10 BLV⁻ and 10 BLV⁺ lactating Holstein dairy cows were enrolled in the current study (Table 1). BLV⁺ cows (as determined by the producer’s BLV milk ELISA results) were not confirmed to have PL but were selected for elevated total leukocyte counts (as determined using a Beckman Coulter counter) and an elevated proportion of circulating B cells [as determined by immunostaining for surface IgM (SIgM) on freshly isolated PBMCs] 1 week prior to the study’s initiation. BLV⁺ cows had a high proviral load (PVL) on d0 (12). BLV⁻ cows were then age and lactation matched to the 10 selected BLV⁺ cows. Both BLV⁻ and BLV⁺ cows were also re-screened for BLV infection using a commercial serum ELISA (NorthStar Cooperative) 1 week prior to the study start. BLV serum ELISAs and endpoint PCR (on DNA extracted from whole blood) to detect BLV provirus were also used on samples collected on the first and last days of the study to confirm BLV status. One BLV⁻ cow seroconverted in between enrollment diagnostics and the start of the study; this cow and her matched BLV⁺ cow were excluded from the final data analysis.

Upon study initiation on day 0 (d0), all cows received the primary KLH inoculation consisting of 1.5 mL KLH cocktail injected subcutaneously into the left side of the neck. The KLH cocktail was composed of 200 µg KLH (Sigma) in 0.75 mL 1× phosphate-buffered saline (PBS) containing 5% bovine serum albumin and 0.75 mL 20 mg/mL adjuvant dimethyldioctadecylammonium bromide (DDA) (Sigma) in 1× PBS (13). On d56, cows received the secondary KLH inoculation with 1.5 mL KLH cocktail subcutaneously injected into the right side of the neck. All protocols were reviewed and approved by the Michigan State University Institutional Animal Use and Care Committee (AUF# 04/15-061-00).

Whole blood was collected by coccygeal venipuncture on d0 before primary inoculation and on days 7, 14, 18, 21, 28, 54, and 56 after primary inoculation; after blood collection on d56, the secondary KLH inoculation was administered and blood was collected on days 60, 67, 70, 77, and 84 after primary inoculation (days 4, 11, 14, 21, and 28 after secondary inoculation). Blood for PVL quantification was collected in Vacutainer blood collection tubes containing the anticoagulant EDTA (Becton Dickinson); 1 mL aliquots of whole blood were stored at −80°C. Blood for antibody quantification and PBMC immunostaining was collected in Vacutainer blood collection tubes containing the anticoagulant ACD (Becton Dickinson). Plasma and PBMCs were isolated as previously described (10). Briefly, aliquots of plasma with 0.1% sodium azide were stored at −80°C and PBMCs were isolated using Percoll density centrifugation.

DNA was extracted from whole blood using the DNeasy Blood and Tissue kit (Qiagen) using a modified protocol. Briefly, 200 µL of whole blood, 40 µL of proteinase K, 218 µL of buffer AL and 218 µL of 100% ethanol were used instead of the recommended kit volumes. Extracted DNA was quantified using a Nanodrop-1000, and A260/280 ratios were used to assess sample purity. DNA was diluted to 30 ng/µL in elution buffer for PVL quantification. BLV PVL was determined using the Coordination of Common Motifs-qPCR to amplify the long terminal repeat (LTR) of the BLV provirus. To normalize genomic DNA input, the single-copy Bola-DRA gene was also amplified (14). In brief, 30 ng of genomic DNA were assayed using TaqMan Gene Expression Master Mix (Applied Biosystems, CA, USA) on the 7500 FAST Real-time PCR System (Applied Biosystems, CA, USA). BLV copy number and BoLA-DRA copy number were calculated using 10–1 × 10⁵ copies of the standard plasmid, which contained a copy of BLV-LTR and BoLA-DRA. Each value was calculated using the algorithm suggested by the manufacturer. PVL was the ratio of BLV copy number to BoLA-DRA copy number multiplied by 100,000 (15). PVL was expressed as BLV copy number/100,000 cells.

