Polyclonal antibodies targeting defined epitopes on cancer biomarker MUC16 Authors: Spencer E Moffitt, Anubhuti Srivastava, Naviya Schuster-Little, Caitlin R McEntee, Jane C March, Chien-Wei Wang, Danielle R Dotson, Holly V Goodson, Rebecca J Whelan Published: 2026 ABSTRACT AbstractMeasurements of the peptide epitope CA125 are crucial in ovarian cancer care. Despite its value as a tool in disease management, CA125 is an imperfect biomarker, with rates of false positive and false negative response that preclude its use as a screening tool in the general population. The monoclonal antibodies that perform capture and recognition in the CA125 test—OC125 and M11—were developed using complex targets as immunogens and recognize epitopes that have been located on mucin16 (MUC16) but otherwise remain undefined at the molecular level. We hypothesized that new antibodies recognizing MUC16 peptides of known sequence could enable the development of assay platforms that overcome the limitations of the current CA125 test. Here, we report the development and characterization of three sets of polyclonal antibodies that recognize known MUC16 peptides. Peptides that appear several times within MUC16’s highly conserved tandem repeat region were used to immunize two sets of rabbits. Affinity-isolated antibodies were characterized by enzyme-linked immunosorbent assay (ELISA) and Western blot. All three peptides were successful antigens, as indicated by the ability of the resulting polyclonal antibodies to bind individually expressed proteins from the MUC16 tandem repeat region. In particular, the polyclonal antibodies raised against peptide 2 (ELGPYTLDRNSLYV) bound to all tandem repeats tested. Peptide 2 antibodies were able to detect intact MUC16 from ovarian cancer cells in a surface plasmon resonance (SPR) assay. In flow cytometry experiments, peptide 2 antibodies bind to MUC16-positive cells (OVCAR3) and do not bind to MUC16-negative cells (OVCAR8). The binding pattern of these polyclonal antibodies opens the possibility of developing new monoclonal antibodies recognizing known epitopes within the tandem repeat region. These reagents may eventually complement or replace the monoclonal antibodies used in the current clinical CA125 test.Graphical Abstract Supplementary InformationThe online version contains supplementary material available at 10.1007/s00216-026-06583-2.Keywords: Polyclonal antibody, Cancer antigen 125, Mucin 16, Epitope, Tandem repeat Introduction According to current estimates from the American Cancer Society, nearly 20,000 new cases of ovarian cancer will be diagnosed in the USA in 2025, and this “silent killer” will claim the lives of > 12,700 individuals [1]. Modest gains in 5-year survival rates since 1975 are heartening but notably have not been shared uniformly across the population. While the 5-year survival rates for white individuals increased from 36 to 51% between 1975 and 2019, 5-year survival rates for Black individuals have remained essentially flat in the same timeframe at 42%, according to data from SEER (Surveillance, Epidemiology, and End Results) [1]. A notable feature of ovarian cancer, regardless of race, is the strong negative correlation between stage at diagnosis and 5-year survival rate. While ovarian cancer cases diagnosed at local stage are associated with a 5-year survival rate exceeding 90%, diagnosis at distant stage is associated with ~ 30% 5-year survival [1]. These statistics illustrate the importance of early-stage detection of ovarian cancer. Unfortunately, only 22% of ovarian cancer cases are diagnosed at the localized stage, compared with 48% diagnosed at distant stage. Taken together, these clinical data highlight the continuing need for the development of informative molecular tests for ovarian cancer to complement or replace existing detection strategies [2–5]. The gold standard biomarker for the clinical management of ovarian cancer is CA125 [6–8]. The current CA125 test (“CA125II”) uses two monoclonal antibodies—OC125 [9, 10] and M11 [11]—as capture and tracer [12]. These antibodies bind as-yet unknown peptide epitopes (referred to collectively as “CA125”) located on the massive glycosylated protein MUC16 [13–16]. MUC16 is present at low levels on healthy ovarian epithelia but is overexpressed in high-grade serous ovarian cancer, the most aggressive and prevalent type of ovarian cancer [17]. Cleavage of MUC16 proximal to the cell surface and circulation in the blood enables its detection through recognition of non-overlapping epitopes by OC125 and M11 antibodies. Generation of signal in the CA125II test requires that both OC125 and M11 recognize and bind their cognate epitopes, which are believed to be located on one or more of the 19 tandem repeat domains found in MUC16 [18]. Our group and others [19–22] have demonstrated, however, that the CA125 peptide epitopes are not uniformly distributed across the tandem repeat domains of the MUC16 protein. Using individually expressed tandem repeat proteins, we have shown that while M11 binds to all tandem repeats tested, OC125 fails to bind to ~ 30% of the tandem repeats [20]. The failure of OC125 to bind to CA125 could result in false negative responses in the CA125II