• adenocarcinoma;
  • cancer vaccine;
  • conjugate vaccine;
  • KH-1;
  • Lewisy synthetic antigen


  1. Top of page
  2. Abstract
  6. Acknowledgements

There is mounting evidence to suggest that immunization-based strategies can be used to mobilize the human immune system against specific carbohydrate antigens displayed on the surface of cancer cells. Following isolation and identification, such antigens can be administered as conjugate vaccines. The tumor-associated carbohydrate antigen KH-1 is 1 such antigen and may serve as a potential target for immunization against adenocarcinoma. However, a serious impediment to the application of a vaccine-based approach involving this antigen is that its availability from natural sources is severely limited. In order to overcome this limitation, we have developed an efficient total synthesis of this complex glycolipid. We have extended our synthesis to reach a structurally related analog in which the ceramide portion of KH-1 is replaced with an allyl substituent. These synthetic advances have led to the preparation of 2 potential vaccine constructs, each based on the conjugation of the KH-1 nonasaccharide and the carrier protein keyhole limpet hemocyanin (KLH). In 1 construct (KH-1-Et-KLH), the nonasaccharide is conjugated to KLH via a simple ethyl linkage, while in the other (KH-1-MMCCH-KLH), conjugation is mediated by a 4-(4-N-maleimidomethyl)cyclohexane-1-carboxyl hydrazide (MMCCH) cross-linker. We report here the immunological properties of these 2 constructs. Mice were immunized with either of the 2 KH-1-KLH vaccine candidates or the KH-1 ceramide, along with the immunological adjuvant QS-21. Immunization with the ceramide served as a negative control and, as expected, failed to stimulate the production of antibodies against the KH-1 glycolipid. The construct in which the KH-1 nonasaccharide is linked to KLH via a simple alkyl chain stimulated significant quantities of IgM antibodies, whereas the construct linked to KLH by MMCCH induced high titers of both IgM and IgG antibodies. Inhibition data demonstrated that antibodies generated in response to immunization with the KH-1-KLH constructs recognize not only the KH-1 antigen but also the Lewisy (Ley) antigen, which, from a structural perspective, is similar to the 4 residues located at the non-reducing end of the KH-1 nonasaccharide. Thus, the KH-1-KLH constructs elicit an immune response that successfully targets 2 adenocarcinoma markers. As assessed by FACS analysis, the antibodies raised were strongly reactive with the KH-1/Ley positive cell line MCF-7 but not with KH-1 and Ley negative melanoma cell lines. Based on the results of our study, a KH-1-KLH plus QS-21 vaccine is being prepared for clinical evaluation. © 2002 Wiley-Liss, Inc.

The development of vaccines based on tumor-associated carbohydrate antigens holds considerable promise for the treatment of metastatic disease.1, 2 Antibody targeting of carbohydrates on the surface of tumor cells is mechanistically similar to antibody targeting of glycoconjugates on the surface of bacterial or viral pathogens as a means to prevent infection. In order to develop carbohydrate vaccines the target antigen must be available for immunological evaluation. However, this line of investigation has been limited due to complexities associated with the isolation of tumor antigens from natural sources. In order to overcome this limitation, we rely on total chemical synthesis to obtain these cell surface antigens. To significantly impact vaccine development, synthetic routes to tumor antigens must proceed in sufficiently high yield to support pre-clinical and, potentially, clinical evaluation. The synthetic approach also benefits from the possibility of providing the antigen in a form that will allow facile construction of conjugate vaccines, in addition to providing a platform for addressing complex issues related to chemical synthesis. Following these guidelines, we have advanced several synthetic and semi-synthetic conjugate vaccines to clinical evaluation.1 KH-1 (1, Fig. 1) is the most structurally complex tumor antigen that has been prepared by total synthesis. Our route (see Fig. 1) to this target is highly efficient and has provided us with quantities of material sufficient for immunological evaluation.

thumbnail image

Figure 1. The logic of glycal assembly as applied to the synthesis of the KH-1 antigen.