Anti-KLH antibodies were quantified from plasma collected on days 0, 7, 14, 21, 28, 56, 60, 67, 70, 77, and 84 after primary inoculation. Flat-bottomed 96-well ELISA plates (Thermo Fisher Scientific) were incubated with 100 µL of 1 µg/mL KLH in 50 mM carbonate/bicarbonate buffer overnight at 4°C. Plates were washed 3× with wash buffer (0.05% Tween-20 in 1× PBS) and blocked with 2% heat-inactivated horse serum in wash buffer for 1 h at 37°C. Plates were then washed 5×, and 100 µL of plasma (diluted 1:50 in blocking buffer) was added to wells and incubated for 1 h at room temperature. Plates were washed 5× and then incubated for 1 h at room temperature with 100 µL anti-bovine IgM, IgG1, or IgG2 conjugated to horseradish peroxidase (Thermo Fisher Scientific) diluted 1:10,000 in blocking buffer as previously described (10). Plates were washed 5× and incubated for 15 min at room temperature in the dark with 100 µL TMB substrate (Sigma). A total of 100 µL stop solution (2 M H₂SO₄) was added, and the optical density was measured at 450 nm using a SpectraMax M5 microplate reader. All samples were run in duplicate, and each plate included blank and naive (not exposed to KLH) controls.

Immunostaining of freshly isolated PBMCs was performed as previously described (10) on days 0, 7, 14, 21, 28, 56, 60, 67, 70, and 77 after primary inoculation. Briefly, 5 × 10⁵ PBMCs were fixed in 4% paraformaldehyde and then labeled via indirect immunofluorescence first with mouse anti-bovine primary antibodies and second with goat anti-mouse secondary antibodies. The freshly isolated B-cell population was characterized using the stain labeled “fresh B” in Table 2, and the freshly isolated T-cell population was characterized using the stain labeled “fresh T” in Table 2.

To investigate T-cell activation, 2 × 10⁶ PBMCs were cultured at 38°C and 5% CO₂ in 1 mL Roswell Park Memorial Institute (RPMI) complete media (RPMI plus 10% fetal bovine serum, 1% penicillin/streptomycin, and 1% fungizone, pH 7.4) in 24-well culture plates (Corning). PBMCs were either cultured in medium alone (NIL) for 18 h, with 200 µg/mL KLH for 18 h, or with 20 µg/mL positive control concanavalin A (CONA) for the final 6 h. All samples were treated with 20 ng/mL brefeldin A at 12 h to prevent cytokine secretion. T-cell activation was measured on days 7, 14, 56, 67, and 77.

To investigate B-cell activation, 5 × 10⁶ PBMCs were cultured at 38°C and 5% CO₂ in 3 mL RPMI complete media in 12-well culture plates (Corning) with medium alone (NIL), 200 µg/mL KLH, or with positive control 20 ng/mL phorbol 12-myristate 13-acetate and 400 ng/mL ionomycin (P/I) for 18 h. B-cell activation was measured on days 18, 54, and 70.

PBMCs were labeled with four-color stains to investigate IFNγ production by T cells (“IFNγ” in Table 2), IL4 production by T cells (“IL4” in Table 2), and BLV expression in B cells (“CD45R0” and “CD25” in Table 2). Immunostaining of cultured PBMCs was performed as previously described (10). Briefly, indirect immunostaining for surface receptors on live PBMCs was done as described for freshly isolated PBMCs. PBMCs were then fixed and permeabilized using a fixation/permeabilization kit according to manufacturer’s instructions (eBioscience), and PBMCs were labeled with a primary mouse anti-bovine or anti-BLV monoclonal antibody and subsequently labeled with a goat anti-mouse secondary antibody. Labeled PBMCs were stored at 4°C overnight until flow cytometry analysis.

Data were acquired using an Accuri C6 flow cytometer equipped with CSampler (Becton Dickinson). Data were compensated using the Accuri C6 software (Becton Dickinson) and then exported for analysis using FCS Express 4 (De Novo Software). PBMCs were initially selected using forward and side scatter gating to exclude debris, and cells were identified gating on lineage-specific markers (CD4, CD8, γδ, or SIgM). Expression of IFNγ, IL4, CD45R0, MHCII, CD5, and CD25 were then subsequently characterized within specific cell types, and marker expression was expressed as either the mean relative percent or the mean fluorescence intensity (MFI). Unless noted, all referenced B cells are SIgM⁺MHCII⁺. CD5 labeling on fresh PBMCs exhibited three populations; the center population was denoted “CD5^dim” and the far right population was denoted “CD5^bright” (Figure S1 in Supplementary Material).