test. At the same time, Hoffman and co-workers have demonstrated that common serum proteins also bind to OC125 and M11, thereby providing a mechanism of false positive response [23]. Improvements in ovarian cancer care could be enabled by the development of alternative affinity agents that recognize MUC16 at locations complementary to the CA125 epitopes [24]. Both OC125 and M11 are murine monoclonal IgG1 antibodies developed against complex immunogens. OC125 was developed by Bast and co-workers through immunization of BALB/c mice with intact and viable OVCA433 cells, a human epithelial ovarian carcinoma cell line that was established from a serous cystadenocarcinoma [10]. M11 was developed by O’Brien and co-workers using partially purified CA125 antigen from ascites fluid as the immunogen [11]. As has been demonstrated through elegant proteomics studies, > 2500 proteins can be identified in ascites fluid [25, 26]. Considering that the identity of the epitopes recognized by OC125 and M11 has not been determined at the amino acid-level, we hypothesized that raising novel antibodies against peptides of known composition derived from highly conserved regions of the parent protein (MUC16) would provide an opportunity to complement or replace OC125 and/or M11 in a serum test. By targeting known peptide sequences within MUC16, novel antibodies could display more consistent binding across all tandem repeats than OC125. As the first step in this campaign, we examined the MUC16 sequence to identify highly conserved peptides that appear multiple times in the tandem repeat region. These peptides were synthesized and used to generate polyclonal antibodies in rabbits. Here, we demonstrate that these polyclonal antibodies display the attributes we sought: the ability to bind individually expressed MUC16 tandem repeats and the ability to bind intact MUC16 from an ovarian cancer cell line (OVCAR3). These findings support the feasibility of developing new monoclonal antibodies against these peptides and their incorporation into a next-generation CA125 serum test. At the same time, polyclonal antibodies are valuable research reagents [27], and we show that peptide 2-binding antibodies in particular may be a valuable new tool in the effort to detect MUC16 reliably through binding to a conserved site complementary to the binding sites of OC125 and M11. Methods Selection of MUC16 peptide antigens The tandem repeat sequences of human MUC16 (accession number Q8WXI7) were aligned using ClustalX [28] and PASTA [29]. WebLogo (version 2.8.2) was used to visualize conserved and variable amino acids [30]. Highly conserved regions (appearing with total or partial fidelity in most or all tandem repeats) were individually examined for their likelihood to be immunogenic. The sequences chosen as peptide antigens, their corresponding numbers and amino acid positions within each tandem repeat, were CRLTLLRPEKD (Peptide 1, positions 59–69), ELGPYTLDRNSLYV (Peptide 2, positions 108–121), and EENMQHPGSRKFNT (Peptide 3, positions 21–34). Generation of polyclonal antibodies Polyclonal antibodies were produced by Biomatik (Kitchener, ON, Canada). First, three peptide antigens were synthesized by Biomatik with user-defined sequences (Peptide 1: CRLTLLRPEKD, Peptide 2: CELGPYTLDRNSLYV, Peptide 3: CEENMQHPGSRKFNT. Peptide 2 and Peptide 3 were synthesized with an N-terminal cysteine to enable KLH conjugation.) Standard immunization was performed, with two rabbits (A and B) being immunized with each peptide. Each of the six rabbits was given a unique identifier, as follows. Rabbit RB4385 was immunized with peptide 1; the polyclonal pool produced by this rabbit is referred to as “1A.” Rabbit RB4386 was immunized with peptide 1; the polyclonal pool produced by this rabbit is referred to as “1B.” Rabbit RB4387 was immunized with peptide 2; the polyclonal pool produced by this rabbit is referred to as “2A.” Rabbit RB4388 was immunized with peptide 2; the polyclonal pool produced by this rabbit is referred to as “2B.” Rabbit RB4389 was immunized with peptide 3; the polyclonal pool produced by this rabbit is referred to as “3A.” Rabbit RB4390 was immunized with peptide 3; the polyclonal pool produced by this rabbit is referred to as “3B.” Pre-immunization serum (0.5 mL/rabbit) and post-immunization antiserum (~ 20 mL/rabbit) were collected. Pre-immunization serum was found by the vendor to not bind to the target peptide antigen in ELISA (OD < 0.2 at 1:1000 dilution). Antigen affinity purification was performed on a portion of the post-immunization serum. The resulting purified antibodies were confirmed by the vendor to bind to their target peptide antigen via ELISA (dilution series 1:8000–1:128,000) and delivered in lyophilized form. Upon receipt, lyophilized proteins were reconstituted in distilled water and stored at 4 °C until use. Before performing affinity assays, the protein concentration in pre-immunization serum, post-immunization serum, and reconstituted affinity-isolated antibody solutions was determined by Qubit fluorescence protein quantification (Thermo, Waltham, MA, USA). Test solutions that were measured in affinity experiments were prepared using an appropriate dilution scheme so that the