int. j. radiat. biol 2002, vol. 78, no. 5, 407 ± 415

Antibody and radionuclide characteristics and the enhancement ofthe eVectiveness of radioimmunotherapy by selective dose deliveryto radiosensitive areas of tumour

A. A. FLYNN*†, R. B. PEDLEY†, A. J. GREEN†, G. M. BOXER†, R. BODEN†,J. BHATIA†, R. MORRIS‡ and R. H. J. BEGENT†

(Received 31 August 2001; accepted 24 November 2001)

Abstract. 1. IntroductionPurpose: Estimating the absorbed dose to tumour relative to

Radioimmunotherapy (RIT) oVers the potential fornormal tissues has often been used in the assessment of thethe selective delivery of radiation to small cancertherapeutic eYcacy of radiolabelled antibodies for radioimmuno-

therapy. Typically, the calculations assume a uniform dose deposits, whilst limiting normal tissue toxicity.deposition and response throughout the tumour. However, the Therapeutic eYcacy has a strong dependence onheterogeneity of the dose delivery and response within tumours tumour type; clinical trials have produced encour-can lead to a radiobiological eVect inconsistent with dose estim-

aging results in lymphomas (Press et al. 1995), butates. The aim was to assess the in� uence of antibody andonly limited responses in solid tumours, such asradionuclide characteristics on the heterogeneity of dose

deposition. colorectal tumours (Lane et al. 1994). Several studiesMaterials and methods: Quantitative images of the temporal and have shown a marked variation in therapeutic eVectspatial heterogeneity of a range of antibodies in tumour were with RIT compared with external beam radiotherapyacquired using radioluminography. Subsequent registration with

for a given radiation dose (Wessels et al. 1989,images of tumour morphology then allowed the delineation ofRoberson and Buchsbaum 1995) . This discrepancy isviable and necrotic areas of tumour and the measurement of the

antibody concentration in each area. A tumour dosimetry model further compounded by the results from other studiesthen estimated the absorbed dose from 131I and 90Y in each area. that have found little relationship between absorbedResults : Tumour-speci� c antibodies initially localized in the viable dose and the eVect in tumour following RIT (Matthayradiosensitive areas of tumour and then penetrated further into

et al. 1991, DeNardo et al. 1999, Koral et al. 2000).tumour with continued tumour accretion. Multivalent antibodiesClearly, there are important aspects of tumour biologywere retained longer and at higher concentrations in viable

areas, while monovalent antibodies had greater mobility. In and dosimetry that need to be addressed to relate thecontrast, non-speci� c antibodies penetrated into necrotic regions absorbed dose to biological eVect.regardless of their size. As a result, multivalent, speci� c antibodies Typically, absorbed dose estimates assume a dosedelivered a signi� cantly larger dose to viable cells compared with

deposition and response throughout the tumour.monovalent antibodies, while non-speci� c antibodies depositedHowever, the tumour vasculature is highly irregularmost of the dose in necrotic areas. There was a signi� cant

diVerence in dose estimates when assuming a unifrom dose and causes the heterogeneous delivery of antibodiesdeposition and accounting for heterogeneity. The dose to the and a variable cellular response. Tumours tend toviable and necrotic areas also depended on the properties of the have well-perfused, radiosensitive cells at the peri-radionuclide where antibodies labelled with 131I generally

phery with a central necrotic core with intermediatedelivered a higher dose throughout the tumour even though thehypoxic, resistant areas ( Jain 1994) . Anti-tumourinstantaneous dose-rate distribution for 90Y was more uniform.

Conclusions : The extent of heterogeneity of dose deposition in antibodies bind to antigens in the sensitive areas oftumour is highly dependent on the antibody characteristics and tumour and deposit most of the dose in those areasradionuclide properties, and can enhance therapeutic eYcacy (Pedley et al. 1990, Boxer et al. 1994) leading tothrough the selective dose delivery to the radiosensitive areas

enhanced cell sterilization. In contrast, non-speci� cof tumour.antibodies penetrate into the necrotic centre wheremost of the dose is deposited and largely wasted(Flynn et al. 1999a). Clearly, there is a strong relation-ship between antibody distribution, the heterogeneity

*Author for correspondence; e-mail: a.� ynn@ucl.ac.uk of dose deposition and the eVect in tumour. There-†CRC Targeting and Imaging Group, Department of Clinical fore, the meaningful interpretation of the observed

Oncology, and ‡Department of Primary Care and Population biological eVect and the optimization of therapySciences, Royal Free and University College Medical School,requires an understanding of the factors thatRoyal Free Campus, University College London, London

NW3 2PF, UK. in� uence the heterogeneity of dose deposition.

International Journal of Radiation Biology ISSN 0955-3002 print/ISSN 1362-3095 online © 2002 Taylor & Francis Ltdhttp://www.tandf.co.uk/journals

DOI: 10.1080/0955300011011781 8

Int J

Rad

iat B

iol D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y M

ichi

gan

Uni

vers

ity o

n 10

/31/

14Fo

r pe

rson

al u

se o

nly.

