ABSTRACT The distribution of carcinoembryonic antigen (CEA) in colorectal cancer (CRC) differs from that in normal colorectal tissue, being found on all borders of the cell membrane and

hence enabling access to intravenous antibody, making CEA a good target for antibody-based therapy. The distinctive anti-CEA antibody, PR1A3, binds only membrane-bound CEA. Humanised PR1A3

(hPR1A3) was assessed both _in vitro_ cytotoxicity and binding assays with colorectal cancer cell lines expressing varying levels of CEA. Human peripheral blood mononuclear cells (PBMCs) and

purified natural killer (NK) cells were used as effectors. The _in vitro_ assays demonstrated hPR1A3 CEA-specific binding and antibody-dependent and CEA-specific killing of human colorectal

mediators of hPR1A3-based CEA-specific killing. The results support the development of hPR1A3 for therapy of colorectal cancer. SIMILAR CONTENT BEING VIEWED BY OTHERS ONCOLYTIC PEPTIDES

DTT-205 AND DTT-304 INDUCE COMPLETE REGRESSION AND PROTECTIVE IMMUNE RESPONSE IN EXPERIMENTAL MURINE COLORECTAL CANCER Article Open access 24 March 2021 ΓΔ T CELLS ARE EFFECTORS OF

IMMUNOTHERAPY IN CANCERS WITH HLA CLASS I DEFECTS Article Open access 11 January 2023 CHEMOTHERAPY AGENTS STIMULATE DENDRITIC CELLS AGAINST HUMAN COLON CANCER CELLS THROUGH UPREGULATION OF

THE TRANSPORTER ASSOCIATED WITH ANTIGEN PROCESSING Article Open access 27 April 2021 MAIN There has been a recent resurgence of interest in the use of monoclonal antibodies for the treatment

of cancer, mostly in the setting of metastatic disease, and already a number of antibodies have been licensed for use (for review see Reichert and Valge-Archer, 2007). Antibody therapy

directed at cell surface components, in particular cell membrane receptors, has the potential to exploit the specificity and sensitivity of the immune system's ability to recognise cell

membrane receptors, and so to achieve selective therapeutic effects without the toxicity of standard chemotherapy. As a result antibody therapy has become a useful adjunct to the treatment

of cancer. This comes despite poor initial results with murine antibodies (Clynes, 2006). The development of chimeric and humanised antibodies has led to the development of a second

generation of antibodies that hardly, if at all, stimulate the development of human anti-mouse antibodies, and which are potent activators of the immune system (Carter, 2006). Currently,

about 19 antibodies are licensed for clinical use, 11 of which are for the treatment of cancer (Reichert and Valge-Archer, 2007). There are, it is estimated, at least a further 150

antibodies in development. Antibody use has had a major impact on the treatment of haematological malignancies with excellent response rates seen using rituximab (anti-CD20) in follicular

and B cell non-Hodgkin's Lymphoma (Olszewski and Grossbard, 2004). Their impact in solid tumours has, however, been less dramatic. A major reason for this must be the limited

penetration of antibodies, being large glycoproteins, into the tumour. Three antibodies have been licensed by the US FDA for use in advanced colorectal cancer. These include antibodies

against EGFR (cetuximab) and VEGF (bevacizumab). A problem with the former is the presence of EGFR on cells of a variety of normal tissues leading to a degree of nonspecificity with respect

to cancer. The EGFR antibody also appears to be effective in only a proportion of cancer patients. It is therefore important to look for other potential antigens to use as targets to broaden

the number of patients for which antibody-based immunotherapy may be effective. Carcinoembryonic antigen (CEA or CEACAM5) was first recognised as a potential human tumour-specific antigen

in the 1960s (Gold and Freedman, 1965; Berinstein, 2002). CEA has characteristics that make it a useful target for antibody therapy in colorectal cancer (CRC). It is overexpressed in the

vast majority of CRCs (Chan and Stanners, 2007). More importantly, however it is always aberrantly expressed in CRC. While the CEA expressed on normal colonic epithelium is inaccessible to

IgG antibody, being found only on the luminal surface of the cell, this expression pattern changes in the neoplastic cell so that CEA is additionally expressed on the basal and lateral

membranes (Hammarstrom, 1999) making it accessible to blood-borne antibody. Although a member of the Ig superfamily, CEA is linked to the cell membrane by a glycophosphatidylinositol (GPI)

anchor (Thompson et al, 1991) and thus has no direct intracellular signaling motif. The functional significance, if any, of its overexpression in colorectal cancer remains unclear. The

biggest disadvantage of CEA as a target is that it is readily cleaved from the cell surface and so shed into the blood stream from tumours, either directly or via the lymphatics, which is

why the level of serum CEA has been used in the clinic as a marker for screening and recurrence, especially of colorectal cancer (Chau et al, 2004). Serum CEA binds to most of the currently

used anti-CEA antibodies hindering them from reaching their target and so largely mitigating against any potential clinical effect. This is, however, not the case for the murine antibody,

mPR1A3 developed in the ICRF laboratories in London by Richman and Bodmer (1987). PR1A3 was shown to target the B3 domain and GPI anchor of the CEA molecule by Durbin et al, 1994 (see Figure

1), and was subsequently humanised by Stewart et al, 1999. Murine PR1A3 has been shown to react poorly with soluble CEA, which lacks the GPI anchor, and has been used in immunoscintigraphy

for the detection of colorectal tumours with a high degree of specificity (Granowska et al, 1989). In contrast to EGFR and VEGF, there are so far no unconjugated, or ‘naked’ antibodies to

CEA being used for the treatment of colorectal cancer. The anti-CEA antibodies that are currently used in pilot trials, are administered as radioconjugates (Wong et al, 2004; Liersch et al,

2007). Strong evidence for the suggestion that the antitumour effects of antibodies are mediated mainly by ADCC (antibody-dependent cellular cytotoxicity) comes from mouse knockout studies

which showed that antibody antitumour effects were largely absent in mice lacking the appropriate Fc receptor (Clynes et al, 1998). In humans, natural killer (NK) cells that express the CD16

