To study MatAbs, samples were collected from 44 client-owned dogs undergoing cesarean sections at the Cornell University Hospital for Animals (CUHA, IACUC approval 2023-0003). Maternal blood was collected from either the cephalic or saphenous vein. Umbilical cord blood samples were taken immediately after the surgical delivery of pups, as follows: the umbilical cord was clamped at either end to detach the placenta from the puppy. Whenever possible, a syringe and needle were used to aspirate the blood from the cord. If this method was not feasible, the clamp closer to the puppy was removed, and a 1.5-ml microcentrifuge tube was used to collect cord blood manually expressed from the blood vessels. Samples of both live births and stillborns were collected and used in the analysis. Colostrum was collected from at least four teats within 1 h of delivery, pooled, and stored in a single 1.5 ml microcentrifuge tube (variation has been reported between teats (7)). All samples were kept at 4°C for up to 24 h until further processing, after which maternal and cord blood samples were centrifuged for 5 min at 8,000 rpm to separate serum. All samples were subsequently stored long term at − 80°C. Basic clinical and demographic data for each dam were recorded, including age, breed, weight, body condition score, parity, litter size, and any known vaccination history.

For the production of CPV virus-like particles (VLPs) required for ELISAs, the Bac-to-Bac Baculovirus Expression system was used in DH10 Escherichia coli bacteria. The plasmid encoding the CPV-VP2 capsid protein was purified using a commercial kit (Monarch Plasmid Miniprep Kit) according to the manufacturer’s recommendations. Spodoptera frugiperda (Sf9) insect cells were transfected with VP2 BacMid using Cellfectin-II transfection reagent (Invitrogen, California, USA), using 8 μl Cellfectin-II and 1 μg BacMid in 200 μl per T-25 flask with serum-free Grace Media. After incubating for 4 h at 27°C, the transfection mixture was removed, and Grace Media with 10% fetal calf serum (FCS) was added, followed by incubation for 4–7 days at 27°C. To harvest the P0 virus stock, the culture media were collected and centrifuged for 5 min, and then the cell culture was centrifuged for 5 min at 500 g. Subsequently, 200 μl of P0 viral stock was added to fresh cultures and incubated for 4–5 days in a humidified incubator at 27°C. To amplify the P1 virus in Sf9 cells, supernatant was collected from transfected Sf9 cells and centrifuged for 5 min at 500×g. A 100-μl aliquot of supernatant was then transferred to previously seeded Sf9 insect cells. Cells were incubated for 4–5 days in a humidified incubator at 27°C, after which the virus was harvested. Protein production was subsequently performed in Hi5 cells infected with P2 virus under the same conditions. VLPs were purified using a cesium chloride (CsCl) gradient, followed by dialysis to remove residual CsCl. VLP production was confirmed by SDS-PAGE with Coomassie staining and Western blotting. Protein concentrations were determined using both a bicinchoninic acid (BCA) assay and a Qubit fluorometric assay.

Vero cells stably expressing canine signaling lymphocytic activation molecule [SLAM; also known as CD₁₅₀, the receptor for CDV (12)] were incubated at 37°C and 5% CO₂ in DMEM containing 50 IU/ml penicillin, 50 μg/ml streptomycin, and 5% FCS (complete media). Cells were infected with CDV (Onderstepoort strain, provided by Diego Diel, Cornell University) at a MOI of 0.01 in 5 ml of infection media with 2% FCS. The virus–media mix was incubated with cells for 1 h at 37°C with gentle side-to-side rocking every 15 min. Complete media was then added, and the flask was incubated and checked daily for cytopathic effect (CPE). Once CPE was observed, the supernatant was collected, centrifuged at 1,600 rpm for 5 min at 4°C, aliquoted, and stored at − 80°C. A TCID₅₀ assay was used to determine viral titer. A total of 2 × 10⁴ Vero-dogSLAM cells were seeded per well of a 96-well plate in a total volume of 50 μl. A 10-fold dilution of the supernatant containing virus was prepared in complete media, and 50 μl per well of each dilution was added in quadruplicate to the seeded cells. The last row of wells remained uninfected as a negative control. The plate was incubated until a clear CPE was visible, and the titer was calculated using the Reed–Muench method.