408 A. A. Flynn et al.

One of the most practical ways of in� uencing the the F(ab’)2 and Fab fragments by enzymatic digestionheterogeneity of dose deposition is by the manipula- (Casey et al. 1996) . Divalent (DFM) and trivalenttion of the properties of the antibody. Four main (TFM) antibodies were produced by chemically cross-antibody characteristics are likely to in� uence the linking the Fab fragments of A5B7 using a maleimidedistribution in tumour: molecular weight, aYnity, linker (Casey et al. 1996) . MFE-23 is a geneticallyavidity and speci� city. Speci� city, aYnity and avidity engineered single-chain Fv (scFv) antibody to CEAdescribe the interaction between the antibody and produced by phage technology (Chester et al. 1994)the target antigen. An antibody is speci� c if it and consisting of the variable light and heavy chainrecognizes an epitope on the antigen; aYnity joined by a synthetic linker. It is the smallest antibodydescribes the strength of this interaction; avidity is fragment that retains a full antigen-binding capacity.also referred to as functional aYnity as it also takes The control antibodies were MOPC (Schultes et al.into account the valency of the interaction. 1999) , a non-CEA binding monoclonal IgG, and

The molecular weight of an antibody restricts its NFE, a non-CEA-binding scFv created by insertingmobility within the tumour. Small antibodies can a point mutation into the antigen recognition sitepenetrate further into the tumour mass compared in MFE-23 (Read et al. 1997). The characteristicswith large antibodies, which remain near the tumour of these antibodies are summarized in table 1.vasculature (Yokota et al. 1992) . Tumour-speci� c anti- Puri� cation of all antibodies, except MFE-23 andbodies preferentially localize in viable areas, while NFE, was carried out by aYnity chromatographynon-speci� c antibodies penetrate into the necrotic using Protein A followed by gel � ltration. MFE-23centre (Flynn et al. 1999a). Antibodies with increased and NFE were puri� ed by immobilized metal ionaYnity and avidity have shown prolonged retention chromatography (IMAC). Purity and molecularin tumour (King et al. 1994, Adams et al. 1998), weight was con� rmed using SDS-PAGE along withprobably due to speci� c binding in viable areas. How- molecular weight markers.ever, due to the interrelationship between antibodycharacteristics, it has proved diYcult to observe theeVect of each characteristic in isolation. Consequently,there are few data on how each antibody characteristic 2.2. Antibody radiolabellingin� uences the degree of heterogeneity of dose depos-

Each antibody was labelled with 125I to a speci� cition in tumour. Heterogeneity of dose deposition inactivity of 60 kBq mg Õ 1 . All antibodies were labelledtumour is also dependent on the properties of theusing the chloramine-T method (Greenwood andradionuclide. Radionuclides with long-range emissionsHunter 1963) , except for MFE-23 and NFE, whichgive a more uniform pattern of dose deposition (Flynnwere labelled using the iodogen method (Fraker andet al. 1991), but a considerable amount of the energySpeck 1978) . The radiolabelled antibodies were steril-can escape the tumour (Bardies and Chatel 1994). Inized by passing them through a 0.22 mm acrodisccontrast, short-range emitters give a relatively higher� lter (Gelman Sciences). The percentage incorpora-local dose that may be greater than is required fortion of the isotope was analysed by thin-layer chroma-cell sterilization (Humm and Cobb 1990). Again, it istography, and the immunoreactivity of the labellednecessary to understand the factors that in� uenceproduct was checked by applying a dilution to aantibody distribution in tumour before the bestCEA-aYnity column and measuring the percentageradionuclide can be selected.bound.The present paper aims to assess the in� uence of

antibody characteristics and radionuclide propertieson the dose deposition in tumour. It is proposed thatthere may be a signi� cant therapeutic advantage

Table 1. Antibodies and their characteristics.with antibodies that are retained in viable radiosensit-ive areas of tumour. Furthermore, the impact of the Molecular weightchoice of the radionuclide is assessed based on the Antibody Speci� city Valency (kDa)properties of 131I and 90Y. Matching the antibody

A5B7-IgG 1 2 150with the radionuclide may further enhance therapy.A5B7-F(ab’)2 1 2 100A5B7-Fab 1 1 50

2. Materials and methods MFE-23 1 1 27TFM 1 3 150

2.1. Antibodies DFM 1 2 100MOPC Õ 0 150The monoclonal antibody A5B7 was raised against NFE Õ 0 27

CEA (Harwood et al. 1986) and was used to produce

Int J

Rad

iat B

iol D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y M

ichi

gan

Uni

vers

ity o

n 10

/31/

14Fo

r pe

rson

al u

se o

nly.