Fc_γ_receptor are thought to be the main cell type that could mediate ADCC of tumours (Arnould et al, 2006). Several criteria, therefore, need to be fulfilled before an antibody can be

considered for therapy (Reichert and Valge-Archer, 2007). Humanised antibodies are more potent activators of the human immune system than their murine equivalents. Internalisation of

antibody-antigen complexes, leading to a loss of antibody complexes available for binding to Fc receptor-bearing cells, will diminish the effectiveness of antibody-based killing. This study

has three aims. First to analyse the ability of humanised hPR1A3 to react with a panel of colorectal cancer cell lines expressing high and low levels of CEA. Second to explore the use of

hPR1A3 in an _in vitro_ cytotoxicity model and to compare human PBMC with partially purified NK cells as effectors for ADCC. The third aim was to investigate whether soluble CEA inhibited

the ADCC activity of hPR1A3 _in vitro_. MATERIALS AND METHODS CELL LINES The gastric carcinoma cell line MKN45 (Motoyama et al, 1979) was obtained from Cell Services LIF, Cancer Research UK.

All other cell lines were colorectal cell lines: HCT-116 (Brattain et al, 1981) and SKCO-1 (Fogh, 1975) were originally obtained from ATCC; GP5d (Solic et al, 1995) and HT55(Watkins and

Sanger, 1977) were obtained from ECACC; LS174T (Tom et al, 1976) was obtained from BH Tom, NW University Med Centre, Chicago, Ill, USA, PC/JW (Paraskeva et al, 1984) was obtained from C.

Paraskeva, Directors Lab, CRUK, London and C70 (Browning et al, 1993) was established in the Cancer and Immunogenetics Laboratory. All cell lines were maintained in culture in

Dulbecco's modified Eagle's (E4) medium with 1% L-glutamine, 10% fetal calf serum (FCS) and 1% penicillin/streptomycin. They were incubated at 37°C in a humidified 10% CO2

environment. For the chromium-release and EuTDA assays the cells were suspended in 2% RPMI-1640 medium with 1% glutamine and 10% FCS. PERIPHERAL BLOOD MONONUCLEAR CELLS Peripheral blood

mononuclear cells (PBMCs) were isolated either from fresh whole blood from healthy laboratory volunteers, having taken informed consent, or from leucodepletion filters obtained from single

donors following blood donation (Courtesy of Cristina Navarrete, Colindale National Blood Service, London, UK). White cells were eluted from the filter using 5 mM ethylenediaminetetraacetic

acid. Fresh whole blood was mixed with an equal volume of an RPMI-1640/citrate solution (40 ml 3.3% sodium citrate, 2 ml 5 mM 2-mercaptoethanol, 200 ml RPMI-1640+hepes) as an anticoagulant.

Both sources of PBMC were then processed in an identical fashion. Following Ficoll/hypaque density centrifugation, the PBMC layer was withdrawn from the interface and washed once with

RPMI-1640 to remove excess Ficoll, spinning at 800 G (Boyum, 1968) and then a second time, spinning at 200 G for 10 min to remove platelets. The resulting PBMCs were resuspended in

RPMI-1640/10% FCS/1% glutamine (complete RPMI-1640), kept at room temperature and used within 12 h of preparation. ANTIBODIES PR1A3 The original is a murine IgG1_κ_ monoclonal antibody to

CEA (Richman and Bodmer, 1987) that was later humanised (Stewart et al, 1999). Both murine (mPR1A3) and humanised (hPR1A3) antibodies were obtained from the Biotherapeutics Development Unit,

Clare Hall, Cancer Research UK, London, UK. ANTI-CD16 Two different clones of this antibody were used: MEM154 (Biovendor Laboratory Medicine Inc.) and 3G8 (BD Biosciences, Pharmingen, USA).

Both are murine monoclonal IgG1_κ_ antibodies against the human Fc_γ_IIIA receptor (CD16A; FCGR3A). F(ab’)2 derived from the 3G8 clone was obtained from Ancell Corp, Bayport, MN, USA.

ANTI-PROSTATE SPECIFIC MEMBRANE ANTIGEN Murine monoclonal antibody 107-1A4 to PSMA was kindly provided by Robert Vessella (Univ Washington, USA). POLYCLONAL RABBIT ANTI-MOUSE ANTIBODY This

was obtained from DAKO A/S, Denmark, and was diluted to 1 : 100 in RPMI-1640/1% FCS. ANTI-_Β_-GALACTOSIDASE ANTIBODY clone 4C7 (Durbin and Bodmer, 1987), was obtained from the Monoclonal

Antibody Service, Clare Hall, Cancer Research UK, London, UK, and used at a final concentration of 0.4 _μ_g ml−1 in RPMI-1640/1% FCS. AUA-1 This was obtained from the Monoclonal Antibody

Service, Clare Hall, Cancer Research UK, London, UK (Epenetos et al, 1982). It has been shown to be an anti-EpCam antibody (Spurr et al, 1986). FITC-CONJUGATED MURINE ANTI-HUMAN IGG AND

FITC-CONJUGATED RABBIT ANTI-MOUSE IGG These were obtained from Sigma-Aldrich, Poole, UK. CD3-FITC, CD16-PE, CD56-APC, IGG1-FITC (ISOTYPE CONTROL), IGG1-PE (ISOTYPE CONTROL) AND IGG1-APC

(ISOTYPE CONTROL) These were obtained from BD Biosciences Pharmingen, Oxford, UK. CHROMIUM-RELEASE CYTOTOXICITY ASSAY Peripheral blood mononuclear cells were prepared as above and suspended

at a concentration of 1 × 107 ml−1 in RPMI-1640 supplemented with 10% FCS and 1% glutamine (complete medium). 2 × 106 target cells were centrifuged at 400 G for 5 min, decanted and labelled

by resuspending the pellet in 0.1 ml of 7.4MBq of Na251CrO4. This cell suspension was incubated at 37°C for 60–100 min depending on the optimal labelling time for the particular cell line.