Statistical analysis was performed using SAS 9.4 (SAS Institute). Antibody and fresh PBMC data were analyzed using repeated measures ANOVA with BLV and time as fixed effects and post hoc Bonferroni correction for pairwise comparisons. Cultured PBMC data were analyzed using repeated measures ANOVA with BLV and stimulant as fixed effects and post hoc Bonferroni correction for pairwise comparisons. A Tukey test was used to analyze the difference in MHCII or CD25 MFIs between different B-cell subsets. Outliers were detected using Grubbs test; both outliers and their matched cow were eliminated from analysis. Significance was determined as p < 0.05.

To characterize the strength of the humoral immune response in BLV⁺ cows, the relative quantities of KLH-specific IgM, IgG1, and IgG2 were measured. As expected, both BLV⁺ and BLV⁻ cows produced KLH-specific IgM, IgG1, and IgG2 after primary and secondary KLH inoculations (p < 0.001 for all isotypes). Interestingly, BLV⁺ cows had less IgM (p = 0.0045) after primary KLH inoculation at all time points, including d0. However, BLV⁺ cows made IgM at levels equal to BLV⁻ cows after the secondary KLH inoculation (p = 0.7742) (Figure 1A). In contrast to IgM production, BLV⁺ and BLV⁻ cows produced equivalent levels of IgG1 (p = 0.6715) (Figure 1B) and IgG2 (p = 0.9437) (Figure 1C) after both primary and secondary KLH inoculations. These data support recent evidence that BLV infection interferes with IgM antibody production after both a primary immune challenge and a common vaccine booster injection (10).

To investigate how circulating B-cell populations changed in response to immune stimulation, freshly isolated PBMCs were immunostained to examine both CD45R0 and CD5 B-cell populations in BLV⁺ and BLV⁻ cows. Unless noted in the text, all B cells are SIgM⁺MHCII⁺. Surprisingly, the SIgM⁺ B-cell population from BLV⁺ cows did steadily increase over time (p < 0.05) after both primary and secondary KLH inoculations; however, this trend was not observed in BLV⁻ cows. As expected, BLV⁺ cows had significantly more circulating B cells than uninfected, age-matched cows (p < 0.0001) (Figure 2C). Some B-cell populations did not shift in response to KLH inoculation: in both BLV⁺ and BLV⁻ cows, the mean relative percent of MHCII⁺SIgM⁺ B cells and CD45R0⁺ B cells, as well as the CD5 MFI on B cells, all remained constant. However, these populations were altered in BLV⁺ cows. While BLV⁺ cows exhibited a higher proportion of MHCII⁺SIgM⁺ B cells (p = 0.0084), they also demonstrated a large reduction in CD45R0⁺ B cells compared to uninfected, age-matched controls (p < 0.0001) (Figure 2A). When analyzing CD5 expression on B cells, two distinct positive populations were observed, which we denoted CD5^dim and CD5^bright (Figure S1 in Supplementary Material). Interestingly, B cells from BLV⁺ cows had higher CD5 expression on CD5^dim+ B cells (p = 0.0002) but lower CD5 expression on CD5^bright+ B cells (p = 0.0017) (Figure 2B).

Although the surface expression of CD5 did not change over time, both CD5^dim+ and CD5^bright+ B-cell populations responded to KLH inoculation (p < 0.05); while the CD5^dim+ B-cell population steadily increased after inoculation, the CD5^bright+ B-cell population sharply declined. In addition, BLV⁺ cows displayed a lower CD5^dim+ B-cell population (p = 0.0043) (Figure 2D), but the CD5^bright+ B-cell population was equal between BLV⁺ and BLV⁻ cows (p = 0.9611) (Figure 2E).