Download figure to PowerPoint

KH-1 antigen was originally identified with monoclonal antibodies AH6 and has been shown to be over-expressed in human adenocarcinomas.3–6 Structural studies of the isolate revealed a novel species belonging to the Ley series of glycolipids that contained epitopes corresponding to both the Ley tetrasaccharide and the Lex trisaccharide (Fucα1[RIGHTWARDS ARROW]2Galβ1[RIGHTWARDS ARROW]4(Fucα1[RIGHTWARDS ARROW]3)GlcNAcβ1[RIGHTWARDS ARROW] 3Galβ1[RIGHTWARDS ARROW]4(Fucα1[RIGHTWARDS ARROW]3)GlcNAcβ1[RIGHTWARDS ARROW]3Galβ1[RIGHTWARDS ARROW]4Glcβ1[RIGHTWARDS ARROW]1Cer). Over expression of KH-1 by a variety of human cancers makes it an attractive target for active immunotherapy, but progress has been hampered by the lack of suitable quantities of KH-1 for vaccine preparation. Current advances in the synthesis of complex carbohydrates has enabled the total synthesis of the KH-1 nonasaccharide7–9 and, consequently, has permitted the fashioning of KH-1 vaccine candidates and allowed, for the first time, examination of the immunological properties of these constructs in experimental animals.


  1. Top of page
  2. Abstract
  6. Acknowledgements

Antigens and reagents

Protected KH-1 nonasaccharide (13, Fig. 1) was prepared by total synthesis using the logic of glycal assembly.7, 8 Allyl group introduction at the reducing end of the nonasaccharide, followed by global deprotection (not shown in Fig. 1), then gave allyl glycoside 14 (Fig. 2). Incorporation of the allyl substituent facilitated conjugation with protein carriers according to the procedure outlined in Figure 2. Keyhole limpet hemocyanin (KLH) was obtained from PerImmune, Inc. (Rockville, MD). Bovine serum albumin (BSA), sodium cyanoborohydride and human serum complement were obtained from Sigma Chemical Co. (St. Louis, MO). Monoclonal antibody BR96 (MAb BR96) was kindly provided by Dr. I. Hellstrom.6 Goat anti-mouse IgG and IgM, conjugated with alkaline phosphatase and goat anti-mouse IgM-fluorescence-isothiocyanate (FITC) were obtained from Southern Biotechnology Associates, Inc. (Birmingham, AL).

thumbnail image

Figure 2. Conjugation strategy used for the preparation of the vaccine constructs 16 and 18.

Download figure to PowerPoint

Animals and adjuvants

Six-week-old female CB6F1 mice were obtained from the Jackson Laboratory (Bar Harbor, Maine). Adjuvant QS-21, a purified saponin fraction,10 was obtained from Aquila Biopharmaceuticals, Inc. (Framingham, MA), now known as Antigenics Inc. (New York, NY).

Vaccine preparation

KH-1-Et-KLH (16) and KH-1-MMCCH [4-(4-N-maleimidomethyl)cyclohexane-1-carboxyl hydrazide]-KLH (18) conjugates (Fig. 2) were prepared as described previously for STn(c)-KLH conjugate vaccine.11, 12 The KH-1 KLH conjugation procedure is initiated by ozonolytic cleavage of the allylic double bond of the KH-1 allyl glycoside (14 [RIGHTWARDS ARROW] 15, Fig. 2). The resulting aldehyde (15) is then reacted with native KLH in the presence of sodium cyanoborohydride, resulting in reductive amination occurring between the ϵ-amino groups of lysine on the carrier protein and the aldehyde substituent of the oxidized KH-1 allyl glycoside (15 [RIGHTWARDS ARROW] 16, Fig. 2).

The KH-1-MMCCH-KLH conjugate (18) was prepared as previously described11 using a recently developed bifunctional cross-linker MMCCH method, which is outlined as follows. The aldehyde group generated by ozonolysis of the KH-1 allyl glycoside is first condensed with the hydrazide functional group of MMCCH (15 [RIGHTWARDS ARROW] 17) and the maleimide group of the resulting species is then reacted with the sulfhydryl groups on thiolated KLH (17 [RIGHTWARDS ARROW] 18), which is obtained by treatment of KLH with 2-iminothiolane. The KH-1/KLH molar ratio was determined by the high-pH anion-exchange chromatography with pulsed amperometric detection (HPAEC-PAD) assay for carbohydrate after acid hydrolysis and Bio-Rad assay for protein. The KH-1-KLH epitope ratio of KH-1-Et-KLH and KH-1-MMCCH-KLH was 141:1 and 688:1, respectively.