Optimal labelling time had previously been determined by choosing the labelling time with the highest maximal lysis:background radioactivity ratio. The cells were then washed twice with

RPMI-1640 and suspended in complete RPMI-1640 at a concentration of 1 × 105 cells per ml. Target cells (100 _μ_l (1 × 104)) and PBMCs (100 _μ_l (1 × 106)) were added to microcentrifuge tubes

to give an effector: target ratio of 100 : 1. Various antibody concentrations (20 _μ_l) were then added to the relevant tubes. Triton (120 _μ_l of 5%) was added to 100 _μ_l of target cells

to obtain maximum release values. All tubes were made up to the same volume using complete RPMI-1640. The tubes were spun at 200 G for 2 min and the pellet of combined target and effector

cells was incubated at 37°C for 4 h in the presence of antibody. The tubes were then spun at 200 G for 5 min and 35 _μ_l of the supernatant added to 100 ul Optiphase Supermix (Perkin Elmer,

Boston, MA, USA) in a 96-well plate. The 51Cr concentration in each well was then determined using a Microbeta plate reader. FLUORESCENCE-BASED EUTDA CYTOTOXICITY ASSAY Five microlitres of

BATDA (Blomberg et al, 1996) (Perkin Elmer, Boston, MA, USA) were added to 2 × 106 target cells suspended in complete RPMI-1640 and incubated for 10–25 min at 37°C depending on the optimal

labeling time for the particular cell line, determined as described for the 51Cr-release assays. The relative concentrations of target cells and antibody in the microcentrifuge tubes were

similar to those used for the 51Cr release assay. However, varying effector:target cell ratios were used for the fluorescence-based assay. After spinning at 200 G for 2 min the tubes were

incubated at 37°C for 2 h and then spun again at 200 G for 5 min. A total of 20 _μ_l of the resulting supernatant was added to 200 _μ_l of Europium in black 96-well plates and the resulting

fluorescence was then read in a time-resolved fluorometer using an excitation wavelength of 340 nm and an emission wavelength of 615 nm. CEA ELISA ASSAY CEA levels were determined using a

_β_-galactosidase/anti-_β_-galactosidase ELISA (Durbin and Bodmer, 1987). Cells were plated onto a poly-L-lysine coated 96-well Nunc-Immuno PolySorp plate at a concentration of 2.5 × 104

cells per well. Murine PR1A3 was used as the CEA-detecting antibody with rabbit anti-mouse antiserum as the secondary antibody. The GAG complex of _β_-galactosidase with

anti-_β_-galactosidase antibody was made by dissolving _β_-galactosidase (_E.coli β_-galactosidase lyophilised powder; Sigma-Aldrich, Poole, Dorset, UK) at a concentration of 500 U ml−1 in

100 mM TRIS/100 mM MgCl2/100 mM 2-mercaptoethanol with 300 _μ_g ml−1 of 4C7 and incubating this complex at 4°C overnight. The complex so formed was then added at a dilution of 1 : 750 in

RPMI-1640/1% FCS. The GAG complex binds to free antigen-binding sites on the rabbit anti-mouse IgG antibody, which is already bound to the PR1A3 attached to the CEA on the cells. The

substrate 4-methylumbelliferyl-B-D-galactoside (Sigma-Aldrich, Poole, Dorset, UK) was first prepared at a concentration of approximately 0.3 mg ml−1 in a buffer of 1 mM MgCl2/100 mM

2-mercaptoethanol . After stirring for 30 min, the solution was filtered to remove excess substrate. This substrate solution was added to each well to start the reaction. After incubation

for 40 min in the dark at room temperature, fluorescence was measured using an excitation wavelength of 365 nm and an emission wavelength of 445 nm. FLOW CYTOMETRIC ANALYSIS OF CEA

INTERNALISATION The concentration of cells to be analysed was adjusted to 1 × 106 ml−1 and the cells washed once with phosphate-buffered saline (PBSA) containing 2% FCS and then centrifuged

at 400 G for five minutes. The resulting pellet was resuspended in cold PBSA containing 2% FCS and incubated on ice with either 100 _μ_l of hPR1A3 at 20 _μ_g ml−1, 100 _μ_l of AUA-1 at 15 

_μ_g ml−1 or a medium control for 30 min. The cells were next washed again, with PBSA containing 2% FCS and incubated in a water bath at 37°C for 0, 1, 2 or 3 h. The cells were then

incubated with a 1 : 50 dilution of FITC-conjugated anti-human-IgG for hPR1A3 and a 1 : 100 dilution of FITC-conjugated anti-murine IgG for AUA-1 on ice in the dark for 30 min before being

washed again and resuspended in PBSA containing 2% FCS. The cells were then passed through a fluorescence-activated cell sorting (FACS) Calibur flow cytometer and the results analysed using

CellQuest software. NK CELL ENRICHMENT Fresh PBMC or eluted PBMC from leucodepletion filters were enriched for NK cells using a Human NK Cell isolation kit (Miltenyi Biotec) following the

manufacturer's instructions. This involved adding the NK cell Biotin-Antibody cocktail (10 _μ_l per 107 cells) which contained antibodies against T cells, B cells, stem cells, dendritic

cells, monocytes, granulocytes and erythroid cells. Following incubation for 10 min at 4°C, the NK Cell magnetic Microbead Cocktail was added. This was left for 15 min at 4°C and the cells

then washed with MACS buffer (PBSA/0.5% FCS/2 mM EDTA) and centrifuged. The cell pellet was resupended in 0.5 mls MACS buffer and then passed through an LS magnetic column to remove cells

that bound the antibodies in the cocktail. The cells that passed through were then collected and suspended in RPMI-1640 with and without IL-2 (Peprotech) at 10 ng ml−1 and incubated

overnight at 37°C. Aliquots pre and post sorting were taken for FACS analysis with the antibodies CD3, CD56 and CD16 and the appropriate isotype controls. Thus, 1 × 106 cells were suspended

in 2 ml of FACS buffer (PBSA/1%FCS/1% sodium azide) and centrifuged. The supernatant was removed and the cells resuspended in the residual buffer. The antibodies were then added and left for

20 min at 4°C before washing and then fixing with 300 _μ_l of 2% paraformaldehyde. The fixed cells were analysed on a FACS Calibur flow cytometer. Gating of the lymphocyte population in the

forward versus side scatter plot revealed a purity of at least 85% with respect to CD16 and CD56 binding. The resulting NK cells were used in cytotoxicity assays with hPR1A3 against SKCO-1,

which is known to express CEA. EFFECT OF SOLUBLE CEA ON PR1A3-INDUCED ADCC Purified CEA, obtained from human liver colorectal metastasis, was purchased from Chemicon and diluted in

RPMI-1640 for use in ADCC assays with NK cells (CD3−/CD56+/CD16+) that were isolated from PBMC from a healthy volunteer. CEA was added to achieve final concentrations of 2 and10 _μ_g ml−1.