409Antibody and radionuclide properties can improve RIT

2.3. Animal studies ANOVA. This was carried out by comparing thevariation due to ‘antibody’ and ‘time’ with the vari-The CEA-expressing human colorectal carcinoma ation of viable:necrotic ratio for all antibodies.cell line LS174T (Tom et al. 1976) was grown as a Variation due to antibody was subdivided into thexenograft model in the � anks of nude mice weighing variation due to molecular weight, avidity and aYn-approximately 20–25 g. Each radiolabelled antibody ity. This was carried out using the general linearwas administered via the tail vein when the tumours model in Minitab.reached 0.5–1.0 cm3. Groups of four mice were culled

as selected time points after administration and tumoursamples were taken. Mice were given food and water 2.6. Dosimetryad libitum, the water containing 0.1% potassium iodide The activity distribution was represented by ato block the thyroid uptake of iodine. spherical tumour model (Flynn et al. 2001a) compris-

ing a viable rim surrounding a central necrotic core.The total activity (A(t )) at time t in the tumour was2.4. Image acquisitiongiven by:

Tumours were � xed in 10% neutral formalin forA (t )/g 5 A0 e Õ lt (%ID(t )/g )/100, (1)48 h, embedded in paraYn wax and sectioned at

3 mm. Multiple sections were cut, dewaxed in inhibi- where A0 was the injected activity and was set to givesol and exposed on Molecular Dynamics (MD, UK) a � xed level of recoverable damage to marrow (Flynnphosphor plates. Intimate contact was achieved by et al. 2001b) and l was the radionuclide decayusing MD Exposure Cassettes. The latent images constant. The percentage of the injected antibodyformed were converted to quantitative digital images dose per g tumour (%ID(t )/g) was taken from ausing an MD 425 PhosphorImager before and after previous study of antibody biodistribution atuse; the remaining images and background noise selected time points ranging from 3 to 144 h afterwere erased from storage phosphor plates using the administration (Flynn et al. 2001b). Two separateMD Image Eraser. After scanning, all tissue sections activity distributions were considered: uniform andwere stained with haematoxylin and eosin to compare heterogeneous. In the uniform case, the activity perthe radiolabelled antibody distribution with tissue g was equal in both the viable and hypoxic areas. Inmorphology. The stained sections were then digitized the heterogeneous case, redistribution of the antibodyusing a HP desk scanner (Hewlett Packard, Palo in the viable and necrotic regions was considered:Alto, CA, USA). the activity per g in the viable region relative to that

in the necrosis was equal to the ratio given by theradioluminographs. In both cases, the dose-rate in2.5. Data analysis the viable and necrotic areas of tumour was thenestimated based on the properties of 131I and 90YImages were analysed using MD ImageQuant

and Interactive Data Language software. Each radio- using a dosimetry model (Flynn et al. 2001a) wherethe tumour was represented by a spherical modelluminograph (RLG) and corresponding stained

histological section was registered using the cross- with a necrotic core and a viable shell. The dose-rate in the viable and necrotic compartments wascorrelation method (Flynn et al. 1999b). This allowed

accurate delineation of histological features by calculated using the self-dose fraction in each com-partment and the cross-dose fraction betweende� ning regions of interest (ROI).

Histological examination of the tumour sections compartments. The total dose in the viable andnecrotic areas was given by the area under the dose-showed that each section could be easily delineated

into two major zones. The � rst contained mostly rate curve and estimated using the trapezium rule.After the last time point, the antibody was assumedviable tumour, the second mostly necrotic tumour.

Using images of the stained histological section and to clear as from the blood and the area under thecurve was given by the integral of the clearanceRLG, a ROI was drawn around these two areas on

the histological image (� gure 1). This ROI was then function (Flynn et al. 2001b). Statistical comparisonof dose estimates was performed using a paired t-test.copied into the same position on the RLG. When

there were separate necrotic or viable areas in atumour section, an ROI was drawn around each 3. Resultsarea and the pixel values were grouped for each zone.

The in� uence of the antibody characteristics on Radioluminography acquired the images of anti-body distribution in tumour sections at selected timethe antibody concentration in viable relative to

necrotic tumour was assessed using a multifactor points after administration of 125I-labelled antibodies.

Int J

Rad

iat B

iol D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y M

ichi

gan

Uni

vers

ity o

n 10

/31/

14Fo

r pe

rson

al u

se o

nly.

410 A. A. Flynn et al.

Figure 1. Registration of images of antibody distribution (A) and tumour morphology (B) enabled the accurate delineation of viable(V) and necrotic (N) areas of tumour. Regions drawn on the morphology image were copied onto the image of antibodydistribution.

