Timing ofmeasurements

Timing of measurementsIntravenous administration* Peak-15 mins after the end of theinfusion* Trough-just before the next dose

Intramuscular administration* Peak-1 hour after the dose* Trough-just before the next dose

Blood samples for serum aminoglycoside measurements are generallytaken immediately before a dose (the trough concentration) and soon afterthe next dose (the peak concentration). Although this is the order in whichthe samples are taken, it is generally assumed for dosage calculations thatthe trough concentration is that measured after the peak rather than beforeit.The time of sampling for the peak concentration depends on the route of

administration. If the aminoglycoside is given by intravenous infusionsamples should be taken 15 minutes after the end of the infusion. If theintramuscular route is chosen they should be taken one hour after theinjection.

The sources of data presented in the graphs are: B M Orme and R E Cutler, Clin Pharmacol Ther 1969;10:543-50 for creatinine clearance v kanamycin clearance; R D Moore, P S Lietman, andC R Smith,J7Infect Dis 1987;155:93-9 for relationship between peak concentration: MIC andrate of response; andM Wenk, S Vozeh, and F Follath, Clin Pharmacokinetics 1984; 9:475-92 fortissue accumulation of aminoglycosides. The data are reproduced with permission of the journals.

Dr J K Aronson is clinical reader in clinical pharmacology and Dr D JM Reynolds is clinicallecturer in clinical pharmacology, Radcliffe Infirmary, Oxford.

MRC Centre andAddenbrooke's Hospital,Cambridge CB2 2QHStephen J Russell, clinicianscientistMeirion B Llewelyn, MRCtrainingfellowRobert E Hawkins, CRCsenior clinical research fellow

BMJ 1992;305: 1424-9

This is the third of three articlesexamining the developmentand clinical application ofmonoclonal antibodies

The success of monoclonal antibodies in clinicalpractice is dependent on good design. Finding asuitable target is the most important part as otherproperties of the antibody can be altered by geneticengineering. Antibodies that target lymphocyteantigens offer less toxic immunosuppressive treat-ment than currently available drugs and the firstmonoclonal antibody approved for human use is animmunosuppressive agent for treating rejection ofrenal transplants. Human trials of monoclonal anti-bodies to treat septic shock have been done andantibodies are also being developed to target commonpathogens such as herpes simplex virus. Althoughmonoclonal antibodies against cancer have beenmuch heralded, their success has been limited by thepoor access to the inside of tumours. Treatmnent ofblood cancers has been more successful and ahuman antibody against B cell malignancies is beingclinically tested. As knowledge about natural immuneresponses and antibody engineering increases manymore monoclonals are likely to feature in clinicalpractice.

The 1990s will be a testing time for monoclonalantibodies. Potential clinical applications include thetreatment of cancer, autoimmune disease, transplantrejection, viral infection, and toxic shock. The Centrefor Exploitation of Science and Technology has esti-mated that the total world market for monoclonalantibodies will reach $1000 million by 1994, rising to$6000 million by the year 2000.' It remains to be seenwhether the clinical promise of monoclonal antibodieswill be realised on such a grand scale, but antibodytherapy is likely to be much in evidence in manyclinical settings over the next few years. Clinicians willtherefore need to familiarise themselves with some ofthe issues relating to use of clinical antibodies.As described in the previous article in this series,

monoclonal antibodies can be generated against most

Principles ofantibody therapy

Stephen J Russell, Meirion B Llewelyn, Robert E Hawkins

MonoclonalAntibodies in Medicine

target antigens, purified, split into fragments, andconjugated to radionuclides, toxins, enzymes, ordrugs.2 The genes can be cloned and reconstructed togive new versions of the antibody with decreasedimmunogenicity, improved affinity, reduced size, ornovel effector domains. For a given clinical applicationthe first task is to choose an appropriate target antigenand then to optimise the therapeutic antibody generatedagainst the chosen target.

Rather than catalogue every antibody with clinicalpotential or every disease that has responded tomonoclonal antibodies, the aim of this review is tounderline the principles goveming antibody therapyand to illustrate these with specific examples.

Target antigensAntibodies can neutralise toxins; block the inter-

action of growth factors, hormones, intercellularadhesion molecules, or viruses with their cognatecellular receptors; and coat bacteria, viruses, or cells,marking them for phagocytosis, antibody dependentcellular cytotoxicity, or complement mediated lysis.Target antigens can therefore be circulating or on

the cell surface. Selecting a suitable target for a givendisease depends not only on the aims of treatment buton the precise tissue distribution of the target antigen,its function, and its fate after it has complexed with thetherapeutic monoclonal antibody. Finding a suitabletarget antigen is probably the most important factordetermining the ultimate success or failure of antibodytherapy. Provided the target has been well chosen itmay be possible to modify the corresponding mono-clonal antibody in various ways to enhance its thera-peutic potential.

PharmacokineticsInfused antibodies are diluted almost immediately in

the total plasma volume and then diffuse more slowly

1424 BMJ VOLUME 305 5 DECEMBER 1992

on 9 Decem

ber 2020 by guest. Protected by copyright.

http://ww

w.bm

j.com/

BM

J: first published as 10.1136/bmj.305.6866.1424 on 5 D

ecember 1992. D

ownloaded from

across the walls of small blood vessels into the inter-stitial fluid (distribution phase). The half life of thecirculating antibody is determined by the rate at whichit is metabolised and excreted (elimination phase). Thedegree to which the target antigen is bound varies withthe total time of exposure, the concentration, and thekinetic properties of the monoclonal antibody.