This concentration far exceeds any concentration that would be found in the serum of a colorectal cancer patient. A level above 5 ng ml−1 is generally accepted as being raised. Patients with

CEA levels above 15 ng ml−1 have been found to have a worse prognosis (Wiratkapun et al, 2001). The cytotoxicity assay was then carried out as described above. FACS ANALYSIS FOR COMPETITIVE

INHIBITION OF HPR1A3 BINDING IN THE PRESENCE OF SOLUBLE CEA Soluble CEA at a final concentration of 10 _μ_g ml−1 was added to hPR1A3 (final concentration 20 _μ_g ml−1). The mixture was

incubated in a 1.7 ml eppendorf microtube for 45 min at room temperature. Humanised PR1A3 alone, at a concentration of 20 _μ_g ml−1, and medium alone were similarly incubated in a microtube.

Cells to be assayed with these various mixtures were adjusted to a concentration of 1 × 106 ml−1 in RPMI-1640 complete medium and washed with FACS buffer (PBSA/1%FCS/1% sodium azide). The

resulting pellet was resuspended in 10 _μ_l FACS buffer and then incubated on ice with 100 _μ_l 20 _μ_g ml−1 mPR1A3, hPR1A3/CEA (as prepared above) or medium control for 30 min. The cells

were then washed twice with FACS buffer and incubated on ice in the dark for 30 min with a 1 : 50 dilution of FITC-conjugated anti-human-IgG (for hPR1A3). The resulting labelled cells were

washed again with 2 ml of FACS buffer and centrifuged at 400 G. The supernatant was removed completely and the cells resuspended in 300 _μ_l of 2% paraformadehyde in PBSA. The labelled cells

were analysed using a FACS Calibur flow cytometer as described before. DATA ANALYSIS Percentage lysis of the cell lines in the cytotoxicity assays was calculated as (experimental

release−background release)/(maximum release−background release) × 100. Percentage specific lysis was calculated as (experimental release−antibody independent release)/(maximum

release−antibody independent release) × 100. The s.e.m. of multiple experiments was calculated using Graphpad Prism software, San Diego, CA, USA. Standard normal distribution tests were used

to assess the significance of the differences found. RESULTS HPR1A3 BINDS TO MEMBRANE-BOUND CEA AND THE ANTIBODY–ANTIGEN COMPLEX IS NOT INTERNALISED AFTER 3 H Humanised PR1A3 showed

specific binding to MKN45, a high CEA-expressor. After three hours incubation of MKN45 at 37°C with hPR1A3, no change was observed in the amount of antibody detected on the cell surface

(Figure 2A). This is in contrast to the results obtained in similar experiments using the anti-EpCAM monoclonal antibody, AUA-1 (Figure 2B). In that case there is already a significant

reduction in the cell surface expression of EpCAM after incubation for 1 h at 37°C indicating a fairly rapid internalisation of the surface EpCAM/anti-EpCAM complex. These data show that the

CEA/anti-CEA(PR1A3) complex is not significantly internalised even after 3 h of incubation at 37°C (Figure 2). HPR1A3 CAUSES DOSE-DEPENDENT LYSIS OF THE HIGH CEA-EXPRESSING CELL LINE, MKN45

MKN45 was used as a high CEA expressing cell line in cellular cytotoxicity assays with 51Cr using different antibody concentrations. In the absence of antibody, PBMCs effected a low but

significant level of spontaneous killing. The level of killing increased with increasing concentrations of antibody, as would be expected for ADCC (Figure 3). COLORECTAL CANCER CELL LINES

EXPRESS VARYING LEVELS OF CEA The level of CEA expression on a subset of cell lines was assessed using an ELISA assay with mPR1A3, and is given in Table 1 as mean arbitrary fluorescence

units (±s.d.). These data correlate well with those obtained from FACS analysis, RT–PCR and micro-array expression data (see Supplementary Figure 1). THE LEVEL OF HPR1A3-MEDIATED ADCC

DEPENDS ON THE LEVEL OF CEA EXPRESSION HCT-116 was identified by FACS, ELISA and RT–PCR assays to have no or at most a very low level of CEA expression. The killing of the high expressing

cell line MKN45 was therefore compared with that of HCT-116 by hPR1A3 in the presence of human PBMC using a EuTDA assay. While MKN45 was killed as expected, there was no increase in lysis of

HCT-116 above spontaneous killing even with the highest concentration of antibody (Figure 4A). The variation in hPR1A3-based ADCC lysis between cell lines expressing different levels of CEA

was assessed using PBMC and a EuTDA assay. The results shown in Figure 4B indicate a good correlation between CEA levels and the degree of hPR1A3-mediated killing (Figure 4).

HPR1A3-MEDIATED ADCC-BASED KILLINGS DEPENDS ON THE FC PORTION OF THE ANTIBODY Since the murine IgG1 isoform does not associate strongly with the human Fc_γ_IIIA receptor (Lubeck et al,

1985), mPR1A3 was compared with hPR1A3 in killing assays on MKN45. The results (Figure 5) show that murine PR1A3 did not kill above background, in contrast to hPR1A3. This adds to the

support for the specificity of hPR1A3-based killing and suggests its dependence on appropriate interaction with an Fc_γ_ receptor (Figure 5). HPR1A3-DEPENDENT AND SPONTANEOUS KILLING ARE

BOTH INHIBITED BY AN ANTI-CD16 ANTIBODY, BUT ONLY ANTIBODY-DEPENDENT KILLING IS INHIBITED BY AN F(AB’)2 OF ANTI-CD16 Since the NK effector cells in PBMC, which are presumed to mediate the

majority of antibody-dependent killing, do so via the CD16 (Fc_γ_IIIA) receptor (Titus et al, 1987; Lanier et al, 1988; Moretta et al, 1989), the blocking effects on hPR1A3 killing of MKN45

by anti-CD 16 and a F(ab’)2 of the same antibody were investigated (Figure 6). The results of blocking experiments with two different concentrations of the reagents are shown in Figure 6.