Groups of 5 mice were vaccinated with PBS, 10 μg of KH-1 ceramide alone or with KH-1-Et-KLH or KH-1-MMCCH-KLH vaccines containing 3 μg of KH-1 nonasaccharide plus 10 μg QS-21. Vaccines were administered subcutaneously 3 times at 1 week intervals and 2 more times at 10 week intervals.


ELISAs were performed as detailed previously for globo-H.13 Anti-KH-1 antibody titers were measured by coating ELISA plates with KH-1 ceramide at 0.2 μg/well. Serially diluted antisera were then added to wells and incubated for 1 hr at room temperature. Goat anti-mouse IgM or IgG conjugated with alkaline phosphatase were used as secondary antibodies. The antibody titer was defined as the highest serum dilution showing an absorbance of 0.1 or greater over that of normal mouse sera.

Flow cytometry

Cells from the KH-1 positive breast cancer cell line MCF-7 served as targets.14 The cells were incubated on ice with 20 μl of 1:20 diluted pre-vaccination sera, antisera or MAb BR96 (5 μg/ml) for 30 min. After washing, 20 μl of 1:15 goat anti-mouse IgM labeled with fluorescein-isothiocyanate (FITC, Southern biotechnology Associates, Inc., Birmingham, AL) was added, mixed and incubated for 30 min on ice. After washing, the positive population and mean fluorescence intensity of the stained cells were analyzed by Flow Cytometry (FACScan, Becton and Dickinson, San Jose, CA).15

Inhibition assays

Suitably diluted (1:50–1:100) antisera was mixed with various concentrations of structurally related and unrelated carbohydrate antigens (Table I). The mixture was incubated at room temperature for 30 min and transferred to an ELISA plate coated with KH-1-ceramide. ELISAs were performed as described above. Percentage inhibition was calculated based on the difference in absorbance between the serum without and with the potential inhibitor present: % Inhibition = (Absorbance of serum without inhibitor − Absorbance of serum with inhibitor) /Absorbance of serum without inhibitor.

Table I. Structure of Carbohydrate Antigens Used for Inhibition Assays
  • 1

    Synthetic carbohydrates.



  1. Top of page
  2. Abstract
  6. Acknowledgements

Antibody response to vaccination with KH-1 and KH-1-KLH conjugates

IgM and IgG antibodies against KH-1 were not detected in sera of mice pre-vaccination nor in the sera of mice from control groups. ELISA titers of all immunized mice are listed in Table II. Immunization with 10 μg of KH-1 ceramide plus 10 μg QS-21 did not induce detectable amounts of IgM or IgG antibodies against KH-1. In contrast to the nonconjugated form, immunization with the KH-1-Et-KLH construct or the KH-1-MMCCH-KLH construct along with QS-21 stimulated the production of antibodies in all mice. In mice immunized with KH-1-Et-KLH/QS-21, only IgM antibodies were detected, whereas in those immunized with KH-1-MMCCH-KLH/QS-21 both IgM and IgG antibodies were observed. Significantly, the IgM titers that resulted from the KH-1-MMCCH-KLH construct were substantially higher than those of the KH-1-Et-KLH construct.

Table II. Serological Analysis by ELISA and FACS Using SERA Obtained from Mice Immunized with KH-1 Vaccines
MouseELISA1FACS (% positive)2
Pre-vaccination sera10 days after 3rd vaccination10 days after 3rd vaccination
IgMIgGIgMIgGIgM (%)IgG (%)
  • 1

    ELISA plates were coated with 0.2 μg of KH-1 ceramide/well.

  • 2

    FACS was conducted with MCF-7 cells. Mouse sera used for FACS were 1:10 dilution. 63.3% positive reaction was observed with anti-Ley BR-96 (5 μg/ml) at 1:10 dilution by FACS analysis.

Vaccine: KH-1 ceramide (1) plus QS-21
Vaccine: KH-1-Et-KLH conjugate (16) plus QS-21
Vaccine: KH-1-MMCCH-KLH conjugate (18) plus QS-21

Sera from mice vaccinated with the KH-1-KLH conjugates along with QS-21 also exhibited a strong reactivity with KH-1-expressing MCF-7 breast cancer cells, as determined by flow cytometry (Table II). Sera from those mice immunized with KH-1 ceramide along with QS-21 exhibited no significant reactivity.