REACHING THE TARGET

How easily infused monoclonal antibody can reach itdepends on the target's location. Intravascular targetsare readily accessible, but other targets are less easilyreached because exit from the vascular system isrestricted. To gain access to extravascular targets (forexample, cancer cell surface antigens) the antibodymust pass through the endothelial lining of a capillaryor postcapillary venule. The macromolecular perme-ability of these vessels decreases progressively withincreasing molecular weight such that molecules largerthan 40 kDa (the molecular weight of IgG is 150 kDa)escape from the plasma only slowly. Smaller antibodyfragrnents, particularly Fv reagents (molecular weight25 kDa), penetrate the interstitial fluid space morereadily than whole IgG. High molecular weight proteinscan escape from the microvessels through gaps betweenadjacent endothelial cells, which are particularlyabundant in inflamed tissues. The discontinuousenidothelial lining of the sinusoidal circulations of liver,spleen, and bone marrow also allows free passage ofmacromolecules such as IgG.

BINDING REACTIION

The degree to which a therapeutic antibody binds toits target antigen is govemed by the concentration ofantigen, the concentration of antibody to which it isexposed, the duration of exposure, and the intrinsicproperties of antibody and antigen that determine theirrates of association and dissociation.

TIhe equation is simple when attempting, forexample, to neutralise a circulating toxin. From theknown concentration of toxin in plasma and the affinityof the antitoxin antibody the dose required for effectiveneutralisation can be calculated. For cell surfaceantigens the analysis is less simple. The concentrationof antibody to which the target cells are exposed andthe duration of exposure are determined by the rate atwhich the antibody enters the interstitial fluid and therate at which it is eliminated from the body. Further-more, in contrast with soluble target antigens, cellassociated targets are effectively multivalent so thataffinity is no longer the only factor determining the rateof dissociation of cell bound antibody. Dissociationslows greatly when an IgG molecule is anchored to thetarget cell surface through both of its antigen bindingsites rather than through a single site. Provided thetarget antigen is expressed at sufficiently high density,bivalent molecules such as intact IgG or F(ab)2 canbind with much greater avidity than can the smallerunivalent Fab and Fv antibody fragments. Complexedantibody may also be taken into the cytoplasm of thetarget cell, effectively preventing further dissociation.Additionally, a cluster of many cells displaying thesame target artigen (a tumour deposit, for example)may behave as an antigen "sink" from which theantibody escapes only very slowly. This is becausedissociation of a bound antibody molecule will befollowed immediately by rebinding to the same or aneighbouring cell.

ANTIBODY CLEARANCE

Ultimately, all infused antibody will be eliminatedfrom the body. The Fc portion of IgG is thought todetermine its catabolic rate, which (in humans) is fasterfor murine antibodies than for human antibodies.4Smaller antibody fragments pass relatively easily from

the glomerular capillaries into the renal tubules and arerapidly excreted unchanged in the urine; this greatlyshortens their plasma half life. Thus the circulatinghalf life of IgG is measured in days, and that of singlechain Fv fragments (scFv) in minutes, while F(ab)2and Fab fragments have intermediate half lives.3

Clearance of antibody which has been retained in thetissues is slower. Retention of antibodies by tissue dueto specific interaction of the antibody with its targetantigen is welcome, but non-specific binding tohomologous or non-homologous antigens also occurs.Moreover, IgG may be retained in liver, spleen, andbone marrow through the interaction of its Fc portionwith Fc receptors on resident macrophages. Fab andF(ab)2 fragments tend to accumulate in the kidneys.3Persistence of antibodies in normal host tissues may betroublesome, leading, for example, to excessive toxicityof a radiolabelled therapeutic cancer antibody.For certain applications prolonged or repeated

administration of monoclonal antibodies is needed toobtain real clinical benefit. But a host immune responseagainst the infused antibody can greatly shorten itscirculating half life.4 Murine and other non-humanantibodies and chimeric antibodies linked to non-human proteins are particularly prone to host re-sponses. For example, a strong human antimouseantibody response can sequester infused mouse anti-body to give circulating immune complexes that arerapidly phagocytosed by reticuloendothelial cells,reducing the potency of treatment. This problem canbe circumvented by using human or humanised anti-bodies, although the patient may still mount animmune response against idiotypic (antigen bindingsite) or allotypic (Fc) determinants of the therapeuticantibody. As a general rule it is preferable to usehuman or humanised antibodies to rodent antibodies.New methods for screening large human phage anti-body libraries will prove useful for this purpose.