These data show that anti CD16 completely blocked both spontaneous and hPR1A3-induced killing. To ensure that this was a direct effect of blocking the CD16 antibody receptor and not simply a

consequence of adding a second antibody which might compete for Fc-receptor sites, we added a non-specific antibody against prostate membrane-specific antigen instead of anti-CD16 and this

did not block hPR1A3-mediated killing of MKN45. In marked contrast to the blocking effects of whole CD16 antibody, the F(ab’)2 of the anti-CD16, which lacks the Fc-portion of the antibody

required for binding to the CD16 receptor on effector cells, abolished only antibody-dependent killing but did not affect the spontaneous lysis. PURIFIED NK CELLS ARE ABLE TO ELICIT ADCC

WITH HPR1A3 AT MUCH LOWER EFFECTOR:TARGET RATIOS THAN ARE UNFRACTIONATED PBMC To establish that the major killing effect seen with hPR1A3 was actually due to NK cells as conventionally

defined, purified NK cells were used in assays of hPR1A3 killing of the high CEA expressing colorectal cell line, SKCO-1. NK cells were enriched from PBMCs obtained from buffy coat (NBS)

using the Human NK Isolation kit, as described in Materials and Methods. The extent of enrichment for NK cells is illustrated by the FACS analysis shown in Figure 7A, using antibodies to

CD16 and CD56. The second panel clearly documents the extensive purification of the CD16 and CD56 positive NK cells. The results of ADCC assays with hPR1A3 using either these purified NK

cells or unfractionated PBMCs as effectors, and SKCO-1 as target cells, are shown in Figure 7B. The difference in the effectiveness of killing by NK cells as compared to unfractionated PBMCs

is striking. Whereas in both cases there is some spontaneous killing in the absence of antibody, the NK cell antibody-specific killing is already very clear at the lowest effector:target

ratio of 10. There is, on the other hand, barely detectable specific killing with the PMBCs even at a 50 : 1 effector: target ratio. Further data (see Supplementary Figure 2) demonstrate

that ADCC can occur in the presence of purified NK cells even at effector: target ratios as low as 1 : 1. These results provide strong support for the assumption that the hPR1A3

antibody-specific killing of CEA expressing human cell lines is almost entirely due to the action of the human NK cells as effectors. NEITHER HPR1A3 BINDING TO, NOR HPR1A3-MEDIATED KILLING

OF MKN45 ARE BLOCKED BY SOLUBLE CEA The binding to the high expressing cell line SKCO-1 of hPR1A3 (20 _μ_g ml−1) on its own, or after preincubation with soluble CEA (10 _μ_g ml−1) was

investigated using a FACS analysis. The results, illustrated in Figure 8A, show that there is no reduction in reactivity of hPR1A3 with MKN45 after preincubation with CEA, as compared to

hPR1A3 alone. This corroborates for hPR1A3 the earlier results obtained by Durbin et al, 1994 with mPR1A3, which showed that soluble CEA does not block this antibody's binding to

membrane bound CEA (Figure 8). To show the same lack of effect of soluble CEA on hPR1A3-specific killing, ADCC assays were carried out using fresh blood-derived NK cells and SKCO-1 as target

cells with hPR1A3 on its own, or after preincubation with soluble CEA. The results, illustrated in Figure 8B, clearly show that even preincubation with 10 _μ_g ml−1 of soluble CEA does not

reduce the specific killing effect of hPR1A3. DISCUSSION Our study has clearly shown that the humanised PR1A3 antibody can be used for targeted killing of colorectal cancer lines that

express cell surface CEA. Both the direct binding of the antibody to cells, and the extent of its ADCC activity against cells are dependant on the level of surface expression of CEA. The

fact, as we have also shown, that CEA is not significantly internalised adds another advantage to CEA as a target for naked antibody therapy. This effectively increases the exposure times to

Fc_γ_ receptor-bearing cells by promoting attachment to antibody-coated target cells. We have confirmed, as was shown previously for the murine version of PR1A3, that the binding of hPR1A3

to surface bound CEA is not inhibited by soluble CEA, and in addition have shown that the same is true for its ADCC activity. This property of PR1A3 accounts for the low false-positive rate

of lymph node detection in immunoscintigraphy of colorectal cancers with PR1A3 in patients (Granowska et al, 1989), given that it has been shown that soluble CEA drains into lymphatics and

so can become sequestrated into regional lymph nodes in the absence of cancer cells (Kubo et al, 1992). Human NK cells are known to express CD16 (homologous to the Fc_γ_ IV receptor in mice

(Nimmerjahn and Ravetch, 2005)) and are thought to play an important role in responses to antibody therapy (Liljefors et al, 2003). We have shown that the ADCC activity of hPR1A3 is

dependent on its Fc domain. The evidence for this comes both from the fact that mPR1A3, though it binds to surface CEA on human cells, has no ADCC activity using human PBMC as effectors, and

that the ADCC activity of hPR1A3 is blocked both by whole antibody to CD16 and by a CD16 F(ab’)2 fragment. The fact that enrichment of human NK cells can elicit hPR1A3-dependent ADCC at

very low effector:target ratios when compared with unfractionated PBMC, provides strong evidence that the ADCC activity of hPR1A3 is actually mediated by human NK cells. That the addition of

the complete CD16 IgG, (but not the F(ab’)2 fragment) also abolishes antibody independent killing may be explained either by antibody binding to CD16-bearing cells reducing their mobility

in the wells due to formation of clusters of CD16+ cells or that killer cells may destroy CD16+ cells which have bound the antibody. The affinity of hPR1A3 for CEA appears to be relatively

low as reflected by the higher concentrations of PR1A3 needed to mediate ADCC as compared to cetuximab (see Supplementary Figure 2). However, this may be a potential advantage in the

treatment of solid tumours, since higher affinity antibodies may penetrate less into tumours due to the majority of binding taking place at the outer most part of a tumour (Adams et al,