Inhibition assays

The inhibition results for IgM and IgG antibodies from 5 mice that received the MMCCH vaccine are shown in Tables III. The inhibition of antibody reactivity with structurally related and unrelated carbohydrate antigens for serum from 1 mouse is shown in Figures 3 and 4. The results indicate that neither the KH-1 allyl glycoside nor the Ley allyl glycoside inhibit, on average, greater than 70% of IgM reactivity with KH-1 ceramide, whereas IgG antibody reactivity was essentially completely inhibited by both KH-1 allyl glycoside and Ley allyl glycoside. Less than 10% inhibition of both IgM and IgG reactivity with KH-1ceramide was seen with the other structurally related and unrelated carbohydrates. The lower level of inhibition observed for the IgM antibodies by KH-1 allyl glycoside and Ley allyl glycoside, relative to that observed in the case of IgG antibodies, may be a reflection of the inherent differences in receptor affinity for the 2 classes of antibodies. The 10 lower affinity binding sites on IgM antibodies may preferentially bind to KH-1 ceramide displayed in high density on the ELISA plate.

Table III. Inhibition of ELISA-Based Reactivity with KH-1 Ceramide Using Immune SERA After Vaccination with KH-1-MMCCH-KLH Plus QS-211
MousePotential inhibitor
KH-1 allyl glycosideLey allyl glycosideLexLeaLebLeb-LexLex-Lex
  • 1

    Grading scale: Percent inhibition of 85–90% was given a grade of ++++, 70–85% +++, 40–70% ++, 20–40% + and below 20% −.

thumbnail image

Figure 3. Inhibition of IgM reactivity with KH-1 ceramide using sera after vaccination with KH-1-MMCCH-KLH plus QS-21.

Download figure to PowerPoint

thumbnail image

Figure 4. Inhibition of IgG reactivity with KH-1 ceramide using sera after vaccination with KH-1-MMCCH-KLH plus QS-21.

Download figure to PowerPoint


  1. Top of page
  2. Abstract
  6. Acknowledgements

Tumor cells are often characterized by the over-expression of various cell-surface carbohydrate-based molecules. Of these, glycolipids are especially tempting targets for eliciting an antibody-mediated immune response since they are intimately associated with the cell surface lipid bilayer, which could potentially facilitate cell lysis by a variety of mechanisms. Unfortunately, application of an immunologically based approach of this type is severely hampered by the fact that isolation of the appropriate glycolipids in sufficient quantities for vaccine production is not generally possible. Complicating matters even further is the fact that, from a structural perspective, the molecules in question are typically highly complex and, as a result, attempts to access them via total synthesis is an extremely challenging undertaking. KH-1 is a prime example of such a molecule and serves to clearly emphasize these points; it is a structurally complex cell surface glycolipid, available in only minute quantities from natural sources. We have recently overcome limitations related to the availability of KH-1, however, by developing a total synthesis of it, based on our glycal assembly methodology.7, 8 Through an extension of our synthetic work, we were able to prepare 2 related constructs conjugated to KLH (16, 18, Fig. 2) for evaluation as potential cancer vaccines. Using these compounds, we have demonstrated that, upon co-administration with QS-21, an appropriate antibody response is evoked. The most promising of the 2 candidates (18) examined in our study resulted in the production of both IgM and IgG antibodies, which were shown to be strongly reactive with KH-1 positive MCF7 breast cancer cells.

Contained within the KH-1 nonasaccharide is a terminally situated epitope that corresponds to Ley and an internally located 1 corresponding to Lex. Not surprisingly, therefore, is the fact that KH-1 is recognized by monoclonal antibodies against both Ley and Lex. While Ley and Lex have a normal tissue distribution largely restricted to the secretory borders of a variety of epithelial tissues (a location apparently poorly accessible to the immune system), Lex is also expressed on peripheral blood polymorphonucleocytes.16, 17 From the outset of our investigation, we were wary of the possibility that antibody production in direct response to the Lex epitope of KH-1 would be detrimental to any immuno-therapeutic approach that involved KH-1. Fortunately, as demonstrated here through inhibition studies involving Ley allyl glycoside, Lex and anti-sera raised in response to our synthetic KH-1-based constructs, it appears that virtually all of the IgG antibodies and the majority of the IgM antibodies induced, were against the terminal Ley epitope. No significant response to the internal Lex epitope was observed. This finding is consistent with the well-established paradigm that anti-carbohydrate antibodies react primarily with the terminal residues of oligosaccharides.18