Effector mechanismsWhen the goal of treatment is neutralisation of a

toxin or blockade of a ligand-receptor interaction thetherapeutic antibody requires no special effectordomain and, depending on the effective valency of thetarget antigen, should function well as a monovalent(single chain Fv or Fab) or bivalent (F(ab)2) fragment.More commonly, however, the aim is to destroy aspecific population of target cells. Phagocytosis, anti-body dependent cellular cytotoxicity, and complementfixation are the natural effector pathways activated bythe Fc portion of cell bound antibody. Murine anti-bodies do not recruit human effector functions well butthis can be completely remedied by chimerisation orby grafting on complementarity determining regions(CDR grafting) to humanise the antibody. Humaneffector functions are recruited more efficiently by ratantibodies of the IgG2b subclass than by murineantibodies.5 Smaller antibody fragments without an Fcportion (single chain Fv, Fab, F(ab)2) can be artificiallygiven altemative effector mechanisms including radio-active metals, plant and bacterial toxins, enzymes, andcytotoxic drugs.The most suitable effector mechanism depends on

several factors. A high density of IgG on the target cellis required to activate complement because it isinitiated by crosslinking the Fc portions of twoadjacent cell bound antibody molecules. Moreover, Fcmediated recruitment of phagocytes, antibodydependent cellular cytoxicity, and complement is notpossible unless bound antibody stays on the surface ofthe target cells.

For some conditions it may be more appropriate touse antibodies with artificially linked effector functions.Cells which rapidly intemalise bound antibody or

BMJ VOLUME 305 5 DECEMBER 1992 1425

on 9 Decem

ber 2020 by guest. Protected by copyright.

http://ww

w.bm

j.com/

BM

J: first published as 10.1136/bmj.305.6866.1424 on 5 D

ecember 1992. D

ownloaded from

which express the target antigen at low density may bekilled more effectively by antibodies conjugated todrugs, toxins, or radionuclides. Radioimmunocon-jugates also have the advantage that their radiation canpenetrate several cell diameters into the tissues-thismay be useful for cancer therapy as the monoclonalantibody cannot penetrate deep into the tumour. Thebox summarises the available antibody effectormechanisms.

Clinical use ofmonoclonal antibodiesIMMUNOSUPPRESSION

Immune responses to self or foreign antigens can

lead to autoimmune destruction of tissue or rejection oftransplanted organs. Immunosuppressive therapiesinclude corticosterioids, cyclosporin, cytotoxic drugs,and polyclonal antilymphocyte antisera, all of whichhave high toxicity and are sometimes ineffective.Monoclonal antibodies offer a realistic altemative tothese immunosuppressive drugs, and this is perhapstheir most useful current application. Potential targetsfor immunosuppressive monoclonal antibodies includelymphocyte differentiation antigens, cytokines,cytokine receptors, and cell adhesion molecules.The first monoclonal antibody to be approved for

human therapy (OKT3) is an immunosuppressivemurine reagent which binds to T lymphocytes and isuseful for treating rejection of renal transplants.7 Incommon with many other immunosuppressiveantilymphocyte monoclonal antibodies it does notstimulate a strong antimouse response. The toxicity ofOKT3 is worse with the first dose, which triggersrelease of cytokines from targeted cells and leads insome cases to hypotension, weight gain, and breath-lessness, progressing occasionally to pulmonaryoedema. Many other immunosuppressive monoclonalantibodies have been shown to have activity in humans.Among the most promising are antibodies against thelymphocyte antigens CD4, Tac, and CDw52 (seebelow), all of which have now been humanised by CDRgrafting,'0 and several monoclonal antibodies whichblock adhesion ofimmune and inflammatory cells.Monoclonal antibodies against CD4 inhibit the

function of helper T cells and have been used withvarying success to treat acute rejection of renal allo-grafts, rheumatoid arthritis, inflammatory boweldisease, systemic lupus erythematosus, psoriasis,relapsing polychondritis, systemic vasculitis, andmycosis fungoides."'' Tac monoclonal antibodies

recognise high affinity interleukin 2 receptors ofactivated lymphocytes and do not bind to restinglymphocytes. They can therefore block ongoing antigenspecific immune responses highly specifically withoutdamaging resting. lymphocytes. Murine Tac mono-clonal antibodies were shown to prevent early rejectionof renal allografts, but antimouse responses weredetected in 810/0 of patients after one month oftreatment."8 Hum.anised Tac a-ntibody (Tac-H) wasrecently compared with the murine antibody inprimates given cardiac allografts.'9 The humanisedantibody had a lon-ger circulating half life (103 v 38 h),was less immunogenic (0% v 100% antiantibodyresponses before day 33), and produced a longer graftsurvival than the rnurine antibody.Immunosuppressive monoclonal antibodies will

undoubtedly contribute to the therapeutic optionsagainst autoimmune disease and rejection of trans-plants but their precise role has yet to be defined. Moredetailed understanding of the underlying immuno-pathogenic mechanisms in many autoimmune con-ditions and vasculitic states will help future explorationof the therapeutic potential of monoclonal antibodiesin these conditions.

INFECTION

Agammaglobulinaemic patients suffer from re-current bacterial sinopulnonary infection, meningitis,and bacteraemia.2 Viral infections are no more severein patients with agamm.aglobulinaemia than in healthypeople, suggesting that T cells are the most importantinitial defence, but lasting immunity is lacking somultiple bouts of chickenpox and measles may occur.These observations suggest that antibodies should beable to prevent bacterial and viral infections. Indeed,regular administration of purified pooled humanimmunoglobulin provides good protection for patientswith agammaglobulinaemia (and hypogamma-globulinaemia or dysgammaglobulinaemia).