2001). Intermediate affinity antibodies may, thus, be predicted to have greater penetration into solid tumours. Much research is now directed at conjugating antibodies with radiolabels or

toxins. Although this has had some success in experimental animal models, there remain significant problems in the true _in vivo_ situation, including especially the development of an immune

response against the toxins or enzymes linked to a therapeutic antibody. We suggest that the appropriateness of CEA as a therapeutic target, together with our evaluation of antibody

hPR1A3's mediated ADCC activity makes this antibody a very attractive target for clinical development as a naked antibody. The main challenge may be to enhance PR1A3's ADCC

activity, and this may be achieved by glycoengineering its Fc hinge region (Umana et al, 1999), which has been shown to be a very effective method for enhancing the effectiveness of

antibody-mediated ADCC _in vitro_. As previously discussed, only a small percentage of antibody administered intravenously actually reaches the cells of a solid tumour ((Allum et al, 1986;

Delaloye et al, 1986; Epenetos et al, 1986; Colcher et al, 1987; Welt et al, 1990). While a small number of antibody molecules reaching their tumour target may be sufficient to elicit

immune-based killing by ADCC, it seems unlikely that such small amounts of antibody reaching a tumour could have much effect in blocking function, since this would require at least the

majority of the antibody's targets to be covered. This emphasises the potential importance of immune mechanisms, even for therapy with antibodies against targets such as EGFR and ErbB

with known functions, and so the importance of enhancing ADCC for effective treatment, rather than improving the blocking of function. The fact that CEA has no obvious function that might be

blocked by antibody does not mitigate against its use for naked antibody-based therapy on the assumption that the primary mechanism is immune and through ADCC. We believe that the results

we have presented here suggest that the naked anti-CEA humanised antibody PR1A3, glycoengineered to increase its efficacy in ADCC, may be an excellent candidate for therapy of colorectal and

other solid tumours that express significant levels of CEA. CHANGE HISTORY * _ 16 NOVEMBER 2011 This paper was modified 12 months after initial publication to switch to Creative Commons

licence terms, as noted at publication _ REFERENCES * Adams GP, Schier R, McCall AM, Simmons HH, Horak EM, Alpaugh RK, Marks JD, Weiner LM (2001) High affinity restricts the localization and

tumor penetration of single-chain fv antibody molecules. _Cancer Res_ 61: 4750–4755 CAS  PubMed  Google Scholar  * Allum WH, MacDonald F, Anderson P, Fielding JW (1986) Localisation of

gastrointestinal cancer with a 131 I labelled monoclonal antibody to CEA. _Br J Cancer_ 53: 203–210 Article  CAS  Google Scholar  * Arnould L, Gelly M, Penault-Llorca F, Benoit L, Bonnetain

F, Migeon C, Cabaret V, Fermeaux V, Bertheau P, Garnier J, Jeannin JF, Coudert B (2006) Trastuzumab-based treatment of HER2-positive breast cancer: an antibody-dependent cellular

cytotoxicity mechanism? _Br J Cancer_ 94: 259–267 Article  CAS  Google Scholar  * Berinstein NL (2002) Carcinoembryonic antigen as a target for therapeutic anticancer vaccines: a review. _J

Clin Oncol_ 20: 2197–2207 Article  CAS  Google Scholar  * Blomberg K, Hautala R, Lovgren J, Mukkala VM, Lindqvist C, Akerman K (1996) Time-resolved fluorometric assay for natural killer

activity using target cells labelled with a fluorescence enhancing ligand. _J Immunol Methods_ 193: 199–206 Article  CAS  Google Scholar  * Boyum A (1968) Separation of leukocytes from blood

and bone marrow. Introduction. _Scand J Clin Lab Invest Suppl_ 97: 7 CAS  PubMed  Google Scholar  * Brattain MG, Fine WD, Khaled FM, Thompson J, Brattain DE (1981) Heterogeneity of

malignant cells from a human colonic carcinoma. _Cancer Res_ 41: 1751–1756 CAS  PubMed  Google Scholar  * Browning MJ, Krausa P, Rowan A, Hill AB, Bicknell DC, Bodmer JG, Bodmer WF (1993)

Loss of human leukocyte antigen expression on colorectal tumor cell lines: implications for anti-tumor immunity and immunotherapy. _J Immunother_ 14: 163–168 Article  CAS  Google Scholar  *

Carter PJ (2006) Potent antibody therapeutics by design. _Nat Rev Immunol_ 6: 343–357 Article  CAS  Google Scholar  * Chan CH, Stanners CP (2007) Recent advances in the tumour biology of the

GPI-anchored carcinoembryonic antigen family members CEACAM5 and CEACAM6. _Curr Oncol_ 14: 70–73 Article  CAS  Google Scholar  * Chau I, Allen MJ, Cunningham D, Norman AR, Brown G, Ford HE,

Tebbutt N, Tait D, Hill M, Ross PJ, Oates J (2004) The value of routine serum carcino-embryonic antigen measurement and computed tomography in the surveillance of patients after adjuvant

chemotherapy for colorectal cancer. _J Clin Oncol_ 22: 1420–1429 Article  CAS  Google Scholar  * Clynes R (2006) Antitumor antibodies in the treatment of cancer: Fc receptors link opsonic

antibody with cellular immunity. _Hematol Oncol Clin North Am_ 20: 585–612 Article  Google Scholar  * Clynes R, Takechi Y, Moroi Y, Houghton A, Ravetch JV (1998) Fc receptors are required in

passive and active immunity to melanoma. _Proc Natl Acad Sci USA_ 95: 652–656 Article  CAS  Google Scholar  * Colcher D, Esteban JM, Carrasquillo JA, Sugarbaker P, Reynolds JC, Bryant G,

Larson SM, Schlom J (1987) Quantitative analyses of selective radiolabeled monoclonal antibody localization in metastatic lesions of colorectal cancer patients. _Cancer Res_ 47: 1185–1189

