, metabolite ratios, rather than absolute values,are used to demonstrate changes occurring in tissues. It can bedifficult, however, to ascertain which metabolite is having a greateraffect on the change in ratio. Use of an external standard to quantifyeach NMR peak provides a more specific and robust method forassessing changes in individual metabolites. This technique requiresknowledge of the B1 profile the coil, tissue T1 values, and carefulcalibration; however, it enables us to report individual metabolitechanges with confidence.
The hepatic phosphoenergetic changes we detected support theobservations of Campbellet al. (15). We observed a significantdepletion in hepatic NTP levels with an associated rise in the Pi:NTPratio 48 h after partial hepatectomy, which nearly recovered by 72 h.Because the NTP change was significant and the Pi change was not,it is apparent that the NTP depletion provided a greater contribution to
Fig. 5. Comparison of Ki67 expression after 70% hepatectomy in adult Fischer rats,both with and without simultaneous FUDR therapy.A, two photomicrographs (3200)showing hepatic Ki67 expression 48 h after operation in a hepatectomized rat (left panel)and a hepatectomized rat treated with FUDR (right panel).B, chart demonstratingpercentage of nuclei that are Ki67 positive over time in untreated, hepatectomized animals(f) and in hepatectomized animals treated with FUDR (E).p, P , 0.05.
Fig. 6. Results of survival study. Hepatectomized adult Fischer rats treated with “earlyFUDR” (initiated 72 h after hepatectomy,solid line)versus“immediate FUDR” (initiatedat hepatectomy,broken line).P 5 0.0005.
3804
31P-NMR STUDY OF LIVER REGENERATION AND CHEMOTHERAPY
the Pi:NTP rise. These31P-NMR changes correlated with direct mea-sures of hepatic DNA synthesis and proliferation as determined byincorporation of [3H]thymidine and expression of the nuclear antigenKi67. Previous studies have examined the effects of fluoropyrimidineson liver regeneration (43, 44). In our study, introduction of a courseof FUDR therapy at the time of operation partially inhibited theenergy-requiring process of DNA synthesis in the regenerating liver,as determined by [3H]thymidine incorporation and Ki67 immunohis-tochemistry. Corresponding hepatic NTP levels and Pi:NTP ratioswere affected in these animals as well. The significant NTP depletionobserved at 48 h in hepatectomized animals was not witnessed in thehepatectomized group treated with FUDR, presumably because lessATP was being used forde novoDNA synthesis. These observationssupport the premise that31P-NMR can detect energy fluctuationsreflective of DNA synthesis and hepatocyte proliferation.
Numerous studies have evaluated the process of liver regenerationin animal models using varied techniques such as [3H]thymidineincorporation and labeling of various antigens expressed in prolifer-ating cells such as Ki67, proliferating cell nuclear antigen, and bro-modeoxyuridine (11, 45–47). Some variation in results is observeddepending on the model being examined and the percentage of liverresected, but these methods are reliable and reproducible. [3H]Thy-midine is a radiolabeled nucleotide that incorporates into DNA as it isbeing synthesized; therefore, it is intimately related to S-phase. TheKi67 antigen is expressed in all phases of the cell cycle except G0 andearly G1 (45); therefore, Ki67 expression may persist despite absenceof [3H]thymidine incorporation. This basic difference in the informa-tion provided by these techniques can account for some inconsistencyin the data at the various time points. The information obtained by[3H]thymidine incorporation is a better representation of DNA syn-thesis and probably correlates more closely to the changes in hepaticNTP levels observed by31P-NMR.
In addition to evaluating phosphoenergetic changes,31P-NMR pro-vides information on phospholipid precursor levels. In our study, thetotal PME concentration did not change significantly after hepatec-tomy alone. This peak is comprised of PC, PE, and sugar phosphates.Somein vivo studies have reported elevated PME ratios in the daysafter partial hepatectomy (14, 48), whereas others have not (41). Allof these studies were performed by surgically exposing the liver fordirect application of the surface coil, which may confound results, asmentioned previously.In vitro studies have shown that PE is elevatedafter partial hepatectomy (14, 41, 48); however, PE may compriseonly a small fraction of total PMEs, making the change difficult todetectin vivo. Additionally, PC levels have been shown to be depen-dent on diet (41), and although PE levels may be elevated, total PMEconcentration may be unchanged or even reduced based on the drop inPC associated with pair-feeding and decreased food intake in thepostoperative period.