Registered images of tumour morphology and anti- necrotic areas of tumour, when the deposition isheterogeneous, for each antibody labelled with 131Ibody distribution measured the amount of antibody

in the viable relative to the necrotic areas of tumour and 90Y. There was no signi� cant diVerence betweenthe doses delivered by antibodies along with 131I and(� gure 1). Figure 2 shows the distributions of a multi-

valent speci� c (A(ii)) and non-speci� c (B(ii)) antibody 90Y ( p 5 0.56) when assuming a uniform dose. Whenaccounting for heterogeneity, there was evidence ofrelative to tumour morphology (A(i), B(i)). The

speci� c antibody preferentially localized in the viable, a diVerence in dose to viable ( p 5 0.05) and necroticareas ( p 5 0.03) of tumour with 131I comparedradiosensitive areas where most of the dose was

delivered, while the non-speci� c antibody penetrated with 90Y.A signi� cantly greater dose was deposited in theinto the necrosis. Also shown is the dose-rate distribu-

tion generated from b-point dose kernels for 131I viable areas using speci� c antibodies ( p 5 0.001) whilenon-speci� c antibodies delivered a higher dose to the(A(iii), B(iii)) and 90Y (A(iv), B(iv)) for the same two

antibodies. This was in� uenced by the range of the necrosis ( p 5 0.08). Consequently, the actual dosedelivered to the viable areas, by speci� c antibodies,b emission, with longer ranges giving a more uniform

instantaneous dose-rate. was signi� cantly greater than the estimated dosewhen assuming a uniform distribution ( p 5 0.03).Table 2 shows the viable:necrotic ratio for all

antibodies. The highest ratio was obtained when The highest doses to the viable areas were deliveredby the multivalent antibodies DFM, A5B7-F(ab’)2TFM was used. In fact, TFM gave the highest ratios

until the � nal timepoint. The smaller monovalent and TFM. Indeed, the dose to the viable tumourareas was more than � ve times greater for DFManti-CEA antibodies penetrated quickly into the nec-

rosis and, therefore, had low ratios. The non-speci� c compared with MFE-23 and more than 16 timesgreater than MOPC.antibodies were not retained in the viable areas,

penetrated quickly into the necrosis and, therefore,had a ratio <1. 4. DiscussionThe analysis of variance regarding time and anti-body characteristics against the viable:necrotic ratio It has been shown that radioluminography is a

powerful technique for quantifying the variation ofshowed there was a signi� cant change in the ratioover time ( p<0.0001) . Similarly, a change in anti- antibody distribution in tumour over time. For this

purpose, we simpli� ed the tumour structure by de� n-body aYnity ( p 5 0.0005) and valency ( p 5 0.032) hada signi� cant eVect on the retention in viable areas. ing two major regions, each distinctive in its morpho-

logical structure and response to radiation. The � rstA change in molecular weight had no signi� canteVect ( p 5 0.807). area contained radiosensitive viable tumour cells, the

second mostly necrotic cells, although hypoxic cellsTable 3 shows the absorbed dose in tumour, whenthe deposition is uniform, and in the viable and may also persist here. It has been shown that the

Int J

Rad

iat B

iol D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y M

ichi

gan

Uni

vers

ity o

n 10

/31/

14Fo

r pe

rson

al u

se o

nly.

411Antibody and radionuclide properties can improve RIT

Figure 2. Typical distribution in tumour from speci� c A(ii)) and non-speci� c (B(ii)) antibodies compared with the morphology(A(i), B(i)). Also shown is the dose distribution from 131I (A(iii), B(iii)) and 90Y (A(iv), B(iv)).

dose delivered to these areas is highly dependent on with the highest viable:necrotic ratios, with the biva-lent antibodies scoring the next highest, and thethe antibody and radionuclide used.

Eight antibodies were used that were speci� c to lowest being MFE-23, which penetrated quickly intothe necrosis, as reported with another scFv (YokotaCEA, and two which were non-speci� c. Of the anti-

CEA antibodies, the trivalent TFM was associated et al. 1992) . Given that the activity in the viable

Int J

Rad

iat B

iol D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y M

ichi

gan

Uni

vers

ity o

n 10

/31/

14Fo

r pe

rson

al u

se o

nly.

412 A. A. Flynn et al.

Table 2. Viable:necrosis ratios for each antibody at selected time points.

Time after administration (h)

Antibody 1 3 6 24 48 72

A5B7-IgG – 3.37 Ô 1.07 – 2.25 Ô 0.52 0.92 Ô 0.32 0.88 Ô 0.25A5B7-F(ab’)2 – 3.74 Ô 0.82 – 1.95 Ô 0.29 1.9 Ô 1.38 0.58 Ô 0.17A5B7-Fab – 1.66 Ô 1.07 – 1.1 Ô 0.27 0.67 Ô 0.09 0.59 Ô 0.13MFE-23 1.38 Ô 0.52 1.58 Ô 0.83 1.12 Ô 0.56 0.72 Ô 0.98 – –TFM – 4.11 Ô 2.05 – 3.25 Ô 2.01 2.66 Ô 2.47 0.66 Ô 0.27DFM 2.7 Ô 1.37 3.52 Ô 1.65 – 2.92 Ô 0.52 1.83 Ô 0.42 0.54 Ô 0.24MOPC – 0.78 Ô 0.04 – 0.68 Ô 0.12 0.55 Ô 0.05 0.6 Ô 0.14NFE 0.59 Ô 0.07 0.44 Ô 0.16 – – – –

Table 3. Absorbed dose (Gy) in tumour when the dose deposition is uniform and in viable and necrotic areas when accounting forheterogeneity for antibodies labelled with 131I and 90Y. Also shown is the ratio of dose in viable relative to necrosis.