CAS  PubMed  Google Scholar  * Delaloye B, Bischof-Delaloye A, Buchegger F, von Fliedner V, Grob JP, Volant JC, Pettavel J, Mach JP (1986) Detection of colorectal carcinoma by

emission-computerized tomography after injection of 123I-labeled Fab or F(ab’)2 fragments from monoclonal anti-carcinoembryonic antigen antibodies. _J Clin Invest_ 77: 301–311 Article  CAS 

Google Scholar  * Durbin H, Bodmer WF (1987) A sensitive micro-immunoassay using beta-galactosidase/anti-beta-galactosidase complexes. _J Immunol Methods_ 97: 19–27 Article  CAS  Google

Scholar  * Durbin H, Young S, Stewart LM, Wrba F, Rowan AJ, Snary D, Bodmer WF (1994) An epitope on carcinoembryonic antigen defined by the clinically relevant antibody PR1A3. _Proc Natl

Acad Sci USA_ 91 (10): 4313–4317 Article  CAS  Google Scholar  * Epenetos AA, Canti G, Taylor-Papadimitriou J, Curling M, Bodmer WF (1982) Use of two epithelium-specific monoclonal

antibodies for diagnosis of malignancy in serous effusions. _Lancet_ 2: 1004–1006 Article  CAS  Google Scholar  * Epenetos AA, Snook D, Durbin H, Johnson PM, Taylor-Papadimitriou J (1986)

Limitations of radiolabeled monoclonal antibodies for localization of human neoplasms. _Cancer Res_ 46: 3183–3191 CAS  PubMed  Google Scholar  * Fogh J, Trempe G (1975) _Human Tumour Cells

In Vitro_. Plenum Press: New York Book  Google Scholar  * Gold P, Freedman SO (1965) Demonstration of tumor-specific antigens in human colonic carcinomata by immunological tolerance and

absorption techniques. _J Exp Med_ 121: 439–462 Article  CAS  Google Scholar  * Granowska M, Jass JR, Britton KE, Northover JM (1989) A prospective study of the use of 111In-labelled

monoclonal antibody against carcino-embryonic antigen in colorectal cancer and of some biological factors affecting its uptake. _Int J Colorectal Dis_ 4: 97–108 Article  CAS  Google Scholar

  * Hammarstrom S (1999) The carcinoembryonic antigen (CEA) family: structures, suggested functions and expression in normal and malignant tissues. _Semin Cancer Biol_ 9: 67–81 Article  CAS

  Google Scholar  * Kubo A, Nakamura K, Katayama M, Hashimoto S, Teramoto T, Kodaira S (1992) Pharmacokinetic analysis of antibody localization in human colon cancer: comparison with

immunoscintigraphy. _Ann Nucl Med_ 6: 21–27 Article  CAS  Google Scholar  * Lanier LL, Ruitenberg JJ, Phillips JH (1988) Functional and biochemical analysis of CD16 antigen on natural killer

cells and granulocytes. _J Immunol_ 141: 3478–3485 CAS  PubMed  Google Scholar  * Liersch T, Meller J, Bittrich M, Kulle B, Becker H, Goldenberg DM (2007) Update of carcinoembryonic antigen

radioimmunotherapy with (131)i-labetuzumab after salvage resection of colorectal liver metastases: comparison of outcome to a contemporaneous control group. _Ann Surg Oncol_ 14: 2577–2590

Article  Google Scholar  * Liljefors M, Nilsson B, Hjelm Skog AL, Ragnhammar P, Mellstedt H, Frodin JE (2003) Natural killer (NK) cell function is a strong prognostic factor in colorectal

carcinoma patients treated with the monoclonal antibody 17-1A. _Int J Cancer_ 105: 717–723 Article  CAS  Google Scholar  * Lubeck MD, Steplewski Z, Baglia F, Klein MH, Dorrington KJ,

Koprowski H (1985) The interaction of murine IgG subclass proteins with human monocyte Fc receptors. _J Immunol_ 135: 1299–1304 CAS  PubMed  Google Scholar  * Moretta A, Tambussi G, Ciccone

E, Pende D, Melioli G, Moretta L (1989) CD16 surface molecules regulate the cytolytic function of CD3CD16+ human natural killer cells. _Int J Cancer_ 44: 727–730 Article  CAS  Google Scholar

  * Motoyama T, Hojo H, Suzuki T, Oboshi S (1979) Evaluation of the regrowth assay method as an _in vitro_ drug sensitivity test and its application to cultuered human gastric cancer cell

lines. _Acta Medica et Biologica_ 27: 49–63 CAS  Google Scholar  * Nimmerjahn F, Ravetch JV (2005) Divergent immunoglobulin g subclass activity through selective Fc receptor binding.

_Science_ 310: 1510–1512 Article  CAS  Google Scholar  * Olszewski AJ, Grossbard ML (2004) Empowering targeted therapy: lessons from rituximab. _Sci STKE_ 2004: pe30 PubMed  Google Scholar 

* Paraskeva C, Buckle BG, Sheer D, Wigley CB (1984) The isolation and characterization of colorectal epithelial cell lines at different stages in malignant transformation from familial

polyposis coli patients. _Int J Cancer_ 34: 49–56 Article  CAS  Google Scholar  * Reichert JM, Valge-Archer VE (2007) Development trends for monoclonal antibody cancer therapeutics. _Nat Rev

Drug Discov_ 6: 349–356 Article  CAS  Google Scholar  * Richman PI, Bodmer WF (1987) Monoclonal antibodies to human colorectal epithelium: markers for differentiation and tumour

characterization. _Int J Cancer_ 39: 317–328 Article  CAS  Google Scholar  * Solic N, Collins JE, Richter A, Holt SJ, Campbell I, Alexander P, Davies DE (1995) Two newly established cell

lines derived from the same colonic adenocarcinoma exhibit differences in EGF-receptor ligand and adhesion molecule expression. _Int J Cancer_ 62: 48–57 Article  CAS  Google Scholar  * Spurr

NK, Durbin H, Sheer D, Parkar M, Bobrow L, Bodmer WF (1986) Characterization and chromosomal assignment of a human cell surface antigen defined by the monoclonal antibody AUAI. _Int J