For virus-specific ELISAs, high-binding ELISA plates (Greiner, North Carolina, USA) were coated with either 15 ng of CPV-VLPs in phosphate-buffered saline (PBS) per well or 1 × 10⁵ TCID₅₀ CDV particles. Plates were incubated overnight at 4°C, washed with PBS containing 0.1% Tween 20 (“PBS-T”), and blocked with 5% milk in PBS-T. After washing with PBS-T, serially diluted dam serum (1:800 starting dilution), cord serum (1:100 starting dilution), or colostrum (1:1,600 starting dilution) was added, followed by incubation for 2 h at 37°C. After washing with PBS-T, either anticanine IgG (H+L) conjugated to horseradish peroxidase (HRP) (Cat. No. PA1-29738, Invitrogen) or anticanine IgG2 (Cat. No. A40-121P, Bethyl Laboratories, Texas, USA) was added for 1 h. Following a final wash step, tetramethylbenzidine (TMB) substrate (Abcam, Massachusetts, USA) was added for 10 min. This reaction was stopped using 1 M H₂SO₄, and absorbance was measured at optical density, 450nm (OD450) (BioTek Cytation 7 Cell Imaging Multimode Reader, Agilent, California, USA). Each sample was tested in duplicate. Every plate included a positive control (pooled serum samples from five healthy, vaccinated dogs), and a virus-specific IgG-negative control (Marshall Bioresources, New York, USA) was included on every plate. To correct for background signal, the optical density (OD) values from wells coated with PBS only were subtracted from the OD values of wells coated with VLPs or virus. The corrected OD450 values were plotted against dilution in GraphPad Prism, and nonlinear regression analysis with interpolation was used to calculate endpoint titers. The endpoint titer was defined as the reciprocal of the dilution at which the OD450 crossed the detection threshold, calculated as the mean OD450 of all PBS-coated wells plus three standard deviations. Results were only accepted if the nonlinear regression fit for both the standard curve and the sample curve had a coefficient determination (R²) ≥ 0.9.

Indirect ELISA assays were developed in-house to quantify total IgG in clinical samples. Rabbit anticanine IgG (Cat. No. SA5-10309, Invitrogen) was coated onto ELISA plates at 1 μg/ml in PBS overnight at 4°C. Wells were washed three times with PBS-T and then blocked with 5% milk in PBS-T. After washing with PBS-T, serially diluted dam serum, cord serum, or colostrum was added. Each sample was tested in duplicate. For the standard curve, serially diluted canine IgG (Cat. No. 0129-01, Southern Biotech), starting at 5 μg/ml, was added in duplicate to the plate. Samples were incubated for 2 h at 37°C. After washing, the TMB substrate was added, and the reaction was stopped using 1 M H₂SO₄. Absorbance was measured at OD450. Background signal was corrected by subtracting PBS-only values, as performed for virus-specific ELISAs. Corrected OD450 values were plotted against the dilution factor using GraphPad Prism, and nonlinear regression analysis with interpolation to the standard curve was performed to calculate the IgG concentration in each sample. Results were accepted only if the nonlinear regression fit for both the standard curve and the sample curve had an R² ≥ 0.9.

Based on preliminary data and paired sample comparisons, recruitment of 25 dogs was estimated to provide 80% power (α = 0.05) to detect a standardized difference of 0.6 in antibody transfer ratios between maternal and neonatal samples, a magnitude comparable to that reported in human studies (13, 14). To ensure robust analysis across a naturally occurring cohort, matched samples were ultimately collected from 44 dams, exceeding the minimum requirement. Puppies from the same litter were treated as biological replicates, and data were aggregated at the litter level to avoid pseudo-replication and eliminate the need for hierarchical model analysis. All statistical analyses were conducted using GraphPad Prism (version 10.3.1). End-point titer outcomes of CPV- and CDV-specific IgG, as well as total IgG concentrations, were nonnormally distributed; therefore, all were analyzed using the Kruskal–Wallis test followed by Dunn’s multiple comparisons test. Transfer ratios between dam and colostrum IgG, and between dam and cord IgG, were analyzed using Mann–Whitney U tests. Correlation analyses included nonlinear regression, Pearson’s correlation for normally distributed variables, and Spearman’s rank correlation for nonnormally distributed variables. Statistical significance was defined as p ≤ 0.05 for all comparisons.

A total of 44 female dogs (“dams”) presenting to Cornell University Hospital for Animals for cesarean sections were enrolled in this study. The dams had a mean age of 3.6 years old (range 1–8 years), and multiple breeds were represented, including brachycephalic breeds (American Bulldog [four], French Bulldog [eight], English Bulldog [two], bully cross [2], pitbull [1], Chihuahua [1], shih tzu cross [1]) and nonbrachycephalic breeds (labrador [four], corgi [three], Golden Retriever [three], Leonberger [two], others [12]). All cesarean sections were emergency procedures except for a single elective procedure. Demographic and clinical details for each dam are presented in Table 1.