131I 90Y

Mean Viable Necrotic Mean Viable NecroticAntibody dose tumour tumour Ratio dose tumour tumour Ratio

A5B7-IgG 34.60 35.54 33.81 1.05 27.86 29.58 26.62 1.11A5B7-F(ab’)2 50.56 70.25 47.34 1.48 47.26 57.72 42.56 1.36A5B7-Fab 21.48 24.87 15.32 1.62 22.66 23.46 14.63 1.60MFE-23 13.14 16.58 13.78 1.20 15.59 17.06 14.71 1.16TFM 34.33 53.1 28.57 1.86 32.44 43.36 28.77 1.51DFM 61.78 89.33 54.72 1.63 58.37 73.39 51.78 1.42MOPC 6.44 5.53 9.03 0.61 7.16 5.84 7.45 0.78NFE 15.62 11.20 23.95 0.47 17.87 12.56 19.29 0.65

region for the speci� c antibodies is 1.6–4.1 times lower pressure than the viable areas. This appears tocontradict some evidence of raised interstitial pres-that in the necrosis at 3 h, the radiosensitive viable

region of the tumour receives a signi� cantly larger sure at the centre of tumours ( Jain and Baxter 1988)but agrees with the predictions of a mathematicaldose-rate. For the non-speci� c antibodies, there was

1.3–2.3 times the activity in the necrosis compared model describing tumour growth that shows a reduc-tion in interstitial pressure in necrosis (Please et al.with the viable region, demonstrating a rapid penet-

ration into the tumour. This shows that a larger dose 1998) compared with the viable areas.The temporal change in the concentration gradientis delivered to the necrotic region of the tumour

where it is probably wasted. also aVects the viable:necrotic ratio. As the gradientfalls and reverses, the ratio also falls to <1. However,All antibodies showed some degree of penetration

into the tumour. Antibodies are forced towards the higher ratios may be maintained if higher-aYnityand -avidity antibodies are used. Higher-aYnity anti-centre of the tumour by a concentration gradient

between the circulating antibody and the tumour. bodies have a stronger interaction with the targetantigen, while the avidity eVect has a number ofIndeed, it is this gradient that is used in mathematical

models to explain the transport of antibodies from possible explanations. The probability of an inter-action is increased and the binding is stronger ifthe blood into tumours ( Jain 1987, Strand et al. 1993,

Baxter et al. 1994) . The magnitude of this gradient more than one arm is bound. Alternatively, a mono-valent antibody bound to an antigen that has beenchanges with time. Initially, the gradient is high

when the antibody concentration in the blood is shed from the cell surface eVectively behaves as anon-speci� c protein, while multivalent antibodies stillhigh, and antibodies penetrate into the tumour. As

the antibody clears from the circulation, the gradient have the ability to bind to � xed antigen. Nevertheless,aYnity and avidity restrict antibody movement awaydecreases and there is a reduced rate of tumour

loading. Once the gradient has reversed, the amount from the viable tumour, which could potentially leadto an enhanced biological eVect. In contrast, molecu-of antibody in the viable areas reduces. However,

antibodies that penetrated into the necrosis seem to lar weight did not signi� cantly restrict the mobilityof antibodies in tumour.be trapped, which suggests that necrosis may be at a

Int J

Rad

iat B

iol D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y M

ichi

gan

Uni

vers

ity o

n 10

/31/

14Fo

r pe

rson

al u

se o

nly.

413Antibody and radionuclide properties can improve RIT

There was no signi� cant diVerence in the amount distribution in the viable and necrotic areas of 90Y-and 131I-labelled antibodies is likely to be small andof IgG in the viable and necrotic regions at late

timepoints. This was primarily consistent with the have little in� uence on the pattern of the dosedeposition in tumour.hypothesis that the long circulating half-life of the

antibody maintained the concentration gradient and The eVect that the heterogeneous nature of tumourpathophysiology has on systemically delivered thera-forced the antibody into the tumour when the antigen

sites adjacent to blood vessels were occupied. This peutics has largely been ignored, since detailedinformation cannot be readily obtained. It has beensupports the idea that higher amounts of protein may

prolong the positive gradient, improve localization shown that this heterogeneity can make a signi� cantdiVerence to the dose estimates and depends on(Thomas et al. 1989) and lead to a more uniform

tumour dose (Sgouros 1992). many factors. Antibody aYnity and valency restrictpenetration into the tumour but facilitate theThe multivalent antibodies DFM, A5B7-F(ab’)2

and TFM labelled with 131I delivered the highest improved retention in viable areas. An increase inthe amount of antibody administered and in thedose to the viable tumour cells, while the non-speci� c

antibody MOPC labelled with 90Y delivered the clearance half-life can also in� uence the heterogen-eity but would increase normal tissue toxicity.lowest dose. Indeed, viable tumour cells received a

dose that was >15 times greater with DFM-131I Furthermore, radionuclides can also enhance theeVect of heterogeneity and should be selected tocompared with MOPC-90Y. For speci� c antibodies,

the dose to the viable tumour cells was � ve times match the antibody kinetics.This paper gives important information on thegreater for DFM compared with MFE-23, illustrating

the scale of the eVect produced by optimizing the heterogeneous distribution of antibodies in tumoursand helps to explain the variation of reportedchoice of antibody and radionuclide.