Interestingly, a significantly elevation of PME was observed in ourstudies at 48 h after hepatectomy, when a course of FUDR therapywas administered. Previousin vivo NMR studies of the effects ofantineoplastic treatment on tumor metabolism have noted changes inboth PE and PC levels. Specifically, PE:PC was noted to increase afteradministration of both chemotherapy and radiation (49–52). 5-Fluorouracil has been shown to induce a 70% increase in PE in amammary carcinoma, with a minimal (nonsignificant) decrease in PC(51). These studies, and others have interpreted these changes to beattributable to decreased cellular proliferation or increased cell death.Thus, the increased PME observed in the FUDR-treated hepatecto-mized animals may reflect chemotherapy-induced damage to theregenerating liver. Proton-decoupled phosphorus-31 CSI could beused to clarify findings in the regenerating liver by improving reso-
lution of PE and PC peaks, as well as by increasing signal-to-noiseratio.
Perhaps the most interesting and clinically relevant observation isthe significant improvement in survival seen in animals in whichFUDR therapy is initiated 72 h after partial hepatectomy. HepaticNTP levels, [3H]thymidine incorporation, and Ki67 expression allnormalize substantially by this time point. These data suggest that thecurrent practice of waiting 4 weeks after hepatic resection to initiateadjuvant chemotherapy may be unnecessary, and that31P-NMR canbe used to identify hepatic high-energy phosphate changes reflectiveof DNA synthesis during the process of liver regeneration. It must beemphasized that we continue to caution against early use of chemo-therapy immediately after NTP normalization. Clinical trials will beneeded to determine how soon after NTP normalization adjuvantchemotherapy can safely be administered. Clearly though, if a patienthas prolonged depression of NTP levels, chemotherapy should not beadministered. Hepatic31P-MR-spectroscopy may also be used tomonitor toxicity of adjuvant chemotherapy after liver resection.
ACKNOWLEDGMENTS
We thank Karen Witty-Blease and Katia Manova for helping perform Ki67immunohistochemistry and Yong-jia You for technical support.
REFERENCES
1. Fong, Y., Blumgart, L. H., and Cohen, A. M. Surgical treatment of colorectalmetastases to the liver. CA Cancer J. Clin.,45: 50–62, 1995.
2. Fong, Y., Cohen, A. M., Fortner, J. G., Enker, W. E., Turnbull, A. D., Coit, D. G.,Marrero, A. M., Prasad, M., Blumgart, L. H., and Brennan, M. F. Liver resection forcolorectal metastases. J. Clin. Oncol.,15: 938–946, 1997.
3. Willett, C. G., Tepper, J. E., Cohen, A. M., Orlow, E., and Welch, C. E. Failurepatterns following curative resection of colonic carcinoma. Ann. Surg.,200: 685–690, 1984.
4. Panis, Y., Ribeiro, J., Chretien, Y., and Nordlinger, B. Dormant liver metastases: anexperimental study. Br. J. Surg.,79: 221–223, 1992.
5. Fisher, E. R., and Fisher, B. Experimental studies of factors influencing hepaticmetastases. Cancer (Phila.),12: 929–932, 1959.
6. Paschkis, K. E., Cantarow, A., Stasney, J., and Hobbs, J. H. Tumor growth in partiallyhepatectomized rats. Cancer Res.,15: 579–582, 1955.
7. Grage, T., and Moss, S. Adjuvant chemotherapy in cancer of the colon and rectum:demonstration of effectiveness of prolonged 5-FU chemotherapy in a prospectivelycontrolled randomized trial. Surg. Clin. N. Am.,61: 1321–1329, 1981.
8. Mamounas, E., Rockette, H., Jones, J., Wieand, D., Fisher, B., and Wolmark, N.Comparative efficacy of adjuvant chemotherapy in patients with Duke’s BversusDuke’s C colon cancer: results from four NSABP adjuvant studies. Proc. Am. Soc.Clin. Oncol., Abstract, 205, 1996.