IgG is the major antibody isotype transferred from mother to offspring in all mammalian species (15, 16). For puppies, the transfer of MatAbs specific for life-threatening pathogens is critically important for neonatal health. Therefore, we first aimed to quantify IgG for two significant canine pathogens: CPV and CDV. We also reasoned that most dams would be seropositive to these viruses due to core vaccination recommendations (17); thus, studying the transfer of these MatAbs would be feasible. Virus-specific IgG was quantified in dam serum and compared with MatAbs detectable in colostrum. We further compared virus-specific IgG in dam serum with that detected in serum separated from cord blood, representing the first reported use of cord blood sampling to quantify MatAbs delivered transplacentally in dogs. This cord sampling strategy enabled precise measurement of placental antibody transfer, overcoming limitations of prior methods that relied on blood sampling from newborn puppies. Direct sampling of puppies is more ethically sensitive, technically challenging, and potentially confounded by early colostrum intake.

Virus-specific IgG titers for all samples in our cohort were quantified using in-house ELISAs. Serial dilutions of each sample were generated, and the endpoint titer was defined as the highest dilution at which antibody binding to the viral antigen remained detectable. Endpoint titers are presented as the reciprocal of this final positive dilution. A small number of dams had undetectable titers of IgG specific for CPV (three dams, 6.8%) or CDV (six dams, 13.6%), with two dams having no detectable IgG to both viruses (4.5% dams). Although the presence of seronegative individuals in our breeding cohort was a concern, these rates were lower than those reported in a previous US study of hospitalized dogs, which found 19% and 50% seronegativity for CPV and CDV, respectively (18). We omitted colostrum or cord samples from cases with undetectable virus-specific IgG from further study. Additionally, in a small proportion of cases, it was not possible to collect the full clinical dataset (dam serum, colostrum, cord blood). The number of samples studied for each virus-specific IgG is listed in Table 2.

IgG titers for each dam with detectable virus-specific IgG are presented in Figures 1A, B, with each point representing a single dam and her litter. The average (mean) cord IgG titer within each litter is shown as a single point. We found that virus-specific IgG in colostrum was significantly higher (p = 0.0003 and p = 0.0029, respectively, determined by Kruskal–Wallis test) than IgG circulating in dams for both viruses, as summarized in Table 2. In contrast, the mean cord IgG titer for each litter was significantly lower than that in the corresponding dam serum. To assess individual-level variation in MatAb transfer varied at the individual level, we calculated transfer ratios for every colostrum and cord sample for which the titer was above the lower limit of quantification. The results are presented in Figure 1C. This provides clear evidence that IgG is actively concentrated in the colostrum, resulting in levels that are 8.1-fold higher for CDV and 10.7-fold higher for CPV compared with those in the dam’s serum. Analysis of cord samples showed that approximately 5% virus-specific maternal IgG is able to cross the placenta.

Endpoint titers are presented as the reciprocal dilution at which antibody binding to the antigen was detectable. Transfer ratios were calculated as the IgG endpoint titer in colostrum or cord divided by the corresponding IgG titer in the matched dam serum.

To extend this analysis, we performed CPV-based ELISAs to quantify different MatAb isotypes and subclasses. Analysis of a subset of six groups of samples (dam, colostrum, cord) confirmed that IgM and IgA titers were very low in all sample types compared with the IgG isotype; therefore, further quantification of these isotypes was not pursued (data not shown). In several species, specific IgG subclasses are preferentially transferred to neonates (19, 20). However, in dogs, there is only limited analysis of the roles of different IgG subclasses in dogs at any life stage, and no prior studies have evaluated transfer rates of maternal IgG subclasses. To begin evaluating transfer differences, we adapted our in-house indirect CPV-specific IgG ELISA to quantify CPV-specific IgG2 antibodies. IgG2 has been reclassified as a combination of IgG subclass B and IgG subclass C (21); however, commercially available secondary antibodies cannot distinguish between these subsets (22). We measured CPV-specific IgG2 MatAbs in a randomly selected subset of samples (29 dams, with matched colostrum and cords) and demonstrated that there was no preferential transfer of CPV-specific IgG2 relative to all IgGs in our cohort (Supplementary Figure S1).

Previous studies of MatAb transfer have largely focused on total IgG rather than virus-specific MatAbs (7, 9, 23, 24). Therefore, we included total IgG quantification in our study to enable comparisons with both previous reports and our virus-specific MatAb analysis. IgG titers from 24 groups in the cohort (selected for having adequate remaining sample volume) are presented in Figure 2A, with each point representing the IgG titer of a single dam, a colostrum sample, or the average cord IgG titer within one litter. The mean total IgG titers in dam serum, colostrum, and cord were 7.9, 19.5, and 0.59 mg/ml, respectively. These findings closely match previously reported values, which quantified average total IgG in dam circulation as 8.1 mg/ml and in colostrum as 19.3–20.8 mg/ml (7, 24). Cord titers were also comparable to serum levels obtained from puppies at birth, reported as 0.3 mg/ml (25, 26) and 1.2 mg/ml (23). This confirms that umbilical cord blood is a viable, noninvasive alternative to direct neonatal blood sampling for evaluating the efficiency of placental IgG transfer.