Antibodies labelled with 131I performed better than responses in RIT. The complexity of tumour biologyalso has important implications for other therapeuticthose with 90Y. Primarily, this was due to the fact

that a greater activity of 131I can be injected before strategies and can cause a biological eVect that islargely unrelated to that produced under conditionsdelivering an equal dose to the blood. In addition,

the fraction of the total energy emitted from 90Y of uniformity. The eYcacy of conventional chemo-therapeutics is likely to be in� uenced by both hetero-absorbed in the tumours used in this study was much

less than for 131I. However, it is likely that 90Y would geneous delivery and response. External beamradiotherapy may also bene� t from setting the beamperform comparatively better in larger tumours and

may be more eVective in a clinical setting. However, conformation so that the dose distribution matchesthe pattern of viable tumour areas rather thanmore information is required on the antibody kinetics

in human tumours to address this issue. treating the tumour as a whole.Furthermore, the temporal change in the distribu-Although 90Y gives a more uniform dose-rate

distribution due to its longer range, some antibodies tion is of great importance to the outcome of multi-phase strategies where the eVective timing ofwith 90Y delivered a higher dose to the viable cells

relative to the necrotic cells. Primarily, this was due subsequent phases is dependent on there being afavourable distribution of the previous phase.to the shorter half-life. Initially, viable cells receive a

high dose when the antibody is preferentially local- Treatments involving repeated RIT would bene� tgreatly from knowledge about the temporal changeized in the viable area, then as antibody clears from

the viable areas and is retained in the necrosis, the in dose deposition in tumour. Likewise, when RIT isused in combination with external beam radiation,necrosis gives a higher proportion of the dose. In the

case of 90Y, however, this proportion is reduced due chemotherapeutics or antivascular agents. Othertumour-targeted strategies such as antibody-directedto lower activity caused by decay.

Dosimetry calculations for each antibody and enzyme prodrug therapy (ADEPT) (Bhatia et al.2000) would also pro� t from this information as theradionuclide were based on data from antibodies

labelled with 125I. Although the kinetics of 125I- prodrug should be administered at the time whenthe antibody–enzyme conjugate is localized in viablelabelled antibodies are likely to be similar to those

labelled with 131I, this may not be the case for 90Y. tumour areas to achieve maximum therapy.In conclusion, the extent of the heterogeneity ofIndeed, it has been shown that diVerences occur

between antibodies labelled with 131I and those dose deposition in tumour is dependent on thecharacteristics of the antibody and radionuclide. Thislabelled with 90Y at the major sites of catabolism

such as the kidney, liver and spleen (Casey et al. information gives a conceptual basis for consideringdosimetry and could lead to improved dose estimates.1996) . However, no diVerence was observed in

tumour. Consequently, any diVerence in the Accounting for the eVect of dose heterogeneity in

Int J

Rad

iat B

iol D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y M

ichi

gan

Uni

vers

ity o

n 10

/31/

14Fo

r pe

rson

al u

se o

nly.

414 A. A. Flynn et al.

tal cancer xenograft model. International Journal of Radiationrelation to the spatial heterogeneity of tumourOncology Biology and Physics, 43, 183–189.response may enhance the understanding of the

Flynn, A. A., Green, A. J., Boxer, G., P edley, R. B. andrelationship between the absorbed dose and the Begent, R. H. J., 1999b, A comparison of image registra-eVect and enable the optimization of therapy by tion techniques for the correlation of radiolabelled anti-

body distribution with tumour morphology. Physics inselecting the most eVective antibody–radionuclideMedicine and Biology, 44, N151–159.combinations.