9. Kemeny, N., Cohen, A. M., Huang, W., Shi, W., Brennan, M. F., Bertino, J. R.,Turnbull, A. D. M., Sullivan, D., Stockman, J., Blumgart, L. H., and Fong, Y.Randomized study of hepatic arterial infusion and systemic chemotherapyversussystemic chemotherapy alone as adjuvant therapy after resection of hepatic metastasesfrom colorectal cancer. Proc. Am. Soc. Clin. Oncol., Abstract, 1999.
10. Fabrikant, J. The kinetics of cellular proliferation in regenerating liver. J. Cell Biol.,36: 551–565, 1968.
11. Assy, N., Gong, Y., Zhang, M., Pettigrew, N., Pashniak, D., and Minuk, G. Y. Useof proliferating cell nuclear antigen as a marker of liver regeneration after partialhepatectomy in rats. J. Lab. Clin. Med.,131: 251–256, 1998.
12. MacIntosh, E. L., Gauthier, T., Pettigrew, N. M., and Minuk, G. Y. Hepatic fibrosisas a predictor of hepatic regenerative activity after partial hepatectomy in the rat.Hepatology,17: 307–309, 1993.
13. Wolf, H. K., and Michalopoulos, G. K. Hepatocyte regeneration in acute fulminantand nonfulminant hepatitis: a study of proliferating cell nuclear antigen expression.Hepatology,15: 707–713, 1992.
14. Farghali, H., Rilo, H., Zhang, W., Simplaceanu, V., Gavaler, J. S., Ho, C., and vanThiel, D. H. Liver regeneration after partial hepatectomy in the rat: sequential eventsmonitored by31P-nuclear magnetic resonance spectroscopy and biochemical studies.Lab. Investig.,70: 418–425, 1994.
15. Campbell, K. A., Wu, Y. P., Chacko, V. P., and Sitzmann, J. V.In vivo 31P NMRspectroscopic changes during liver regeneration. J. Surg. Res.,49: 244–247, 1990.
16. Brown, T. R., Kincaid, B. M., and Ugurbil, K. NMR chemical shift imaging in threedimensions. Proc. Natl. Acad. Sci. USA,79: 3523–3526, 1982.
17. Higgins, G., and Anderson, R. Experimental pathology of the liver. I. Restoration ofthe liver of the white rat following partial surgical removal. Arch. Pathol.,12:186–202, 1931.
18. Jaffe, J. Diurnal mitotic periodicity in regenerating rat liver. Anat. Rec.,120: 935,1954.
3805
31P-NMR STUDY OF LIVER REGENERATION AND CHEMOTHERAPY
19. Zakian, K., D’Angelica, M., Matei, C., Fong, Y., and Koutcher, J. A quantitativeassessment of liver metabolites in jaundice using three-dimensional phosphoruschemical shift imaging. Proceedings of the International Society of Magnetic Reso-nance Medicine, Sixth Scientific Meeting, p. 1859, 1998.
20. van der Veen, J. W. C., de Beer, R., Luyten, P. R., and van Ormondt, D. Accuratequantification ofin vivo 31P NMR signals using the variable projection method andprior knowledge. Magn. Reson. Med.,6: 92–98, 1988.
21. Rudin, M., and Sauter, A.In vivo phosphorus-31 NMR: potential and limitations.In:P. Diehl, E. Fluck, H. Gunther, R. Kosfeld, and J. Seelig (eds.),In Vivo MagneticResonance Spectroscopy, Ed. 1, pp. 161–190. Berlin: Springer-Verlag, 1992.
22. Kaplan, E., and Meier, P. Nonparametric estimation from incomplete observations.J. Am. Stat. Assoc.,53: 457–462, 1958.
23. Mantel, N. Evaluation of survival data and two new rank order statistics arising in itsconsideration. Cancer Chemother. Rep.,50: 163–170, 1966.
24. Cox, D. Regression models and life tables. J. R. Stat. Soc.,34: 197–220, 1972.25. Murphy-Boesch, J., Jiang, H., Stoyanova, R., and Brown, T. R. Quantitation of
phosphorus metabolites from chemical shift imaging spectra with corrections forpoint spread effects and B1 inhomogeneity. Magn. Reson. Imaging,39: 429–438,1998.