concentration of pre-immunization serum, post-immunization serum, or purified antibody was consistent across all solutions measured in an experiment. Expression and purification of tandem repeats MUC16 tandem repeat sequences were obtained using long-read (Nanopore) sequencing as described previously [18] and were cloned in pET-14b vectors (GenScript, Piscataway, NJ, USA) with BamHI and XhoI enzyme restriction sites. The plasmids were transformed into SHuffle T7 Express Escherichia coli cells (New England Biolabs, Beverly, MA) for expression of the N-terminal 6-His tagged protein to enable purification and immobilization. The E. coli cells were grown to late stationary phase at 30 °C in MagicMedia™ E. coli Expression Medium (Thermo) containing 100 μg/mL ampicillin. Cells were lysed by freeze–thaw cycles in liquid nitrogen followed by sonication in lysis buffer (20 mM sodium phosphate, 10 mM imidazole, 300 mM sodium chloride) containing cOmplete™ protease inhibitor (Thermo). Recombinant repeat proteins were purified on HisPur™ Ni–NTA Resin (Thermo), followed by size-exclusion chromatography using a Superdex 75 10/300 GL column on an ÄKTA pure™ 25 chromatography system. Three repeats (R14, R18, R19) were unable to be expressed in soluble form using this approach and were therefore not included in this study. Hereafter, the individually expressed MUC16 tandem repeats are referred to as “recombinant repeat proteins” or “repeats.” The identity of each repeat was confirmed by mass spectrometry, as reported in our previous publication [20]. Figure S1 shows a Coomassie blue–stained gel of the repeats. Concentrations of the repeats in stock solutions were determined by measurements of absorbance at 280 nm on a NanoDrop spectrophotometer (Thermo). Test solutions were prepared by diluting stock solutions to a consistent concentration appropriate to the experiment. ELISA Indirect ELISA was conducted in Pierce Nickel-Coated Plates by first immobilizing recombinant repeat proteins via their 6-His tag to Ni2+ sites on the plate surface. A saturating amount (886.5 ng, 5 × well capacity) of each recombinant repeat was dissolved in PBS before immobilization for 1 h at room temperature (RT). Wells containing HE4 recombinant protein (Fisher) and wells containing no antigen served as negative controls. HE4 was used as a negative control because it—like MUC16—is an ovarian cancer biomarker. Therefore, in samples from an ovarian cancer patient, both MUC16 and HE4 may be elevated. All recombinant repeat protein–antibody combinations were tested in triplicate. After repeat immobilization, three washes with 200 µL PBS-T (0.05% Tween-20) were done before the addition of 100 µL of affinity-isolated polyclonal antibody (1 μg/mL in PBS). The primary antibodies were incubated for 1 h at RT. Three washes with 200 µL PBS-T were done before the addition of 100 µL goat anti-rabbit HRP secondary antibody (Thermo, 1:20,000), followed by incubation for 1 h at RT. After a final set of three washes with 200 µL PBS-T, chemiluminescent signals were developed with SuperSignal ELISA Femto Substrate (Thermo). A SpectraMax M5 plate reader (Molecular Devices, San Jose, CA, USA) was used for detection of relative chemiluminescence units at 425 nm. Indirect ELISA for the pre-immunization and post-immunization serum samples was conducted in a similar manner, the only differences being that the combinations were done in duplicate and after the addition of the serum samples, the number of washes with 200 µL PBS-T (0.05% Tween-20) was increased from three to five and the duration of each wash was increased from 30 s to 5 min. In each case, the total protein concentration added to the ELISA plate as primary antibody was 20 µg/mL. Assuming that 10% of the protein concentration in serum is from antibodies, the serum solutions are prepared such that the antibody concentration is 2 µg/mL. Western blot Recombinant repeat proteins were separated by size via sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS-PAGE) using an applied voltage of 150 V for 75 min on a 16% acrylamide gel and then transferred to a polyvinylidene difluoride (PVDF) membrane using an applied voltage of 20 V for 90 min. After transfer, the membrane was blocked with a solution of 5% non-fat milk in Tris-buffered saline containing 0.05% Tween-20 (TBS-T) for 1 h at RT. The membrane was then incubated overnight at 4 °C with a polyclonal antibody solution (1:500 dilution in 5% non-fat milk TBS-T). After incubation, the membrane was washed with TBS-T for 40 min (4 rounds of 10-min washes). The membrane was then incubated at RT with goat anti-rabbit HRP secondary antibody (Thermo, 1:10,000 dilution in 5% non-fat milk in TBS-T) for 1 h. The membrane was developed using Pierce ECL Western Blotting Substrate (Thermo) and imaged with a ChemiDoc system (Bio-Rad, Hercules, CA, USA). Isolation of MUC16 MUC16 was isolated from conditioned cell media of OVCAR3 cells as previously described [31]. Briefly, low passage number OVCAR3 cells were validated by Short Tandem Repeat profiling (ATCC, Manassas, VA, USA) and confirmed to be free of Mycoplasma contamination using a PCR kit from Genlantis (San Diego, CA, USA). Cells were cultured under standard