Flynn, A. A., Green, A. J., P edley, R. B., Boxer, G. M.,Boden, R. and Begent, R. H. J., 2001a, A mouse modelfor calculating the absorbed beta dose, from 131I and 90YAcknowledgementslabelled immunoconjugates, including a method fordealing with heterogeneity in kidney and tumour. RadiationThe work was supported by an MRC-LINK AgentResearch, 156, 28–35.Grant with Celltech Ltd, the Cancer Research

Flynn, A. A., P edley, R. B., Green, A. J., Boxer, G. M.,Campaign, the Trusthouse Charitable Foundation Boden, R., Dearling, J., Bhatia, J. and Begent,and the Ronald Raven Chair in Clinical Oncology R. H. J., 2001b, The eVectiveness of radiolabelled anti-

bodies for radioimmunotherapy in a colorectal xenograftTrust.model: a comparative study using the linear–quadraticformulation. International Journal of Radiation Biology, 77,507–517.References

Fraker, P . J. and Speck, J. C., 1978, Protein and cell membraneiodinations with a sparingly soluble chloramide,Adams, G. P ., Schier, R., Marshall, K., Wolf, E. J., McCall,

A. M., Marks, J. D. and Weiner , L. M., 1998, Increased 1,3,6,6-tetrachloro-3a , 6a-diphenylglycoluril. Biochemicaland Biophysical Resaearch Communications, 80, 849–857.aYnity leads to improved selective tumor delivery of

single-chain Fv antibodies. Cancer Research, 58, 485–490. Greenwood, F. C. and Hunter, W. M., 1963, The preparationof 131I-labelled human growth hormone of high speci� cBardies, M. and Chatal, J.-F., 1994, Absorbed doses for

internal radiotherapy from 22 beta-emitting radionuc- radioactivity. Biochemistry Journal, 89, 116–123.Harwood, P . J., Britton, D. W., Southall, P . J., Boxer,lides: beta dosimetry of small spheres. Physics in Medicine

and Biology, 39, 961–981. G. M., Rawlins, G. and Rogers, G. T., 1986, Mappingepitope characteristics on carcinoembryonic antigen.Baxter , L. T ., Zhu, H., Mackensen, D. G. and Jain, R. K.,

1994, Physiologically based pharmacokinetic model for British Journal of Cancer, 54, 75–82.Humm, J. L. and Cobb, L. M., 1990, Nonuniformity of tumorspeci� c and nonspeci� c monoclonal antibodies and frag-

ments in normal tissues and human tumor xenografts in dose in radioimmunotherapy. Journal of Nuclear Medicine,31, 75–83.nude mice. Caner Research, 54, 1517–1528.

Bhatia, J., Sharma, S. K., Chester, K. A., P edley, R. B., Jain, R. K., 1994, Barriers to drug delivery in solid tumors.Scienti�c American, 271, 58–65.Boden, R. W., Read, D. A., Boxer, G. M., Michael,

N. P . and Begent, R. H. J., 2000, Catalytic activity of Jain, R. K., 1987, Transport of molecules in the tumor interstit-ium: a review. Cancer Research, 47, 3039–3051.an in vivo tumor targeted anti-CEA scFv::carboxypeptide

G2 fusion protein. International Journal of Cancer, 85, Jain, R. K. and Baxter , L. T., 1988, Mechanisms of heterogen-eous distribution of monoclonal antibodies in tumors:571–577.

Boxer, G. M., Abassi, A. M., P edley, R. B. and Begent, signi� cance of elevated interstitial pressure. Cancer Research,47, 7011–7032.R. H. J., 1994, Localisation of monoclonal antibodies

reacting with diVerent epitopes on carcinoembryonic King, D. J., Turner, A., Farnsworth, A. P . H., Adair, J. R.,Owens, R. J., P edley, R. B., Baldock, D., P roudfoot,antigen (CEA)-implications for targeted therapy. British

Journal of Cancer, 69, 307–314. K. A., Lawson, A. D. G., Beeley, N. R. A., Millar,K., Millican, A., Boyce, B. A., Antoniw, P .,Casey, J. L., King, D. J., Chaplin, L. C., Haines, A. M. R.,

P edley, R. B., Mountain, A., Yarranton, G. T. and Mountain, A., Begent, R. H. J., Schochat, D. andYarranton, G. T., 1994, Improved tumor targeting withBegent, R. H. J., 1996, Preparation, characterisation

and tumour targeting of cross-linked divalent and trivalent chemically cross-linked recombinant antibody fragments.Cancer Research, 54, 6176–6185.anti-tumour Fab’ fragments. British Journal of Cancer, 74,

1397–1405. Koral, K. F., Dewaraja, Y., Clarke, A., Li, J., Zasadny, R.,Rommelfanger, S. G., Francis, I. R., Kaminski, M. S.Chester, K. A., Begent, R. H. J., Robson, L., Keep, P .,

P edley, R. B., Boden, J. A., Boxer, G., Green, A., and Wahl, R. L., 2000, Tumor-absorbed-dose estimatesversus response in tositumab therapy of previouslyWinter, G., Cochet, O. and Hawkins, R. E., 1994,

Phage libraries for generation of clinically useful antibod- untreated patients with follicular non-Hodgkin’slymphoma: preliminary report. Cancer Biotherapy andies. Lancet, 343, 455–456.

