Marco Conceptual

W hen rad io lab elled antibodies are injected in tra v en o u sly into mice bearing hum an tum our xenografts, antibody accum ulates rapidly in tu m our w ith in a few hours and persists for several days w hile levels in the circulation rem ain high. H ow ever, p ersisten t circulating an tibody results in low tum our to blood ratios which both delays the time at w hich

im aging can be perform ed and limits the dose of therapeutic antibody by causing dam age to norm al tissues.

A ntibody fragm ents are cleared m ore rapidly from the circulation and have better tum our penetration m aking them an alternative to intact a n tib o d y for tu m o u r localisation and th erap y . In an im al m odels, radiolabelled F(ab')2 antibodies have superior tum our to blood ratios and

red u ced toxicity com pared w ith intact an tib o d y (Pedley et ah, 1993). H ow ever tw ice the am ount of F(ab')2 m u st be ad m in istered as intact

antibody to achieve sim ilar therapeutic effects as intact antibody. This is because F(ab')2 is cleared m ore rapidly du rin g the first few h ours after

ad m inistration in mice so that a low er absolute am o u n t of an tibody is d eliv ered to tu m o u r. C onjugation of p o ly eth y len e glycol (PEG) to proteins can alter their pharm acological properties resulting in extended plasm a half-life, red u ced im m unogenicity an d antigenicity, increased solubility and resistance to proteolysis (Buchegger et al., 1989). PEG m odification of antibody fragm ents has been show n to enhance tu m o u r a ccu m u latio n an d red u ce n o rm al tissu e to b lo o d ratio s a n d m ay additionally reduce imm unogenicity (Cheung et aL, 1986)

In m an P(ab')2 fragm ents p en etrate m ore rap id ly into tu m o u rs

c o m p ared w ith intact an tib o d y b u t clearance rates are sim ilar an d cum ulative dose rates in tu m our are com parable (Kohler et aL, 1975). It has been su g g ested th a t F(ab')2 m ight be a good com prom ise as a

th e ra p e u tic m o lecu le in m an g iv in g h ig h a b so lu te a m o u n ts of ra d io th e ra p y in tu m o u r w ith rap id p en etratio n . Sm aller a n tib o d y frag m en ts in clu d in g Fab' and single chain Fv an tib o d ies p e n e tra te tum ours w ell b u t their therapeutic efficacy is lim ited by rapid clearance and accum ulation in kidney.

A p o te n tia l a p p ro a ch to red u c e th e toxicity of c irc u la tin g radiolabelled antibodies is to inject a second antibody directed against the

first resulting in the form ation of im m une com plexes th at are rap id ly cleared by the reticuloendothelial system (M iller et aL, 1982; Schlom, 1986). More recent w ork has exam ined the use of avidin or streptavidin w h ich h av e h ig h affinity for b io tin to com plex w ith c irc u la tin g biotinylated antibodies and thereby increase clearance of the com plexes (Schroff et aL, 1985; Pedley et aL, 1993). The presence of circulating tum our antigen m ay also affect clearance and tu m o u r uptake of antibodies. In mice b earin g h u m an pancreatic xenografts, elevated serum CEA w as associated w ith rap id clearance of a n tib o d y -an tig en com plexes an d reduced tu m our localisation to one quarter of th at seen in mice w ith low serum CEA (Pedley et aL, 1989).

Factors w hich m ay influence localisation of antibodies to tum ours include the concentration gradient betw een blood and interstitial fluid su rro u n d in g the tum our and the ability of the antibody to perm eate the interstitial spaces and into the tum our itself. There is m uch variability in uptake of antibody in individual tum ours in patients for reasons that are poorly understood. Injected antibodies are n o t distributed uniform ly in tum ours and do not alw ays penetrate efficiently into some areas even if they are antigen-rich (Lenhard et aL, 1985; O rder et aL, 1985). The basal lam ina m ay p resen t a significant b arrier to ex trav asated an tib o d ies (D v o rak et aL, 1991) and preferential expression of antigens on tu m o u r cells m ay render them inaccessible to antibodies (Pervez et aL, 1988; Boxer

et aL, 1994)

The im portance of the affinity of the antibody for the antigen is unclear. In anim als, specific antibody is retained in tum ours longer than a non-specific antibody and uptake of antibody has been sh o w n to be greater w ith antibodies of high affinity (Spitler et aL, 1987). A lternatively it has been suggested that high affinity antibodies interact strongly w ith antigen at the tu m our surface resulting in reduced penetration (Shawler

et aL, 1985). A ntibody uptake m ay also d epend on the physiochem ical properties of antibodies (Reichmann et ah, 1988)

Tum our vasculature is likely to differ betw een xenografts in mice an d tu m o u rs in m an and m ay account for som e of the differences in localisation that are seen. Approaches to im prove therapy have included th e u se of v aso d ila to rs or in terleu k in -2 to in crease p e rfu sio n or perm eability of tum ours before delivery of im m unoconjugate (De Leij et aL, 1987; S ouham i et aL, 1990) or the a d m in istra tio n of sy n th etic flavonoids such as flavone-8-acetic acid (FAA) w hich reduce blood flow and induce necrosis after antibody has been delivered (K rishnam urthy et aL, 1990).