conditions of temperature, CO2, and humidity in 175 cm2 tissue culture flasks containing RPMI 1640 media supplemented with 10% FBS and 2 mM l-glutamine (Thermo). Cells were grown until they reached 100% confluency at which point spent media was removed and replaced with fresh media devoid of FBS and phenol red. After 5–7 days, conditioned media were collected and centrifuged to pellet cell debris. Supernatant (containing MUC16) was spun through a 100-kDa molecular weight cut-off filter (MilliporeSigma) to concentrate MUC16. Total protein concentration was determined using a Pierce BCA assay (Thermo). CA125 counts were determined using an R&D Systems human CA125/MUC16 assay kit (Biotechne, Minneapolis, MN, USA). Surface plasmon resonance Alto Surface Plasmon Resonance Spectrometer, 16-channel CMD-carboxyl (CBX) cartridges with cartridge fluid, low refractive index solution (8% Glycerol), high refractive index solution (32% glycerol), aliquots of 1-ethyl-3-(3-dimethylaminopropyl)carbodiimide hydrochloride (EDC), aliquots of N-hydroxysuccinimide (NHS), conditioning solution (10 mM HCl), quenching solution (1 M ethanolamine), and immobilization buffer (10 mM sodium acetate at pH 5.0) were purchased from Nicoya Lifesciences (Kitchener, ON, Canada). Regeneration solution (10 mM Glycine–HCl at pH 1.5) and running buffer (PBS from GrowCells, Irvine, CA, USA, supplemented with 0.1% v/v Tween-20) were prepared in-house. Recombinant CA125 (rCA125) expressed in CHO cells was purchased from R&D Systems (Minneapolis, MN, USA). The protein was reconstituted to a final concentration of 250 µg/mL in PBS. All reagents and samples were supplemented with 0.1% v/v Tween-20. Experimental protocols were created using Direct Kinetics with Single-Cycle Titration assay templates in the Nicosystem software. Sample and sensor temperature was held at 25 °C. For the Activities, default droplet settings were used for Startup, Normalize, Clean, and Build Surface. Default droplet times were used for Kinetic Step, except the analyte association droplets, which were extended from 180 to 300 s. Peptide 2 polyclonal antibodies from rabbit B were diluted to a concentration of 5 µg/mL in 10 mM sodium acetate at pH 5.0 and immobilized onto the CBX CMD cartridges using EDC/NHS coupling. The concentration of EDC and NHS used was 200 mM. Intact MUC16 from conditioned cell media and rCA125 were introduced as analytes. SPR binding data was downloaded from the Nicosystem software. The signal from each concentration of analyte was averaged and plotted against CA125 concentration using Igor-Pro software (Wavemetrics Inc., Oregon, USA). A 1:1 binding model was used for curve fitting. Flow cytometry: cell line and culture conditions Ovarian cancer cell lines OVCAR3 and OVCAR8 were cultured in RPMI 1640 supplemented with 10% FBS and 2 mM l-glutamine under standard conditions (in a humidified 5% CO2 incubator at 37 °C). After reaching ~ 90% confluency, cells were detached by incubating with 3 mL 0.25% Trypsin–EDTA solution for 5 min. Cell density and viability were determined by counting cells stained with Trypan blue on a hemacytometer. Cells were then resuspended in staining buffer (PBS with 2% FBS and 2 mM EDTA) at a concentration of approximately 350,000 cells per 50 μL. Flow cytometry: binding assays Four concentrations of the polyclonal antibodies 2B were prepared in staining buffer: 0 μg/mL, 10 μg/mL, 20 μg/mL, and 50 μg/mL. Fifty microliters of each antibody solution was added to 50 μL of resuspended cells (around 350,000 cells) and incubated for 30 min at RT. Cells were washed twice with 200 μL of staining buffer followed by incubation with 100 μL of 1:500 dilution of FITC-labelled anti-rabbit secondary antibody (Thermo) for 30 min at RT. Cells were washed as described before and resuspended in PBS prior to flow cytometry analysis. An “only cells” control was included which was not incubated with either the primary or secondary antibody. Samples were run in duplicate on a NovoCyte Flow Cytometer (Agilent, Santa Clara, CA, USA), and 10,000 events were recorded at a medium flow rate. Results and discussion Selection of MUC16 peptide antigens Multiple sequence alignment reveals that among the 19 individual tandem repeats found in the consensus sequence of MUC16 [18], some positions are conserved while others are highly variable. To identify peptides that could serve as MUC16 peptide antigens, we looked for regions that displayed high conservation over 10–15 amino acids. The sequence alignment table and associated WebLogo are found in Fig. 1. The amino acid positions that were used as the peptide antigens are highlighted with a red bar above their corresponding position numbers. Figure 2A shows a ribbon diagram of one tandem repeat (R9) and the placement of the three peptide antigens, with peptide 1 in green, peptide 2 in purple, and peptide 3 in brown. Figure 2B shows a space-filling model of R9. The location of the three peptide antigens relative to a Clustal X alignment of the 19 tandem repeats. The 19 tandem repeats of human MUC16 (accession number Q8WXI7) were aligned using Clustal X and visualized using Jalview with the Clustal standard color scheme. Red bars at the top show the location of