26. Fong, Y., Sun, R. L., Jarnagin, W., and Blumgart, L. H. An analysis of 412 cases ofhepatocellular carcinoma at a Western center. Ann. Surg.,229: 790–800, 1999.
27. Picardo, A., Karpoff, H. M., Ng, B., Lee, J., Brennan, M. F., and Fong, Y. Partialhepatectomy accelerates local tumor growth: potential roles of local cytokine activa-tion. Surgery,124: 57–64, 1998.
28. Elias, D., Lasser, P., Rougier, P., Nitenberg, G., Theodore, C., and Lumbroso, J. Earlyadjuvant intraportal chemotherapy after curative hepatectomy for colorectal livermetastases–a pilot study. Eur. J. Surg. Oncol.,13: 247–250, 1987.
29. Sutanto-Ward, E., Sigurdson, E. R., Tremiterra, S., Lincer, R., Chapman, D., andNiedzwiecki, D. Adjuvant chemotherapy for colorectal hepatic metastases: role ofroute of administration and timing. Surg. Oncol.,1: 87–95, 1992.
30. Otsuka, H., Harada, M., Koga, K., and Nishitani, H. Effects of hepatic impairment onthe metabolism of fructose and 5-fluorouracil, as studied in fatty liver models usingin vivo 31P-MRS and 19F-MRS. Magn. Reson. Imaging,17: 283–290, 1999.
31. Brauer, M., Lu, W., and Ling, M. Chronic ethanol administration alters hepaticrates of glycerol phosphorylation and glycerol 3-phosphate oxidation: a dynamicin vivo 31P magnetic resonance spectroscopy study. Biochem. Cell Biol.,76:542–552, 1998.
32. Davidson, B. R., Barnard, M. L., Changani, K. K., and Taylor-Robinson, S. D. Livertransplantation: current and potential applications of magnetic resonance spectros-copy. Liver Transplant Surg.,3: 481–493, 1997.
33. Li, C. W., Negendank, W. G., Padavic-Shaller, K. A., O’Dwyer, P. J., Murphy-Boesch, J., and Brown, T. R. Quantitation of 5-fluorouracil catabolism in human liverin vivo by three-dimensional localized 19F magnetic resonance spectroscopy. Clin.Cancer Res.,2: 339–345, 1996.
34. Gonen, O., Murphy-Boesch, J., Li, C. W., Padavic-Shaller, K., Negendank, W. G.,and Brown, T. R. Simultaneous 3D NMR spectroscopy of proton-decoupled fluorineand phosphorus in human liver during 5-fluorouracil chemotherapy. Magn. Reson.Med., 37: 164–169, 1997.
35. Adams, E. R., Leffert, J. J., Craig, D. J., Spector, T., and Pizzorno, G.In vivo effectof 5-ethynyluracil on 5-fluorouracil metabolism determined by 19F nuclear magneticresonance spectroscopy. Cancer Res.,59: 122–127, 1999.
36. Kiyono, K., Shibata, A., Sone, S., Wantanabe, T., Oguchi, M., Shikama, N., Ichijo, T.,Kiyosawa, K., and Sodeyama, T. Relationship of31P MR spectroscopy to the
histopathological grading of chronic hepatitis and response to therapy. Acta Radiol.,39: 309–314, 1998.
37. de Certaines, J. D., Moulinoux, J. P., Benoist, L., Bernard, A. M., and Rivet, P. Protonnuclear magnetic resonance of regenerating rat liver after partial hepatectomy. LifeSci., 31: 505–508, 1982.
38. Fantazzini, P., Lendinara, L., Novello, F., Brosio, E., and Di, N. A. The water protonspin-lattice relaxation time measurements in liver samples from hepatectomized andsham-operated rats. Biochim. Biophys. Acta,803: 250–253, 1984.
39. Sostman, H. D., Gore, J. C., Flye, M. W., Johnson, G. A., and Herfkens, R. J. Timecourse and mechanism of alterations in proton relaxation during liver regeneration inthe rat. Hepatology,5: 538–543, 1985.