The final issue addressed in our study is how the nature of antigen conjugation to KLH and vaccine formulation affects antibody production. Consistent with our prior experience involving other glycolipid antigens13, 17, 18 was that little, if any, antibody production was induced by immunization with a mixture consisting of KH-1 ceramide and QS-21 adjuvant. We have demonstrated previously that this is also the case when KLH is included in the vaccine formulation.19 The vaccine construct in which the KH-1 nonasaccharide was linked to KLH via an ethyl linkage (16) resulted in a relatively low level of IgM antibody production against KH-1 and produced no detectible IgG antibodies. However, the construct that resulted from conjugation of KLH and the nonasaccharide via the MMCCH linker (18) resulted in high titers of both IgM and IgG antibodies and these displayed strong reactivity with KH-1 positive MCF7 breast cancer cells. There are 2 readily identifiable factors, which may contribute to the difference in immunogenicity observed for the 2 vaccine constructs. First, the structure of the linkage that connects the nonasaccharide and KLH is considerably different and, second, there is a difference in the relative amount of nonasaccharide to KLH for each of the constructs. Immunization with each conjugate vaccine was conducted such that approximately the same amount (3 μg) of the KH-1 nonasaccharide was administered; however, the ratio of nonasaccharide to KLH was 688:1 for the KH-1-MMCCH-KLH construct but only 141:1 for the KH-1-Et-KLH construct. (The ratio difference is a reflection of the efficiency of the conjugation procedure used.) This correlation between vaccine immunogenicity and the ratio of antigen molecules to carrier molecules is consistent with our own previous experience (unpublished observations), as well as that of others.20, 21

Based on the studies described above, we are preparing to conduct a phase 1 clinical trial with patients having epithelial cancers, using the KH-1-MMCCH-KLH vaccine construct along with QS-21 adjuvant. This construct not only represents the most complex synthetically derived carbohydrate-based conjugate vaccine examined to date but also exhibits the highly beneficial feature of eliciting an immune response capable of targeting 2 distinct adenocarcinoma markers.


  1. Top of page
  2. Abstract
  6. Acknowledgements

We thank Mr. C. Kandell for preparing Figures 3 and 4. Graduate fellowship support is gratefully acknowledged by H.M.K. (U.S. Army Department of Defense Breast Cancer Graduate Fellowship: DAMD 17-97-1-7119). Postdoctoral fellowship support is gratefully acknowledged by D.M.C. (Natural Science and Engineering Research Council of Canada: PDF-230654-2000 and Alberta Heritage Foundation for Medical Research: 199901330) and L.W.J. (National Institute of Health: F32CA79120).