the peptides used to generate polyclonal antibodies. The WebLogo below the alignment shows the amino acid abundance at each position with more abundant amino acids represented as taller characters A Ribbon diagram of a MUC16 tandem repeat showing the positions of peptide antigens. Shown is MUC16 tandem repeat R9. Peptide 1 (amino acids 59–69) is shown in green. It includes a portion of beta sheet strand and part of the “C-loop.” Peptide 2 (amino acids 108–121) is shown in purple; it is an unstructured coil at the C′ terminal end of the SEA (sea urchin sperm protein, enterokinase, agrin) domain, proximal to the Proline/Serine/Threonine–rich region. Peptide 3 (amino acids 21–34) is shown in brown. It is partially unstructured coil (part of the “first loop”) and partially alpha helix. B Space-filling model of a MUC16 tandem repeat showing hydrophobicity. Shown is MUC16 tandem repeat R9 with hydrophobic regions shown in red and hydrophilic regions shown in blue. For both A and B, the structure of tandem repeat R9 was visualized and annotated using ChimeraX-1.10.1 from a.pdb file generated in I-TASSER Although we chose the peptide antigens to contain the MUC16 consensus sequence as much as possible, there are locations where the peptide antigens differ from the sequences in individual tandem repeats. These differences are shown in Fig. 3, where peptide 1 is highlighted in green, peptide 2 is in purple, and peptide 3 is in brown. Perfect mapping of the peptide antigen is shown with darker color, while lighter background color is used to show imperfect mappings, with no color showing the amino acids that differ. Tandem repeats and peptide antigens. Peptide 1 (CRLTLLRPEKD) is shown in green, peptide 2 (ELGPYTLDRNSLYV) is shown in purple, and peptide 3 (EENMQHPGSRKFNT) is shown in brown. In each case, darker color designates perfect mapping, and lighter colors designate imperfect mapping, with absence of color showing amino acids that differ in sequence between the tandem repeat and the peptide antigen. Sequences are from the repeat region of human MUC16 (accession number Q8WXI7) Enzyme-linked immunosorbent assay (ELISA) Affinity-isolated polyclonal antibodies from six rabbits (two rabbits per peptide antigen) were tested by ELISA for their ability to bind MUC16. Specifically, individual subunits from the MUC16 tandem repeat domain were expressed in E. coli with N-terminal 6-His tags and purified on Ni2+ resin followed by size-exclusion chromatography. These recombinant repeat proteins were then immobilized onto Ni2+-coated plates and probed with each of the polyclonal antibody solutions. Figure 4 shows normalized chemiluminescent ELISA signals for each of the three peptide antigens (1, 2, and 3) and the two rabbits that were immunized. Normalized chemiluminescent ELISA. Signals show that the antibodies generated from two different rabbits immunized with the same peptide have comparable binding patterns and that antibody 2 binds the most uniformly to the tandem repeat region out of the antibodies tested. Signals less than 0.20 are classified as non-binding. Error bars indicate the standard deviation (n = 3). The antibodies tested were (A) 1A, (B) 1B, (C) 2A, (D) 2B, (E) 3A, and (F) 3B. The sequences of peptides used as immunogens are shown above their corresponding data plots Two trends are notable. First, the pattern of binding is similar between the two rabbits (ELISA results from the two rabbits immunized with the same peptide are in the same column in Fig. 4, with the peptide immunogen shown at the top). We interpret this observation to indicate that the immune response initiated by each of the three peptide antigens is reproducible across multiple animal hosts, a necessary condition for the successful generation of polyclonal antibodies. Secondly, although each polyclonal antibody set successfully binds multiple MUC16 tandem repeats, the polyclonals vary in their ability to bind MUC16 tandem repeats. The peptide 1 polyclonal pools bound the fewest tandem repeats (9 out of 16, ~ 56%), followed by the peptide 3 polyclonal pools (11 out of 16, ~ 69%) and the peptide 2 polyclonal pools, which displayed 100% binding from the pool from rabbit B (Fig. 4D) and binding to 15 of 16 from rabbit A (~ 94%, Fig. 4C). The difference is in repeat 11 (R11), for which the binding of 2 A was ~ 0.10, below the value used as the cut-off to define binding (0.20); in 2B, the binding signal was ~ 0.40. The use of 0.20 as the cut-off to define binding in ELISA data is rationalized in our previous publication [32]. Greater insight into the binding patterns observed in ELISA data can be gained by examining binding between the polyclonal antibodies and tandem repeats that differ in their amino acid composition from the peptide used as the antigen. Figure 5 illustrates this trend, with the sequence of the peptide antigen in the far-left column (darker blue background) and the sequences of the 16 repeats tested using ELISA listed in the other columns. Red letters show sites at which a tandem repeat differs from the peptide antigen. Blue background designates repeats that bind to the antibodies; pink background designates non-binding repeats. Amino acid differences and binding patterns