Cancer_ 38: 631–636 Article  CAS  Google Scholar  * Stewart LM, Young S, Watson G, Mather SJ, Bates PA, Band HA, Wilkinson RW, Ross EL, Snary D (1999) Humanisation and characterisation of

PR1A3, a monoclonal antibody specific for cell-bound carcinoembryonic antigen. _Cancer Immunol Immunother_ 47: 299–306 Article  CAS  Google Scholar  * Thompson JA, Grunert F, Zimmermann W

(1991) Carcinoembryonic antigen gene family: molecular biology and clinical perspectives. _J Clin Lab Anal_ 5: 344–366 Article  CAS  Google Scholar  * Titus JA, Perez P, Kaubisch A, Garrido

MA, Segal DM (1987) Human K/natural killer cells targeted with hetero-cross-linked antibodies specifically lyse tumor cells _in vitro_ and prevent tumor growth _in vivo_. _J Immunol_ 139:

3153–3158 CAS  PubMed  Google Scholar  * Tom BH, Rutzky LP, Jakstys MM, Oyasu R, Kaye CI, Kahan BD (1976) Human colonic adenocarcinoma cells. I. Establishment and description of a new line.

_In vitro_ 12: 180–191 Article  CAS  Google Scholar  * Umana P, Jean-Mairet J, Moudry R, Amstutz H, Bailey JE (1999) Engineered glycoforms of an antineuroblastoma IgG1 with optimized

antibody-dependent cellular cytotoxic activity. _Nat Biotechnol_ 17: 176–180 Article  CAS  Google Scholar  * Watkins JF, Sanger C (1977) Properties of a cell line from human adenocarcinoma

of the rectum. _Br J Cancer_ 35: 785–794 Article  CAS  Google Scholar  * Welt S, Divgi CR, Real FX, Yeh SD, Garin-Chesa P, Finstad CL, Sakamoto J, Cohen A, Sigurdson ER, Kemeny N (1990)

Quantitative analysis of antibody localization in human metastatic colon cancer: a phase I study of monoclonal antibody A33. _J Clin Oncol_ 8: 1894–1906 Article  CAS  Google Scholar  *

Wiratkapun S, Kraemer M, Seow-Choen F, Ho YH, Eu KW (2001) High preoperative serum carcinoembryonic antigen predicts metastatic recurrence in potentially curative colonic cancer: results of

a five-year study. _Dis Colon Rectum_ 44: 231–235 Article  CAS  Google Scholar  * Wong JY, Chu DZ, Williams LE, Yamauchi DM, Ikle DN, Kwok CS, Liu A, Wilczynski S, Colcher D, Yazaki PJ,

Shively JE, Wu AM, Raubitschek AA (2004) Pilot trial evaluating an 123I-labeled 80-kilodalton engineered anticarcinoembryonic antigen antibody fragment (cT84.66 minibody) in patients with

colorectal cancer. _Clin Cancer Res_ 10: 5014–5021 Article  CAS  Google Scholar  Download references ACKNOWLEDGEMENTS PJC was supported by generous grants from the CORE charity and the

Jacqueline Seroussi Memorial Foundation for Cancer Research. SQA was supported by a Bobby Moore Research Fellowship, CRUK. MGT was supported by generous grants from the Royal College of

Surgeons of England and a grant from the John Radcliffe Charitable Trust. The overall work in the laboratory is funded by a Cancer Research UK programme grant to WFB. We thank Sylvia

Bartlett and Rose Dorupi for their help and support. AUTHOR INFORMATION Author notes * P J Conaghan and S Q Ashraf: These authors contributed equally to this work. AUTHORS AND AFFILIATIONS *

Cancer Research UK, Cancer & Immunogenetics Laboratory, Weatherall Institute of Molecular Medicine, Headington, Oxford, UK P J Conaghan, S Q Ashraf, M G Tytherleigh, J L Wilding, E

Tchilian, D Bicknell & W F Bodmer * Department of Colorectal Surgery, John Radcliffe Hospital, Headington, Oxford, UK P J Conaghan, S Q Ashraf, M G Tytherleigh & N JMcC Mortensen

Authors * P J Conaghan View author publications You can also search for this author inPubMed Google Scholar * S Q Ashraf View author publications You can also search for this author inPubMed

 Google Scholar * M G Tytherleigh View author publications You can also search for this author inPubMed Google Scholar * J L Wilding View author publications You can also search for this

author inPubMed Google Scholar * E Tchilian View author publications You can also search for this author inPubMed Google Scholar * D Bicknell View author publications You can also search for

this author inPubMed Google Scholar * N JMcC Mortensen View author publications You can also search for this author inPubMed Google Scholar * W F Bodmer View author publications You can

also search for this author inPubMed Google Scholar CORRESPONDING AUTHOR Correspondence to W F Bodmer. ADDITIONAL INFORMATION Supplementary Information accompanies the paper on British

Journal of Cancer website (http://www.nature.com/bjc) SUPPLEMENTARY INFORMATION SUPPLEMENTARY INFORMATION (DOC 134 KB) RIGHTS AND PERMISSIONS From twelve months after its original

publication, this work is licensed under the Creative Commons Attribution-NonCommercial-Share Alike 3.0 Unported License. To view a copy of this license, visit

http://creativecommons.org/licenses/by-nc-sa/3.0/ Reprints and permissions ABOUT THIS ARTICLE CITE THIS ARTICLE Conaghan, P., Ashraf, S., Tytherleigh, M. _et al._ Targeted killing of

colorectal cancer cell lines by a humanised IgG1 monoclonal antibody that binds to membrane-bound carcinoembryonic antigen. _Br J Cancer_ 98, 1217–1225 (2008).

https://doi.org/10.1038/sj.bjc.6604289 Download citation * Revised: 21 January 2008 * Accepted: 31 January 2008 * Published: 18 March 2008 * Issue Date: 08 April 2008 * DOI:

https://doi.org/10.1038/sj.bjc.6604289 SHARE THIS ARTICLE Anyone you share the following link with will be able to read this content: Get shareable link Sorry, a shareable link is not

currently available for this article. Copy to clipboard Provided by the Springer Nature SharedIt content-sharing initiative KEYWORDS * PR1A3 * CEA * ADCC * colorectal cancer