  1. Top of page
  2. Abstract
  6. Acknowledgements
  • 1
    Danishefsky SJ, Allen JR. From the laboratory to the clinic: A retrospective on fully synthetic carbohydrate-based anticancer vaccines. Angew Chem Int Ed Engl 2000;39: 83663.
  • 2
    Ragupathi G. Carbohydrate antigens as targets for active specific immunotherapy. Cancer Immunol Immunother 1996;43: 1527.
  • 3
    Kaizu T, Levery SB, Nudelman E, et al. 1986. Novel fucolipids of human adenocarcinoma: monoclonal antibody specific for trifucosyl Ley (III3FucV3FucVI2FucnLc6) and a possible three-dimensional epitope structure. J Biol Chem 1986;261: 112548.
  • 4
    Nudelman E, Levery SB, Kaizu T, et al. 1986. Novel fucolipids of human adenocarcinoma: characterization of the major Ley antigen of human adenocarcinoma as trifucosylnonaosyl Ley glycolipid (III3FucV3FucVI2FucnLc6). J Biol Chem 1986;261: 1124753.
  • 5
    Kim YS, Yuan M, Itzkowitz SH, et al. Expression of LeY and extended LeY blood group-related antigens in human malignant, premalignant and nonmalignant colonic tissues. Cancer Res 1986;46: 598592.
  • 6
    Hellstrom I, Garrigues HJ, Garrigues U, et al. Highly tumor-reactive, internalizing, mouse monoclonal antibodies to Lewisy-related cell surface antigens. Cancer Res 1990;50: 218390.
  • 7
    Deshpande PP, Danishefsky SJ. Total synthesis of the potential anticancer vaccine KH-1 adenocarcinoma antigen. Nature 1997;387: 1646.
  • 8
    Deshpande PP, Kim HM, Zatorski A, et al. Strategy in oligosaccharide synthesis: an application to a concise total synthesis of the KH-1 (adenocarcinoma) antigen. J Am Chem Soc 1998;120: 160014.
  • 9
    Hummel G, Schmidt RR. A versatile synthesis of the lactoneo-series antigens- synthesis of sialyl dimer lewisx and of Dimer Lewisy. Tetrahedron Lett 1997;38: 11736.
  • 10
    Kensil CR, Patel U, Lennick M, et al. Separation and characterization of saponins with adjuvant activity from Quillaja saponaria Molina cortex. J Immunol 1991;146: 4317.
  • 11
    Ragupathi G, Koganty RR, Qiu DX, et al. A novel and efficient method for synthetic carbohydrate conjugate vaccine preparation: synthesis of sialyl Tn KLH conjugate using a 4-(4-N-maleimidomethyl)cyclohexane-1-carboxyl hydrazide (MMCCH) linker arm. Glycoconjugate J 1998;15: 21721.
  • 12
    Ragupathi G, Howard L, Cappello S, et al. 1999. Vaccines prepared with sialyl-Tn and sialyl-Tn trimers using the 4-(4-maleimidomethyl)cyclohexane-1-carboxyl hydrazide linker group result in optimal antibody titers against ovine submaxillary mucin and sialyl-Tn-positive tumor cells. Cancer Immunol Immunother 1999;48: 18.
  • 13
    Ragupathi G, Park TK, Zhang S, et al. Immunization of mice with a fully synthetic globo-H antigen results in antibodies against human cancer cells: a combined chemical-immunological approach to the fashioning of an anticancer vaccine. Angew Chem Int Ed Engl 1997;36: 1258.
  • 14
    Zhang S, Helling F, Lloyd KO, et al. Increased tumor cell reactivity and complement-dependent cytotoxicity with mixtures of monoclonal antibodies against different gangliosides. Cancer Immunol Immunother 1995;40: 8894.
  • 15
    Zhang S, Zhang HS, Cordon-Cardo C, et al. 1997. Selection of tumor antigens as targets for immune attack using immunohistochemistry. II. Blood group-related antigens. Int J Cancer 1997;73: 506.
  • 16
    Capurro M, Bover L, Portela P, et al. A monoclonal antibody active against human breast cancer specifically recognizes Lewis X hapten. Cancer Immunol Immunother 1998;45: 3349.
  • 17
    Cappello S, Liu NX, Musselli C, et al. Immunization of mice with fucosyl-GM1 conjugated with keyhole limpet hemocyanin results in antibodies against human small-cell lung cancer cells. Cancer Immunol Immunother 1999;48: 48392.
  • 18
    Kabat EA. Antigenic determinants and the size of the antibody-combining site: determinants of cell-mediated immunity. In: KabatEA. Structural concepts in immunology and immunochemistry. New York: Holt, Rinehart and Winston, 1976. 11966.
  • 19
    Livingston PO, Ragupathi G. Carbohydrate vaccines that induce antibodies against cancer. 2. Previous experience and future plans. Cancer Immunol Immunother 1997;45: 109.
  • 20
    Anderson PW, Pichichero ME, Stein EC, et al. Effect of oligosaccharide chain length, exposed terminal group and hapten loading on the antibody response of human adults and infants to vaccines consisting of Haemophilus influenzae type b capsular antigen uniterminally coupled to the diphtheria protein CRM 197. J Immunol 1989;142: 24648.
  • 21
    Pozsagay V, Chu C, Pannell L, et al. Protein conjugates of synthetic saccharides elicit higher levels of serum IgG lipopolysaccharide antibodies in mice than do those of the O-specific polysaccharide from Shigella dysenteriae type1. Proc Natl Acad Sci USA 1999;96: 51947.