Polyclonal human immunoglobulin preparationshave been used for many years to treat and preventseveral viral diseases including hepatitis A and B,chickenpox, measles, and cytomegalovirus infection.Against this background it is surprising that anti-bacterial and antiviral monoclonal antibodies havebeen used very little in clinical practice. This maychange soon because several antiviral and antibacterialmonoclorial antibodies are under development forhuman trials. For example, humanised versions ofmonoclonal antibodies to herpes simplex virus22 andrespiratory syncytial virus23 have been prepared andhuman antibodies to HIV have been isolated byscreening phage libraries.24 Antiviral monoclonalantibodies can block attachment and penetration ofviruses, opsonise virus and virus infected cells forphagocytosis or antibody dependent cytoxicity, andmediate complement lysis of enveloped virus particlesor infected cells. Cocktails of monoclonal antibodieswill probably give greater benefit than single reagents.However, there are important reservations relating

to the use of antibodies for treating viral disease. SinceT cells and not antibodies seem to be essential foreradicating established viral infections, it can be arguedthat antibodies are unlikely to be useful in treatingthese conditions. Moreover, there is eviden-ce thatcertain viral infections may be enhanced by antiviralantibodies, which can facilitate Fc receptor mediatedviral entry into macrophages and some other cells.

TOXIC STATES

The use of monoclonal antibodies to treat septicshock has been reviewed recently.2' Endotoxin, alipopolysaccharide component of the bacterial cellwall, damages vascular endothelium triggering acascade of events that leads to septic shock. Because

BMJ VOLUME 305 5 DECEMBER1992

Effector functions for antibody targetedtherapy

Blocking Ligand-receptor interactions* Cell adhesion* Virus attachment* Cytokine stimulation

Natural Fc mediated* Complement fixation* Antibody dependent

cellular cytotoxicity* Phagocytosis

Artificial Toxins* Plant toxins* Bacterial toxins

RadioisotopesCytotoxic drugsEnzymes* Prodrug activation* Direct toxicity

Bifunctional Crosslinking* Cytotoxic effectors to

targets* Enzymes or toxins to targets

1426

on 9 Decem

ber 2020 by guest. Protected by copyright.

http://ww

w.bm

j.com/

BM

J: first published as 10.1136/bmj.305.6866.1424 on 5 D

ecember 1992. D

ownloaded from

the target antigen is intravascular IgM monoclonalantibodies can be used. HA-1A, a human IgMantiendotoxin monoclonal antibody, reduced 28 daymortality by 39% in 105 patients with Gram negativebacteraemia.2' Although the result seems impressive,the antibody is expensive and it is difficult to designprotocols that avoid treating large numbers of patientswho subsequently prove not to have had Gram negativebacteraemia. Moreover, the initial study has beenseriously criticised and a second placebo controlledclinical trial of HA-lA has been recommended todetermine whether the antibody should be widelyused.2Tumour necrosis factor is one of the central

mediators of septic shock, and monoclonal antibodiesagainst it are protective in animal models. A phase Iclinical trial of one monoclonal antibody againsttumour necrosis factor confirmed its safety, but itsefficacy has yet to be shown in humans.28 Besides theobvious examples of tetanus and diphtheria, othertoxic states which may be amenable to monoclonalantibody therapy include drug overdosage, chemicalpoisoning, and snake or spider bites. Already digoxinFab fragments are well established for the managementof digoxin overdose and monoclonal antibodies arebeing developed for neutralising tricyclic antide-pressants.29

CANCER (SOLID TUMOURS)

Monoclonal antibodies against cancers have beenused for both imaging and treatment. There arenumerous possible target antigens, which fall intoseveral broad categories (box). Except in a few cases,unique tumour specific antigens have not been identi-fied, and studies have focused on target antigens thatare present to a greater or less degree on some normalhost tissues. Examples include oncofetal antigenssuch as carcinoembryonic antigen and c fetoprotein,epidermal growth factor receptors, carbohydrateantigens, and components of the extracellular matrixsuch as mucin. Radioimmunoconjugates accumulatein tumour deposits well enough to produce reasonableimages,30 although the image is not yet good enoughseriously to challenge conventional imaging methodssuch as computed tomography. Treatment of cancerwith monoclonal antibodies has so far been disappoint-ing.34 Early studies used immunogenic murinemonoclonal antibodies that could not recruit humaneffector functions. Humanising these monoclonalantibodies or linking them to radioisotopes, toxins,

False colour scanning electron micrograph of hybridoma cells used to produce monoclonal antibodies fortreating cancer

Target antigens for monoclonalantibodies against cancer

Unique to tumour

Relative abundancein tumour

Confined to tumourand non-essentialnormal tissues

Stromal targets

ImmunoglobulinsT cell receptorsMutated cell surface proteinsGrowth factor receptorsOncofetal antigensDead cell markersDifferentiation antigens