in ELISA. The sequence in the far-left column (darker blue background) is the peptide that was used as the antigen. The antibodies tested were (A) 1 A, (B) 1B, (C) 2 A, (D) 2B, (E) 3 A, and (F) 3B. Columns with blue background show the expressed repeats (R1–R17) that were recognized by polyclonal antibodies against each peptide. Columns with pink background show repeats that were not recognized by these antibodies. Black text designates amino acids that are the same as the peptide antigen. Red text shows amino acids that differ in the repeat from the peptide antigen A difference in sequence partially explains the observed ELISA binding patterns. Repeats with more amino acid differences from the sequence of the peptide antigen (red text) tend to be the repeats that do not bind to the antibodies (pink background). Examining these differences in sequence reveals interesting binding trends. For the peptide 1 polyclonal antibodies, we observe that all repeats in which proline at position 8 is substituted with serine do not bind. We have previously demonstrated [33] a crucial role of this substitution in affecting secondary structure of synthesized 21mer peptides from the “C-loop,” of which this peptide comprises the first 11 amino acids. R12 has the same sequence as peptide antigen 1 but does not bind in ELISA (as shown in Fig. 6, below, it does bind in Western blot). Two other notable non-binding repeats against the peptide 1 polyclonal antibody pool are R11 and R15, in which the expressed repeat differs due to the change of an acidic residue in the peptide antigen to a basic residue (D11R in R11 and E9K in R15); these data suggest that these acidic residues are important in binding. Complicating this interpretation is the fact that both R11 and R15 display binding in Western blot. Further investigation is necessary to explain this discrepancy. As noted above, the polyclonal antibodies generated against peptide 2 display complete binding from one rabbit and nearly complete binding from the other rabbit. R11 was found in ELISA to bind antibodies raised in rabbit B against peptide 2, but not to bind antibodies raised in rabbit A against the same peptide. Polyclonal antibodies from both rabbits bind in Western blot. R11 is the only repeat in which the substitution at position 10 is from N to G (the other sequences with amino acid differences have N changing to D; all these repeats bind in both assays). Among the peptide 3 polyclonal antibodies, it is noteworthy that all the repeats in which glutamine at position 5 is not altered display binding, suggesting that this residue is involved in binding [34]. Western blots show the comparable binding patterns of the antibodies generated from rabbits immunized with the same peptide. The antibodies tested were (A) 1 A, (B) 1B, (C) 2 A, (D) 2B, (E) 3 A, and (F) 3B The binding pattern of pre- and post-immunization sera to immobilized tandem repeats was characterized via indirect ELISA. In all cases, post-immune serum from an individual rabbit displayed higher binding to immobilized tandem repeats than pre-immune serum. Representative data (for antibodies raised in rabbit B against peptide 2) are shown in Figure S2. Western blot Further characterization of the binding between the novel polyclonal antibodies and expressed MUC16 tandem repeats was accomplished via Western blot. Figure 6 shows Western blots generated for all six antibody pools (three peptides, two rabbits per peptide) interacting with the 16 expressed MUC16 tandem repeats. The similarities between the rabbits are notable here, as they were in the ELISA binding. The very slight differences that are observed can be attributed to natural biological variation in the immune response of the two different animals that were immunized. Uncropped Western blot images are included in the Supplemental Information section as S3–S14. It is noted that several of the blots include multiple bands at or slightly below 20 kDa. When expressed with a 6-His tag, the molecular weights of the tandem repeat proteins range from a minimum of 18.2 kDa (R17) to a maximum of 19.9 (R2, R11, and R12), consistent with the location observed in the Western blots. The multiple bands are attributed to smaller tandem repeat proteins resulting from cleavage of the unstructured C-terminal region of each tandem repeat during expression and purification. These cleaved products still contain the peptide epitope and are therefore stained in the Western blot. Supplemental Figure S15 shows a Western blot image in which primary antibody was not included during the blotting process, to assess the possibility of non-specific interaction between the tandem repeat proteins and the secondary antibody. As expected, no signal is observed in this Western blot. Comparison of ELISA and Western blot Supplemental Information Figure S16 compares the results from ELISA and Western blot across the antibody pools and tandem repeats tested. Blue boxes show binding; pink boxes show non-binding. The two methods largely agree, with a small number of discrepancies (described above). The similarity in binding between Western blot, in which protein targets are denatured before binding—and ELISA—which interrogates targets in