We were also interested in the effect of BLV infection on MHCII surface expression. Surprisingly, MHCII surface expression actually fluctuated in response to KLH exposure on all measured B-cell types (p < 0.05). BLV⁺ cows presented lower MHCII surface expression on CD5^dim+ B cells (p = 0.0168) (Figure 2G) and a trending lower MHCII surface expression on all B cells (p = 0.0722) (Figure 2F). In contrast, BLV⁺ and BLV⁻ cows exhibited equivalent MHCII surface expression on CD45R0⁺ (p = 0.7919) (data not shown) and CD5^bright+ B cells (p = 0.5662) (Figure 2H).

To determine if KLH stimulation in vitro induced BLV expression in infected B cells, PBMCs were cultured in the presence of KLH or P/I-positive control stimulation and BLV expression (by immunostaining for viral protein BLV gp51) and B-cell activation (by immunostaining for CD25) were characterized. While B-cell culture was done on d18, 54, and 70, there was no difference in results from different time points. Thus, results presented are from d70.

There was no overall difference between BLV⁺ and BLV⁻ cows when comparing the mean relative percent of CD25⁺ B cells (p = 0.5849), and both KLH and P/I stimulation increased the proportion of CD25⁺ B cells. However, B cells from BLV⁺ cows exhibited only a trending (p = 0.053) rise in CD25⁺ B cells after KLH stimulation; in fact, the proportion of CD25⁺ B cells in KLH-stimulated cultures from BLV⁺ cows was significantly lower than KLH-stimulated cultures from BLV⁻ cows (Figure 3A).

In contrast to the mean relative percent of activated B cells, the surface expression of CD25 on B cells was significantly affected by BLV status (p = 0.0044). Only B cells from BLV⁻ cows demonstrated higher CD25 MFI after P/I stimulation; B cells from BLV⁺ cows actually decreased CD25 surface expression after KLH stimulation, and the surface expression of CD25 was lower on B cells from BLV⁺ cows, in comparison to B cells from BLV⁻ cows, after both KLH and P/I stimulation (Figure 3B).

PBMCs were also cultured to investigate if KLH stimulation could induce BLV expression in B-cell populations from BLV⁺ cows. KLH failed to trigger BLV expression in any measured B-cell subset. Surprisingly, KLH stimulation actually reduced the proportion of BLV-expressing CD45R0⁺ B cells. However, CD45R0 expression on B cells did not have a significant effect on BLV expression overall (p = 0.8236) (Figure 4A). In contrast, CD25 expression on B cells did have a significant effect on BLV expression (p = 0.0013); CD25⁺ B cells under all culture conditions exhibited a much greater proportion of BLV expression in comparison to CD25⁻ B cells (Figure 4B).

We also explored how BLV expression affected B-cell phenotypes in vitro. When comparing B cells from BLV⁻ cows to B cells from BLV⁺ cows that did (BLV⁺) or did not (BLV⁻) express BLV protein gp51, BLV⁺ B cells expressed much higher MHCII in all culture conditions in comparison to both BLV⁻ B cells and B cells from BLV⁻ cows (Figure 4C). When investigating the effect of BLV expression on CD25 surface expression, CD25 MFI increased in P/I-stimulated cultures as expected, but only on B cells from BLV⁻ cows and BLV⁺ B cells from BLV⁺ cows; BLV⁻ B cells from BLV⁺ cows failed to increase the CD25 MFI (Figure 4D). Taken together, these results suggest that B cells from BLV⁺ cows are less reactive to in vitro stimulation than B cells from BLV⁻ cows; BLV expression in vitro is more prevalent in activated B cells; and B cells from BLV⁺ cows demonstrate different phenotypes in comparison to B cells from BLV⁻ cows, both in BLV⁻ and BLV⁺ B-cell subsets.

In order to assess how the BLV PVL changed over time after KLH+ DDA exposure, DNA was extracted from whole blood from BLV⁺ cattle and qPCR was used to measure PVL. The average BLV PVL on d0 before inoculation was almost 96,000 copies/10⁵ cells; after both primary and secondary KLH inoculations, the PVL sequentially increased over time (p < 0.0001) (Figure 5). BLV⁺ cows exhibited the largest increase in PVL on the final day of collection (28 days post-exposure for primary and 21 days post-exposure for secondary); in both cases, the PVL increased by over 18,000 copies/10⁵ cells.