Endothelial activation markersFibroblast activation markers

and drugs (which may increase immunogenicity) has sofar had little impact on their therapeutic efficacy andcan produce serious toxicities. However, it would beinappropriate to discount the potential of these alter-native killing mechanisms.An important limiting factor is the inability of

infused monoclonal antibodies to reach the target cells.Monoclonal antibodies have good access to the tumoursurface as the surface blood vessels of a tumour depositare relatively leaky to macromolecules but thebranches of these vessels which penetrate the tumourparenchyma are not.3' 32 Once on the surface, however,they meet an impenetrable wall of tumour cells heldtogether by tight intercellular junctions, which makesaccess to deeper parenchymal regions of the tumourpoor. It was hoped that smaller versions of theantibody molecule-for example, Fv, single chainFv-would escape more readily from penetratingvessels and permeate better through the parenchymalregions of the tumour. However, the early signs arethat their lack of avidity (they are univalent) and rapidrenal excretion result in lower absolute tumour uptakedespite a better tumour to normal tissue ratio.3The toxicity of cancer monoclonal antibodies has

been variable. With unmodified murine monoclonalantibodies fever, rigors, nausea, and vomiting arecommon after the initial doses, immediate hyper-sensitivity reactions can occur, and symptoms second-ary to circulating immune complexes are sometimesseen after prolonged treatment. Radioimmunocon-jugates usually cause appreciable toxicity to normalbone marrow, and immunotoxins can cause thevascular leak syndrome.

Antibody dependent enzyme prodrug therapy(ADEPT) is a promising research prospect.33 With thistechnique an antibody-enzyme conjugate is admin-istered, which localises to tumour deposits. After afew days during which non-specifically bound mono-clonal antibody is cleared, an inactive prodrug isadministered. The prodrug is converted by monoclonalantibody-linked enzyme in tumour deposits to anactive, tumoricidal drug that is small enough topermeate the deeper regions of the tumour. Humanand humanised antibodies with improved affinity andspecificity are likely to be used increasingly in thefuture. One recent animal experiment has shown thatimproved affinity can give improved anticanceractivity.34 Phage technology could help develop appro-priate antibodies.35 Cocktails of monoclonal antibodiesmay give better results than single antibodies (seebelow).

HAEMATOLOGICAL MALIGNANCIES

Monoclonal antibody treatment for haematologicalmalignancies has been more successful than that forsolid tumours. Activity against disease in bone marrowand spleen has been notable, with nodal diseaseresponding less readily.' One possible explanation isthat the sinusoidal circulations of responsive organs are

BMJ VOLUME 305 5 DECEMBER 1992 1 427

on 9 Decem

ber 2020 by guest. Protected by copyright.

http://ww

w.bm

j.com/

BM

J: first published as 10.1136/bmj.305.6866.1424 on 5 D

ecember 1992. D

ownloaded from

easily permeated by immunoglobulins. Also the bonemarrow and spleen are rich in host effector cells thatcan recognise and kill targets coated by monoclonalantibodies.Murine anti-idiotypic monoclonal antibodies have

been raised against unique surface immunoglobulinsand T cell receptors expressed respectively on B andT cell malignancies. The results of treatment areencouraging3637 but monoclonal antibodies have to betailor made for each patient. Altemative targets includea wide array of leucocyte differentiation antigens. Bcell malignancies, for example, have been treated withmonoclonal antibodies against the lymphocyte antigensCD 19, CD22, CD37, and CDw52.5 33Themonoclonalantibodies inevitably destroy some normal lympho-cytes, but these are regenerated from stem cells, whichare not attacked. Transient antibody related immuno-suppression can, however, be troublesome.

Natural effector functions are effective againstseveral haematological malignancies. CAMPATH-1Gis a rat IgG2b monoclonal antibody that recruitshuman complement and antibody dependent cyto-toxicity and binds to an antigen (CDw52) present onmost normal and malignant lymphocytes. Of 29patients with lymphoid malignancies who receivedCAMPATH-1G, nine attained complete remissions,although disease in lymph nodes was generally resistantto treatment.' A CDR grafted version of this antibody(CAMPATH- I H) was the first humanised monoclonalantibody to enter clinical trials and induced completeremissions in two patients with B cell non-Hodgkin'slymphoma, one with lymph node disease.40 Moreextensive clinical testing ofCAMPATH- 1H is currentlyunder way.Immunotoxins and radioimmunoconjugates have

shown activity against lymphoma but direct com-parisons of altemative effector mechanisms on a singlemonoclonal antibody have not yet been made. Poly-clonal antiferritin antisera have been shown to targetHodgkin's disease deposits more efficiently than anti-ferritin monoclonal antibodies,4' which suggests againthat monoclonal antibody cocktails may be the bestway forward.

OTHER APPLICATIONS

Monoclonal antibodies are being developed forimaging of infarcted myocardium (antimyosin), deepvenous or arterial thromboses (antifibrin), and foci ofinfection of inflammation. Antirhesus monoclonalantibodies have been made for treating rhesus haemo-lytic disease and antiplatelet monoclonal antibodies forprevention of intravascular thrombosis. Monoclonalantibody enzyme conjugates targeted at blood clots arealso under development as novel fibrinolytic reagents.