their native state—suggests that the epitopes of these polyclonal antibodies are linear, rather than conformational, in nature. This observation is expected given the relatively short peptide sequences used as antigens [35]. The instances in which binding is observed in Western blot but not ELISA can be attributed to the fact that the repeat protein is denatured and linearized in Western blot whereas in ELISA, the protein is interrogated in its native conformation. The short peptide sequence that served as the immunogen may be sterically hindered or buried due to the native conformation or immobilization orientation. This observation—which is consistent with the idea that the epitopes of these polyclonal antibodies are linear—does not render those antibodies less useful, rather it provides information about the applications that are appropriate for the antibodies. Polyclonal antibodies that display binding in Western blotting across most or all repeats—the antibodies raised against peptide 2 (ELGPYTLDRNSLYV) and peptide 3 (EENMQHPGSRKFNT)—are better suited to applications in which the antigen is linearized before interaction with antibody. Surface plasmon resonance (SPR) Polyclonal antibodies for peptide 2 from rabbit B bound to all 16 repeats tested in ELISA and Western blot. To further assess the ability of these polyclonal antibodies to interact with intact MUC16, surface plasmon resonance (SPR) binding assays were performed. First, we tested the binding of these antibodies (referred to as “2B”) to individually expressed tandem repeat proteins and determined the values of equilibrium dissociation constants (Kd). For the interaction of 2B antibodies and individually expressed tandem repeat proteins, values of Kd ranged from a minimum of 1.3 ± 0.1 nM to a maximum of 7.8 ± 8.3 nM. These values are comparable to those found for the clinically used monoclonal antibodies, which for M11 ranged from a minimum of 0.13 ± 0.08 nM to a maximum of 16 ± 3 nM and which for OC125 ranged from a minimum of 0.03 ± 0.03 nM to a maximum of 0.88 ± 0.03 nM [19]. It is important to note that Kd values for polyclonal antibodies are not rigorously determined because the polyclonal pool contains antibodies with a range of affinities. Next, we tested the binding of the 2B antibodies to a recombinant protein that contains the 987 amino acids (Met13360-Gln14347) overlapping with the six C-terminal MUC16 tandem repeats. Because this recombinant protein contains CA125 epitopes, it is referred to as recombinant CA125 (rCA125). This protein was expressed in CHO cells and determined by the vendor to be > 90% pure by SDS-PAGE and quantitative densitometry. Concentration-dependent binding with high precision between replicates was observed (Fig. 7A). The value of Kd for the interaction of 2B with rCA125 is estimated to be 89 ± 17 nM. It is worth noting that three of the tandem repeats contained in rCA125 (R14, R18, and R19) are not included in this study as individually expressed proteins because they were not expressed in soluble form. Polyclonal antibody 2B binds recombinant CA125 and intact MUC16. Results from SPR binding assays of polyclonal antibodies generated against peptide 2 from rabbit B with (A) rCA125 and (B) intact MUC16 from conditioned cell media of ovarian cancer cells (OVCAR3). Data points were fit with a 1:1 binding model (solid line) and error bars indicate standard error of the mean (n = 3). Following convention, the concentration of MUC16 (in panel B) is reported in terms of the functional units of the CA125 epitope as determined by immunoassay per milliliter of solution The successful binding of 2B with rCA125 was followed by binding affinity characterization of 2B with intact MUC16 enriched from conditioned cell media (CCM) of OVCAR3 cells. The solution of CCM was first passed through a 100-kDa molecular weight cut-off filter and then passed through a size-exclusion column. MUC16 elutes in the void volume because of its large size. As we demonstrated in our previous publication [36], this sample preparation strategy retains up to 98% of MUC16 in 1.7% of total recovered protein. As with rCA125, concentration-dependent binding and high measurement precision were observed (Fig. 7B). The ability of 2B to detect all individual tandem repeats tested, rCA125, and intact MUC16 demonstrates its potential to be a valuable tool in improving the detection of MUC16. Flow cytometry The specificity of the novel polyclonal antibodies was tested in a flow cytometry–based binding assay. In the experiment with OVCAR3 cells and 2B antibodies, a concentration-dependent increase in fluorescence intensity was observed with increasing concentration of the antibodies, indicating that the antibodies recognize MUC16 on the surface of OVCAR3 cells (Fig. 8A). In contrast, when varying concentrations of 2B antibodies were incubated with MUC16-negative OVCAR8 cells, no such concentration-dependent shift in fluorescence intensity was observed (Fig. 8B). This can be attributed to the absence of MUC16 on OVCAR8 cells [37, 38]. Two negative control peaks, corresponding to only cells and cells with no primary antibody (only FITC-labeled secondary antibody), overlapped in both experiments, illustrating