Although BLV most commonly infects B cells, previous research has suggested that T-cell immunity is also compromised in BLV⁺ cows (6). To investigate the circulating effector and memory T-cell population in BLV⁺ cows after KLH inoculation, we labeled freshly isolated PBMCs to measure CD45R0 expression on CD4⁺, CD8⁺, and γδ⁺ T cells. BLV⁺ cows consistently exhibited lower relative proportions of CD4⁺, CD8⁺, and γδ⁺ T cells (p < 0.05) (Figure 6D), although these did not change over time. However, both CD4⁺CD45R0⁺ (Figure 6A) and CD8⁺CD45R0⁺ (Figure 6B) T-cell populations significantly responded to primary and secondary KLH inoculations (p < 0.05): both populations increased after inoculation, although there was no difference between CD4⁺CD45R0⁺ (p = 0.3816) or CD8⁺CD45R0⁺ (p = 0.4237) populations in BLV⁺ and BLV⁻ cows. In contrast to classical T cells, the γδ⁺CD45R0⁺ population did not significantly change over time after KLH exposure; however, BLV⁺ cows demonstrated a greatly diminished γδ⁺CD450⁺ T-cell population (p < 0.0001) (Figure 6C). These data indicate that while BLV infection does not appear to affect the classical effector/memory T-cell compartment, BLV infection may negatively impact the γδ effector/memory T-cell compartment.

In addition to investigate the circulating effector/memory T-cell compartment, we examined whether T cells from BLV⁺ cows generated IFNγ or IL4 in response to stimulation in vitro. While in vitro T-cell activation was measured on d7, 14, 56, 67, and 77, there was no difference between time points; thus, the IFNγ data presented in Figure 7 is from d56. Overall, KLH stimulation failed to increase the proportion of IFNγ⁺ cells in any T-cell subset and actually decreased the proportion of IFNγ⁺CD8⁺ T cells from both BLV⁺ and BLV⁻ cows. There was no overall difference between BLV⁺ and BLV⁻ cows in the proportion of IFNγ⁺CD4⁺ T cells (p = 0.150) (Figure 7A) or CD8⁺ T cells (p = 0.112) (Figure 7B). Unlike classical T cells, γδ⁺ T-cell IFNγ production was significantly higher in BLV⁺ cows (p = 0.0007) (Figure 7C).

We also measured the IFNγ MFI as a measure of reactivity to in vitro stimulation. While there was no difference in IFNγ MFI from CD4⁺ or γδ⁺ T cells between BLV⁺ and BLV⁻ cows (data not shown), BLV status did have a significant effect on IFNγ MFI from CD8⁺ T cells (p = 0.044): reactive CD8⁺ T cells from BLV⁺ cows were actually producing more IFNγ than reactive CD8⁺ T cells from BLV⁻ cows (Figure 7D).

We also studied IL4 production by T-cell subsets in BLV⁺ cows. Similar to the in vitro IFNγ results, in vitro IL4 production was not different between time points; thus, the IL4 data presented in Figure 8 are from d77. Surprisingly, in vitro IL4 production was not similar to what was observed with in vitro IFNγ production. BLV status had a significant or trending significant effect on the proportion of IL4-producing cells within CD4⁺ (p = 0.0009), CD8⁺ (p = 0.0006), and γδ⁺ T-cell (p = 0.0879) populations. In all three T-cell populations, BLV⁺ cows consistently exhibited a higher proportion of IL4-producing T cells (Figures 8A–C). However, reactive CD8⁺ T cells from BLV⁺ cows actually exhibited lower IL4 expression in comparison to reactive CD8⁺ T cells from BLV⁻ cows (p < 0.0001) (Figure 8D). The IL4 MFI from CD4⁺ and γδ⁺ T cells was not different between BLV⁺ and BLV⁻ cows (data not shown). Taken together, these results imply that T cells from BLV⁺ cows are capable of producing both IFNγ and IL4 after in vitro cell culture; however, the balance in BLV⁺ cows may favor a greater proportion of IL4-producing T cells.