ConclusionsWe have progressed considerably since the early

days of monoclonal antibody therapy but there is stillmuch to leam. Human (or humanised) monoclonalantibodies are preferable to rodent monoclonal anti-bodies for most applications. Cocktails of monoclonalantibodies should be more effective than single anti-bodies and production of such cocktails will be helpedby the advent of human phage antibody libraries.Enhancement of an antibody's affinity is now possibleby phage technology," and the early signs suggest thatit should improve therapeutic efficacy. Definitivestudies comparing the clinical efficacy of variousnatural and artificial effector functions are needed, andthere is scope for boosting natural effector mechanismswith lymphokine therapy. For the future antibodiesand antibody genes may be used increasingly toredirect cytotoxic cells or for targeted delivery of genesand other drugs wrapped up in viruses or liposomes.

We thank Cesar Milstein, Geoff Hale, Kerry Chester, andGreg Winter for encouragement and critical reading of themanuscript. REH is a cancer research campaign seniorclinical research fellow and was formerly a Medical ResearchCouncil training fellow. MBL is funded jointly by the MRCand Celltech. SJR is an MRC clinician scientist with additionalsupport from the Kay Kendall Research Foundation andLouis Jeantet Foundation.

I Savin J. The value of antibody engineenrng techniology to the UK. London: Centrefor Exploitation of Science and Technology, 1990.

2 Hawkins RE, Llewelyn MB, Russell SJ. Adapting antibodies for clinical use.BMJ 1992;305: 1348-52.

3 Milenic DE, Yokota T, Filpula DR, Finkelman MAJ, Dodd SW, Wood JF, etal. Construction, binding properties, metabolism, and tumor targeting of asingle-chain Fv derived from the pancarcinoma monoclonal antibody CC49.Car,cerRes 1991;51:6363-71.

4 Waldmann TA. Monoclonal antibodies in diagnosis and therapy. Science1991;252: 1657-62.

5 Dyer MJS, Hale G, Marcus R, Waldmann H. Remission induction in patientswith lymphoid malignancies using unconjugated CAMPATH-I monoclonalantibodies. Leukaenia and Lyntphonia 1990;2:179-93.

6 Ortho Multicentre Transplant Study Group. A randomised trial of OKT3monoclonal antibody for acute rejection of cadaveric renal transplants. NEnglJMed 1985;313:337-42.

7 Carpenter CB. Immunosuppression in organ transplantation. N Engl J Med1990;322:1224-6.

8 Gorman SD, Clark MR, Routledge EG, Cobbold SP, Waldmann H.Reshaping a therapeutic CD4 antibody. Proc Natl Acad Sci USA 1991;88:4181-5.

9 Queen C, Schneider WP', Selick HE, Payne PW, Landolfi NF, Duncan JF, etal. A humanized antibody that binds to the interleukin 2 receptor. Proc Nat!Acad Sci USA 1989;86:10029-33.

10 Riechmann L, Clark M, Waldmann H, Winter G. Reshaping humanantibodies for therapy. Nature 1988;332:323-7.

11 Reinke P, Miller H, Fietze E, Herberger D, Volk HD, Neuhaus K, et al. Anti-CD4 therapy of acute rejection in long-term renal allograft recipients. Lancet1991;338:702-3.

12 Horneff G, Burmester GR, Emmrich F, Kalden JR. Treatment of rheumatoidarthritis with an anti-CD4 monoclonal antibody. Arthritis Rheum 1991;34:129-40.

13 Emmrich J, Seyfarth M, Fleig WE, Emmrich F. Treatment of inflammatorybowel disease with anti-CD4 monoclonal antibody. Lancet 1991;338:570-1.

14 Hiepe F, Volk HD, Apostoloff E, Baehr RV, Emmrich F. Treatment of severesystemic lupus erythematosus with anti-CD4 monoclonal antibody. Lancet1991;338: 1529-30.

15 Nicolas JF, Chamchick N, Thivolet J, Wijdenes, Morel P, Revillard JP. CD4antibody treatment to severe psoriasis. Lancet 1991;338:32 1.

16 Van der Lubbe PA, Miltenburg AM, Breedveld FC. Anti-CD4 monoclonalantibody for relapsing polychondritis. Lancet 1991;337:1349.

17 Mathieson PW, Cobbold SP, Hale G, Clark MR, Oliveira DBG, LockwoodCM, et al. Monoclonal antibody therapy in systemic vasculitis. N EnglJ Med1990;323:250-4.

18 Sollilou J, Cantarovich D, LeMauff B, Giral M, Robillard N, Hourmant M, etal. Randomised controlled trial of monoclonal antibody against the inter-leukin 2 receptor (33B3. 1) as compared with rabbit antithymocyte globulinfor prophylaxis against rejection of renal allografts. N Engi 7 Med1990;322: 1175-82.

19 Brown PS, Parenteau GL, Dirbas FM, Garsia RJ, Goldman CK, BukowskiMA, et al. Anti-Tac-H, a humanized antibody to the interleukin 2 receptor,prolongs primate cardiac allograft survival. Proc Natl Acad Sci 1991;88:2663-7.