40. Pizzarello, D. J., and Chandra, R. Effects of mitotic activity and water content on theproton relaxation times in partially hepatectomized and sham-operated rat livers.Investig. Radiol.,21: 320–324, 1986.
41. Morikawa, S., Inubushi, T., Kitoh, K., Kido, C., and Nozaki, M. Chemical assessmentof phospholipid and phosphoenergetic metabolites in regenerating rat liver measuredby in vivo and in vitro 31P-NMR. Biochim. Biophys. Acta,1117: 251–257, 1992.
42. Nishida, Y., Tanaka, K., Munakata, T., and Kasai, S. Evaluation of functionalrecovery of regenerating rat liver after partial hepatectomy using31P-NMR spectros-copy. J. Surg. Res.,61: 385–390, 1996.
43. Goodman, J. I., and Potter, V. R. The effect of 5-fluorodeoxyuridine on the synthesisof deoxyribonucleic acid pyrimidines in precancerous rat liver. Mol. Pharmacol.,9:297–303, 1973.
44. Tsukamoto, I., and Shosuke, K. The effects of fluorouracil on thymidylate synthaseand thymidine kinase in regenerating rat liver after partial hepatectomy. Biochim.Biophys. Acta,1074: 52–55, 1991.
45. Gerlach, C., Sakkab, D. Y., Scholzen, T., Dassler, R., Alison, M. R., and Gerdes, J.Ki-67 expression during rat liver regeneration after partial hepatectomy. Hepatology,26: 573–578, 1997.
46. Block, W., Reichel, C., Traber, F., Skodra, T., Lamerichs, R., Kreft, B., Spengler, U.,Sauerbruch, T., and Schild, H. H. Effect of cytochrome P450 induction on phosphorusmetabolites and proton relaxation times measured byin vivo 31P-magnetic resonancespectroscopy and 1H-magnetic resonance relaxometry in human liver. Hepatology,26: 1587–1591, 1997.
47. Lambotte, L., Saliez, A., Triest, S., Tagliaferri, E., Barker, A., and Baranski, A.Control of rate and extent of the proliferative response after partial hepatectomy.Am. J. Physiol.,273: G905–G912, 1997.
48. Murphy, E., Brindle, K., Rorison, C., Dixon, R. M., Rajagopalan, B., and Radda, G.Changes in phosphatidylethanolamine metabolism in regenerating rat liver as mea-sured by31P-NMR. Biochim. Biophys. Acta,1135: 27–34, 1992.
49. Koutcher, J. A., Alfieri, A. A., Devitt, M. L., Rhee, J. G., Kornblith, A. B., Mahmood,U., and Merchant, T. E. Quantitative changes in tumor metabolism, partial pressureof oxygen, and radiobiological oxygenation status postradiation. Cancer Res.,52:4620–4627, 1992.
50. Mahmood, U., Alfieri, A. A., Thaler, H., Cowburn, D., and Koutcher, J. A. Radiationdose-dependent changes in tumor metabolism measured by31P nuclear magneticresonance spectroscopy. Cancer Res.,54: 4885–4891, 1994.
51. Street, J. C., Alfieri, A. A., Traganos, F., and Koutcher, J. A.In vivoandex vivostudyof metabolic and cellular effects of 5-fluorouracil chemotherapy in a mouse mammarycarcinoma. Magn. Reson. Imaging,15: 587–596, 1997.
52. Street, J. C., Mahmood, U., Matei, C., and Koutcher, J. A.In vivo andin vitro studiesof cyclophosphamide chemotherapy in a mouse mammary carcinoma by31P NMRspectroscopy. NMR Biomed.,8: 149–158, 1995.
3806
31P-NMR STUDY OF LIVER REGENERATION AND CHEMOTHERAPY
![using 31P NMR - NREL · some background on the use of 31P NMR for analysis of bio-oils [1]. 2. Scope 2.1 This procedure has been optimized for the quantification of hydroxyls (-OH)](https://static.documents.pub/doc/80x56/5f0f5f5e7e708231d443d59e/using-31p-nmr-some-background-on-the-use-of-31p-nmr-for-analysis-of-bio-oils-1.jpg)