Next, we calculated the transfer ratio of total IgG via colostrum and placenta, as shown in Figure 2B. Similar to virus-specific IgG, our results indicate that puppies receive the majority of maternally derived IgG through colostrum ingestion. Unexpectedly, the concentration of total IgG in colostrum relative to dam sera was significantly lower than that of virus-specific IgG, as shown in Figure 2C. These data suggest that virus-specific IgG is selectively transported into the colostrum for transfer to the puppies. In contrast, the transfer ratios from dam serum to cord for total IgG and virus-specific IgG were highly comparable (total IgG: 0.059, CPV-specific IgG: 0.045, and CDV-specific IgG: 0.060).

Given the substantial variation in antibody titers in colostrum and cord samples relative to dam serum, we sought to identify which factors influence canine MatAb transfer. We first evaluated the relationship between IgG titers in dam serum and those in colostrum and cord samples. While studies in humans, sheep, and pigs have reported a positive correlation between maternal IgG titers and colostrum (27–29), previous studies in dogs did not identify such an association (7). Furthermore, multiple human studies have demonstrated a strong correlation between maternal and cord IgG titers, particularly for virus-specific antibodies (30–32), but this relationship has not previously been examined in dogs.

The correlations between dam serum IgG titers and cord or colostrum IgG titers for our two representative viruses are presented as scatterplots in Figure 3. Nonlinear regression analysis revealed a consistent positive trend across all virus-specific MatAb curves. Since the data were nonparametric, correlations were assessed using Spearman’s rank correlation coefficient (r), which demonstrated highly significant associations for each sample type and virus (p < 0.0001).

The strong association observed between dam and colostrum titers differed from the findings by Mila et al. (7). Notably, that earlier study quantified total IgG rather than virus-specific IgG in clinical samples. To assess whether antibody specificity influenced this relationship, we examined correlations between total IgG titers in each sample, as shown in Figure 3C. A moderate positive correlation was found between dam and colostrum total IgG titers (data normally distributed; Pearson’s correlation r = 0.7574, p = 0.0003), whereas no significant correlation was observed between dam and cord total IgG titers (data nonnormally distributed; Spearman’s correlation r = 0.3442, p = 0.1266).

To identify any demographic or clinical factors associated with MatAb transfer, we analyzed the relationship between IgG titers in each sample and parity, dam age, dam weight, and litter size. Figure 4 presents the results for CPV-specific IgG analysis, with comparable findings observed for CDV-specific IgG analysis (data not shown). Spearman’s rank correlation analysis revealed no significant associations between virus-specific IgG and any of these variables (p > 0.05). Vaccine status was also evaluated, but no differences were observed between dogs confirmed to be up to date with vaccinations (n = 20) and those not up to date (n = 6) (data not shown). A similar analysis using total IgG results (Supplementary Figure S2) likewise showed no correlations between MatAb levels and the factors examined.

This study is the first to report analysis of MatAb titers in canine cord blood samples. As shown in Figure 1, approximately 5% of circulating virus-specific IgG in the dam is transferred across the placenta. Considerable variation was observed across litters, prompting an investigation into intralitter variation in IgG titers. Figure 5 presents the CPV-specific cord IgG titers of 122 puppies from 38 litters. Of the six litters for which no cord results were available, three were from dams lacking CPV-specific IgG, and the remaining three had insufficient or uncollected samples. Although most litters had cord IgG titers that clustered closely together, nine litters exhibited a standard deviation > 1,000.

It was apparent that a proportion of puppies were born with moderately high levels of CPV-specific IgG. We investigated whether this level would be consistent with protection at birth. Previous studies have reported that a hemagglutination inhibition (HAI) titer of > 1:80 corresponds to protection (4) and that HAI and ELISA titers are highly correlated (33). Our standard positive control was submitted to the Cornell Animal Health Diagnostic Centre for quantification by HAI. This established that an HAI titer of 1:80 corresponded to an endpoint titer of 3,239 in our in-house ELISA. In Figure 5, this is highlighted as a dashed line, with samples > 1:80 HAI titer, extrapolated from the ELISA data, falling within the shaded yellow area. These results suggest that 22 (17.2%) puppies were born with sufficient IgG to provide protection against CPV at birth without ingesting any colostrum.