20 Spickett GP, Misbah SA, Chapel HM. Primary antibody deficiency in adults.Lattcet 1991;337:281-4.

21 Webster ADB. Intravenous immunoglobulins. BMJ 1991;303:375-6.22 Co MS, Deschamps M, Whitley RJ, Queen C. Humanized antibodies for

antiviral therapy. Proc NatlAcad Sci USA 1991;88:2869-73.23 Tempest PR, Bremner P, Lambert M, Taylor G, Furze JM, Karr FJ, et al.

Reshaping a human monoclonal antibody to inhibit human respiratorysyncytial virus infection in vivo. Biological Technology 1991;9:266-7 1.

24 Burton DR, Barbas III CF, Persson MAA, Koenig S, Chanock RM, LemerRA. A large array of human monoclonal antibodies to type I humanimmunodeficiency virus from combinatorial libraries of asymptomaticseropositive individuals. Proc NatlAcad Sci USA 1991;88:10134-7.

25 Hinds CJ. Monoclonal antibodies in sepsis and septic shock. BMJ 1992;304:132-3.

26 Ziegler EJ, Fisher CJ Jr, Sprung CL, Straube RC, Sadoff JC, Foulke GE,et al. Treatment of Gram-negative bacteremia and septic shock with HA-lAhuman monoclonal antibody against endotoxin: a randomized, doubleblind, placebo-controlled trial. N EtglJMed 1991;324:429-36.

27 Warren HS, Danner RL, Munford RS. Sounding board. Anti-endotoxinmonoclonal antibodies. NEnglJMed 1992;326:1153-7.

28 Exley AR, Cohen J, Buurman W, Owen R, Hanson G, Lumley J, et al.Monoclonal antibody to TNF in septic shock. Lancet 1990;335:1275-6.

29 Kulig K. Initial management of ingestions of toxic substances. N Engl J Med1992;326: 1677-8 1.

30 Order SE. Presidential address: systemic radiotherapy-the new frontier.Intl Radiat Oncol Biol Phys 1990;18:981-92.

31 Dvorak HF, Nagy JA, Dvorak AM. Structure of solid tumors and theirvasculature: implications for therapy with monoclonal antibodies. CancerCells 1991;3:7i-85.

32 Dvorak HF, Nagy JA, Dvorak JT, Dvorak AMI. Identification and character-ization of the blood vessels of solid tumors that are leaky to circulatingmacromolecules. Am JPathol 1988;133:95-109.

33 Bagshawe KD. Towards generating cytotoxic agents at cancer sites. Br JCancer 1989;60:275-81.

34 Schlom J, Eggensperger D, Colcher D, Molinolo A, Houchebs D, Miller LS,et al. Therapeutic advantage of high-affinity anticarcinoma radioimmuno-conjugates. Cancer Res 1992;52:1067-72.

35 Hawkins RE, Russell SJ, Winter G. Selection of phage antibodies by bindingaffinity: mimicking affinity maturation. J Mol Biol 1992;226:889-96.

36 Brown SL, Miller RA, Homing SJ, Czerwinski D, Hart SM, McElderry R, etal. Treatment of B cell lymphomas with anti-idiotype antibodies alone and incombination with alpha interferon. Blood 1989;73:651-61.

1428 BMJ VOLUME 305 5 DECEMBER 1992

on 9 Decem

ber 2020 by guest. Protected by copyright.

http://ww

w.bm

j.com/

BM

J: first published as 10.1136/bmj.305.6866.1424 on 5 D

ecember 1992. D

ownloaded from

37 Janson CH, Tehrani MJ, Mellstedt H, Wigzell H. Anti-idiotypic monoclonalantibody to a T cell chronic lymphatic leukaemia. Cancer ImmunolImmunother 1989;28:225-32.

38 Vitetta ES, Stone M, Amlot P, Fay J, May R, Till M, et al. Phase Iimmunotoxin trial in patients with B-cell lymphoma. Cancer Res 1991;51:4052-8.

39 Press OW, Eary JF, Badger CC, Martin PJ, Appelbaum FR, Levy R, et al.Treatment of refractory non-Hodgkin's lymphoma with radiolabelled MB- I(anti-CD37) antibody. IClin Oncol 1989;7:1027-38.

40 Hale G, Dyer MJS, Clark MR, Phillips JM, Marcus R, Riechmann L, et al.Remission induction in non-Hodgkin's lymphoma with reshaped humanmonoclonal antibody CAMPATH- I H. Lancet 1988;ii: 1394-9.

41 Vriesendorp HM, Herpst JM, Germack MA, Klein JL, Leichner PK,Loudenslager DM, et al. Phase I-II studies of yttrium-labelled antiferritintreatment for end-stage Hodgkin's disease, including radiation therapyoncology group 87-01. J Clin Oncol 1991;9:918-28.

(Accepted 19 October 1992)

OBITUARY

TNA effcoate

Obituaries must be submittedexclusively to the BMJ; theyshould be under 250 wordsand submitted within threemonths of the person's death.We welcome selfwnrttenobituaries and good quality,recent photographs, and, as amedicaljournal, we encourageauthors to include the causeofdeath.