DeNardo, G. L., DeNardo, S. J., Shen, S., DeNardo, D. A., Radiopharmaceuticals, 15, 347–355.Lane, D. M., Eagle, K. F., Begent, R. H. J., Hope-Stone,Mirick, G. R., Macey, D. J. and Lamborn, K. R., 1999,

Factors aVecting 131-I-Lynn-1 pharmacokinetics and L. D., Green, A. J., Casey, J. L., Keep, P . A., Kelly,A. M. B., Ledermann, J. A., Glaser, M. G. and Hilson,radiation dosimetry in patients with non-Hodgkin’s

lymphoma and chronic lymphocytic leukemia. Journal of A. J. W., 1994, Radioimmunotherapy of metastatic col-orectal tumours with iodine-131-labelled antibody to car-Nuclear Medicine, 40, 1317–1326.

Flynn, A. A., Green, A. J., Boxer, G. M., Casey, J. L., P edley, cinoembryonic antigen: phase I/II study withcomparative biodistribution of intact and F(ab’)2 antibod-R. B. and Begent, R. H. J., 1999a, A novel technique,

using radioluminography, for the measurement of uni- ies. British Journal of Cancer, 70, 521–525.Matthay, K. K., Huberty, J. P ., Hattner, R. S. and Ablin,formity of radiolabeled antibody distribution in a colorec-

Int J

Rad

iat B

iol D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y M

ichi

gan

Uni

vers

ity o

n 10

/31/

14Fo

r pe

rson

al u

se o

nly.

415Antibody and radionuclide properties can improve RIT

A. R., 1991, EYcacy and safety of 131-I metaiodobenzyl- Madiyalakan, R., 1999, Immunotherapy of humanovarian carcinoma with OVARIXT M Mab-B43.13 in aguanaide therapy for patients with refractory neuroblas-

toma. Journal of Nuclear Biology and Medicine, 35, 244–247. human-PBL-SCID/BG mouse model. Hybridoma, 18,47–55.P edley, R. B., Boxer, G. M., Boden, J. A., Southall, P . J.,

Begent, R. H. J., Bagshawe, K. D., Humm, J. and Sgouros, G., 1992, Plasmapheresis in radioimmunotherapy ofmicrotastases: a mathematical modeling and dosimetricalSearle, F., 1990, Preliminary observations on the micro-

distribution of labelled antibodies in human colonic aden- analysis. Journal of Nuclear Medicine, 33, 2167–2179.Strand, S.-E., Zanconico, P . and Johnson, T. K., 1993,ocarcinoma xenografts: relevance to microdosimetry.

British Journal of Cancer, 61, 218–220. Pharmacokinetyic modeling. Medical Physics, 20, 515–527.Thomas, G. D., Chappell, M. J., Dykes, P . W., Ramsden,P lease, C. P ., P ettet, G. and McElwain, D. L. S., 1998, A

new approach to modelling the formation of necrotic D. B., Godfrey, K. R., Ellis, J. R. M. and Bradwell ,A. R., 1989, EVect of dose, molecular size, aYnity, andregions in tumors. Applied Mathematical Letters, 11, 89–94.

P ress, O . W., Eary, J. F., Appelbaum, F. R., Martin, P . J., protein binding on tumor uptake of antibody of ligand:a biomathematical model. Cancer Reserach, 49, 3290–3296.Nelp, W. B., Glenn, S., Fisher, D. R., P orter, B.,

Matthews, D. C. and Gooley, T., 1995, Phase II trial Tom, B. H., Rutzky, L. P ., Jakstys, M. M., Oyasu, R., Kaye,C. I. and Kahan, B. D., 1976, Human colonic adenocarci-of 131-B1 (anti-CD20) antibody therapy with autologous

stem cell transplantation for relapsed B cell lymphomas. noma cells. In Vitro, 12, 180–191.Wessels, B. W., Vessella, R. L., P alme, D. F., Berkopec,Lancet, 346, 336–340.

Read, D. A., Cooper, M. S., Boden, R., Boden, J. A., Chester, J. M., Smith, G. K. and Bradley, E. W., 1989,Radiobiological compatison of external beam irradiationK. A. and Begent, R. H. J., 1997, 99mTc labelling of

MFE-23 (an anti-CEA single chain antibody). British and radioimmunotherapy in renal cell carcinoma xeno-grafts. International Journal of Radiation Oncology Biology andJournal of Cancer, 75, 34. [abst].

Roberson, P . L. and Buchsbaum, D. J., 1995, Reconciliation of Physics, 17, 1257–1263.Yokota, T ., Milenic, D. E., Whitlow, M. and Schlom, J.,tumor response to external beam radiotherapy versus

radioimmunotherapy with 131Iodine-labeled antibody for 1992, Rapid tumor penetration of a single-chain Fv andcomparison with other immunoglobulin forms. Cancera colon cancer model. Cancer Research, 55, 5811–5816.

Schultes, B. C., Zhang, C. S., Xue, L. Y., Noujaim, A. A. and Research, 52, 3402–3408.

Int J

Rad

iat B

iol D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y M

ichi

gan

Uni

vers

ity o

n 10

/31/

14Fo

r pe

rson

al u

se o

nly.