that the secondary antibody does not bind non-specifically to either OVCAR3 or OVCAR8 cells. These flow cytometry experiments provide additional evidence suggesting that the 2B polyclonal antibodies are specific for MUC16. Polyclonal antibody 2B binds MUC16-positive cells and does not bind MUC16-negative cells. Results from flow cytometry assays of polyclonal antibodies generated against peptide 2 from rabbit B incubated with (A) OVCAR3 cells (MUC16-positive) and (B) OVCAR8 cells (MUC16-negative). In both (A) and (B), the color scheme is trace 1 (blue), cells only; trace 2 (red), 0 µg/mL 2B antibody; trace 3 (green), 10 µg/mL 2B antibody; trace 4 (orange), 20 µg/mL 2B antibody; trace 5 (black), 50 µg/mL 2B antibody Conclusion and future directions We conclude that the strategy of generating polyclonal antibodies recognizing conserved sequences in the MUC16 tandem repeat region appears to have been successful. As expected, the antibodies appear to be linear, rather than conformational, in nature. In binding studies against 16 of the 19 MUC16 tandem repeats, the peptide 2 polyclonal antibodies bind more consistently than does OC125, one of the two antibodies used in the current clinical test. This observation opens the exciting possibility that a monoclonal antibody raised against peptide 2 could complement or replace OC125 in clinical assays, providing greater coverage of the intact MUC16 molecule with a concomitant lower false negative rate. The development of such monoclonal antibodies is ongoing. In addition, epitope binning experiments are being conducted to identify the combination of antibodies most suitable for use in a next-generation MUC16 assay. At the same time, polyclonal antibodies generated against peptide 2 may be useful research tools in applications such as Western blotting, immunoprecipitation, and immunohistochemistry and merit further investigation. Supplementary Information Below is the link to the electronic supplementary material. Supplementary file1 (PPTX 10.1 MB) Supplementary file2 (XLSX 12.2 KB) Supplementary file3 (XLSX 12.1 KB) Supplementary file4 (XLSX 12.2 KB) Supplementary file5 (XLSX 12.1 KB) Supplementary file6 (XLSX 12.1 KB) Supplementary file7 (XLSX 12.1 KB) Acknowledgements We thank Professor Robin Orozco (University of Kansas) for helpful discussions. Author contributions Conceptualization: Naviya Schuster-Little, Holly V. Goodson, Rebecca J. Whelan. Investigation: Spencer E. Moffitt, Anubhuti Srivastava, Naviya Schuster-Little, Caitlin R. McEntee, Jane C. March, Chien-Wei Wang, Danielle R. Dotson. Visualization: Spencer E. Moffitt, Anubhuti Srivastava, Naviya Schuster-Little, Rebecca J. Whelan. Writing—original draft: Spencer E. Moffitt, Rebecca J. Whelan. Writing—editing and review: Spencer E. Moffitt, Anubhuti Srivastava, Naviya Schuster-Little, Caitlin R. McEntee, Jane C. March, Chien-Wei Wang, Danielle R. Dotson, Holly V. Goodson, Rebecca J. Whelan. Funding acquisition: Holly V. Goodson, Rebecca Whelan. Supervision: Rebecca J. Whelan. All authors have read and agree to the published version of the manuscript. Funding Seed funding from the Mike and Josie Harper Cancer Research Institute at the University of Notre Dame that enabled the generation of the polyclonal antibodies is gratefully acknowledged. N.S.L. was a fellow in the Chemistry-Biochemistry-Biology Interface (CBBI) program at the University of Notre Dame, supported by training grant T32GM075762 from the National Institute of General Medical Sciences. S.E.M. was a participant in the National Science Foundation Research Experience for Undergraduates at the University of Kansas, supported by NSF CHE-2349329. This work was supported in part by NIH grants from NIGMS and OD (P20GM152280) and NCI (R21 CA267532). Additional support from the Kansas Institute for Precision Medicine NIGMS COBRE (P20 GM130423), the Center for Molecular Analysis of Disease Pathways NIGMS COBRE (P20 GM103638), Tina’s Wish, and the University of Kansas is gratefully acknowledged. H.V.G. acknowledges support from the National Science Foundation (MCB 2027389). Research reported in this publication was supported by the National Institute of General Medical Sciences (NIGMS) of the National Institutes of Health under award number P20GM113117. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health. Data availability The original data presented in the study are openly available in KU ScholarWorks at 10.17161/1808.39351. Declarations Conflict of interest The authors declare no competing interests. Footnotes Published in the topical collection highlighting Emerging Capabilities in Microscale Separations and Bioanalysis with guest editors Mark A. Hayes, Christopher R. Harrison, and Alexandra Ros. Publisher's Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. References Associated Data Supplementary Materials Supplementary file1 (PPTX 10.1 MB) Supplementary file2 (XLSX 12.2 KB) Supplementary file3 (XLSX 12.1 KB) Supplementary file4 (XLSX 12.2 KB) Supplementary file5 (XLSX 12.1 KB) Supplementary file6 (XLSX 12.1 KB) Supplementary file7 (XLSX 12.1 KB) Data Availability Statement The original data presented in the study are openly available in KU ScholarWorks at 10.17161/1808.39351.