Sir NORMAN JEFFCOATEMD, FRCSED, FRCOG

Norman Jeffcoate was a student, protege, and discipleof Professor William Blair Bell, the founder and firstpresident of the College of Obstetricians and Gynae-cologists. His career progress was rapid; he wasappointed to the first whole time chair in obstetrics andgynaecology in the University of Liverpool in 1945.He was a wonderful teacher and a great supporter of

the undergraduates. They received superb tuition in allthe basic skills of the two disciplines from a man whosemeticulous attention to history, detail, and physicalsigns never allowed issues and principles to be clouded.He had an uncanny ability to bring on postgraduates atthe appropriate pace, involving them in researchprojects at just the right time. His main work, Prin-ciples of Gynaecology, was first published in 1957 and isnow in its fifth edition and remains a classic. It hasgiven generations of aspiring obstetrician-gynaecolo-gists a wealth of knowledge presented with great clarityin an uncomplicated style. Understandably, he was ingreat demand worldwide: he gave more than 500 guestlectures and orations and was a visiting professor ateight universities.The Gynaecological Visiting Society started by Blair

Bell had an important role in the foundation of theCollege of Obstetricians and Gynaecologists, and Pro-fessor Jeffcoate was its convenor for seven years. Thusan interest in the college was as natural as his becomingits president in 1969; his influence at the college wasimmense and he worked tirelessly for it, never missingthe chance to promote it and its ideals among hischarges. After his retirement he withdrew from allprofessional activities, though he maintained a keeninterest in college affairs.As with many great men, he had other talents. He

was accomplished at both rugby football and cricket,an artist (in watercolours), and, with his wife, Jose-phine, who died in 1981, a warm and welcoming host.He and Josephine had four sons, one of whom is aprofessor of endocrinology and one a consultantphysician.

Friends and colleagues will miss his wisdom, breath-taking long term memory, loyal support, and friend-ship. His contribution to obstetrics and gynaecologywas immeasurable.-R D ATLAY

Thomas Norman Arthur Jteffcoate, professor of obstetrics andgynaecology at Liverpool University 1945-72 and consultantobstetrician and gSynaecologist to Liverpool hospitals, died 13November. Born Nuneaton, 25 March 1907; educated KingEdward VI School, Nuneaton, and Liverpool University (MB,ChB 1929). Honorary assistant surgeon at Liverpool MaternityHospital 1932-45 and Women's Hospital, Liverpool, 1935-45.Knighted 1970.

S HEROLDMRCPSYCH

Throughout her training Sigrid Herold maintained aninterest in research, making a valuable contribution topositron emission tomography and specialising in its

application to autism, schizophrenia, and cerebro-vascular disease. Most recently her research alsoencompassed psychosocial and epidemiological aspectsof schizophrenia. Her critical mind and practicalresearch abilities were enhanced by the clarity,enthusiasm, and sense of humour with which shecommunicated. She believed strongly in the value ofgood teaching in medicine and facilitated a highstandard of postgraduate psychiatric training at StBemard's Wing, Ealing Hospital.

Outside work Sigrid delighted in being with herfamily. Her openness to affection and generous naturewere always evident. Her interest in art and socialissues, and her good cooking gave rise to enjoyablehospitality. She successfully married a busy andfulfilling work and home life, and through her experi-ence she offered support to other women doctors whoknew her. She died of cholangiocarcinoma and issurvived by her husband, Terry Jones, and 4 year olddaughter, Julia.-Jo BOWEN

Sigrid Herold, a senior registrar on the Charing Cross highertraining scheme in psychiatry since 1990, died 19 J7uly aged 39.Studied medicine at University of Tubingen, West Germany (MD1978). Visiting research fellow at Medical Research Council'sCyclotron Unit, Hammersmith Hospital, 1983-5. Registrar inpsychiatry at Maudsley Hospital, then senior registrar at StBernard's Wing, Ealing Hospital.

A P H RANDLEFRCPI, DPHYSMED

Barley Randle was a pioneer of physical medicine andparticularly concemed with rehabilitating patientswith neurological conditions, including stroke andhead injury. He was especially able in helping patientsto come to terms with their problem and make the bestof it.

Barley found no difficulty in coordinating a teamof consultants, general practitioners, clinical psy-chologists, speech and occupational therapists, andphysiotherapists. In 1966, as a result of his efforts, thefirst employment rehabilitation centre based in ahospital was established at Garston Manor. It was runby the Department of Employment with sharedmedical facilities, the aim being to assess patients forwork or vocational training while they had medicaltreatment. It proved a great success. In 1972 Barleymoved to Cambridge to open a new rehabilitationcentre at Addenbrooke's Hospital. When the unitopened in 1976, after delays and setbacks, it was notentirely to Barley's satisfaction, but it became ademonstration centre.

Barley was a warm, cultured man with a puckishsense ofhumour who got on well with people. I suspectthat his interest in people dictated his resignation fromthe NHS in 1979 and appointment as medical directorof the Papworth-Enham trusts. These charitable trustswere designed to promote work skills by providingassessment, training, housing, and jobs in a villagesetting.

In 1986 he retired to Brighton. Sadly, in 1990 hedeveloped a disorder that fell within his own specialty.

BMJ VOLUME 305 5 DECEMBER 1992 1429

on 9 Decem

ber 2020 by guest. Protected by copyright.

http://ww

w.bm

j.com/

BM

J: first published as 10.1136/bmj.305.6866.1424 on 5 D

ecember 1992. D

ownloaded from