ABSTRACT
Canine multicentric lymphoma (CML) is one of the most common malignancies in dogs. Although breed risk is important, environmental factors such as herbicides have also been implicated. The objective of this study was to determine whether genotoxic exposures to the herbicides 2,4‐D and glyphosate are associated with CML, using dogs from the Golden Retriever Lifetime Study cohort. We measured urinary concentrations of glyphosate and 2,4‐D in golden retrievers with CML and matched unaffected controls at two time points: at the time of diagnosis and 1 year prior to diagnosis. To assess the genotoxic potential of herbicide exposures, we used reverse dosimetry from urinary concentrations to estimate plasma concentrations. We then assessed the genotoxicity of these herbicide concentrations towards healthy canine peripheral blood mononuclear cells (PBMC's) in vitro using the CometChip assay, with and without canine liver microsomes. All dogs had detectable urinary exposures to 2,4‐D (7.3–42.9 ng/mg creat) and glyphosate (0.4–80.7 ng/mg creat), with no differences between cases and controls at either time point. Both 2,4‐D and glyphosate were genotoxic to canine PBMCs at concentrations of 0.10 μM and higher, with no consistent effects of canine liver microsomes on herbicide genotoxicity. No dogs reached estimated genotoxic plasma concentrations for glyphosate, but 4 of 30 golden retrievers with CML (13.3%) and 2 of 30 control dogs (6.7%) reached estimated genotoxic 2,4‐D exposures (p = 0.67).
Keywords: cancer risk, canine, DNA damage, lymphosarcoma, weed killers
1. Introduction
Canine multicentric lymphoma (CML) is the most diagnosed cancer in dogs [1], and shares many biological, histological, and clinical characteristics with human non‐Hodgkin lymphoma (NHL) [2, 3]. Both NHL and CML share geographical distributions [4, 5], suggesting shared environmental risk factors between species.
Both human NHL and CML have been associated epidemiologically with exposure to herbicides [6, 7, 8]. Specifically, 2,4‐dichlorophenoxyacetic acid (2,4‐D), a selective broad‐leaf weed killer, and glyphosate, a broad‐spectrum weed killer found in the common household brand Roundup, have been linked to NHL [9, 10, 11]. Similarly, CML has been associated with commercially applied pesticides [12], and household 2,4‐D use [13], but glyphosate has not been evaluated. These chemicals have been evaluated in human NHL and CML risk due to their ability to damage DNA, that is, their genotoxic potential. Both 2,4‐D and glyphosate are genotoxic to human lymphoid cells [14, 15, 16, 17], but their genotoxicity to canine lymphoid cells, especially at relevant in vivo exposures, is not understood.
Our overall hypothesis is that CML is associated with genotoxic herbicide exposures. The specific aims of this study were to (1) determine whether in vivo urinary 2,4‐D and glyphosate concentrations in golden retriever dogs with CML were higher than those of matched controls; (2) use reverse dosimetry to estimate plasma herbicide exposures; and (3) determine whether estimated plasma herbicide exposures reach genotoxic concentrations, as measured by the CometChip assay in vitro, in more dogs with CML compared to controls.
2. Methods
2.1. Dog Demographics
We used a nested case–control study design that leveraged banked urine samples from an existing longitudinal cohort, the Golden Retriever Lifetime Study (GRLS) [18]. Dogs included in the GRLS cohort were purebred golden retrievers that were enrolled from 6 months to 2 years of age between 2012 and 2015 and followed with yearly vet appointments, owner questionnaires, and annually banked samples of blood and urine [18, 19]. Our case population was 30 randomly selected golden retrievers that had been diagnosed with CML from 2015 to 2022; unaffected controls from the same cohort were matched to cases by age at the time of sampling and sex/neuter status.
Dogs with CML were diagnosed by cytology or histopathology read by a board‐certified veterinary pathologist; immunophenotyping was not required but was included when available. For our study, we had access to urine samples from the time of lymphoma diagnosis (T0) and 1 year prior to diagnosis (T‐1y) for cases, and from comparable dates for matched controls, within a 1–2‐month sampling window to account for seasonality. Dog addresses were subsequently categorised by United States Department of Agriculture (USDA) plant hardiness zones, REF to evaluate for possible differences in growing seasons between cases and controls. Hardiness zones (all within the continental US) were encoded as cold (zones 3–5), moderate (zones 6–8), and warm (zones 9–11).
2.2. Urinary Measurements
Urinary concentrations of 2,4‐D and glyphosate at T0 and T‐1y were measured using commercial ELISAs (ABRAXIS 2,4‐D ELISA Microtiter Plate No. 54003A; ABRAXIS Glyphosate ELISA Microtiter Plate No. 500205/500 087). Urinary creatinine was also measured using ELISA (BioAssay Systems QuantiChrom Creatinine Assay Kit DICT‐500). All analytes were normalised to urine creatinine for each subject to control for individual differences in urine concentration [20].
2.3. Estimated Plasma Exposures
Since 2,4‐D and glyphosate are primarily excreted by the kidneys [16, 20, 21, 22, 23], we were able to use reverse dosimetry to estimate plasma exposures from observed urinary concentrations [24, 25] using the following equations:
Internal dose (μmole) = urine concentration (μmole/L)×estimated daily urine volume (L)*. We assumed an estimated daily urine volume of 30 mL/kg for each dog [26].
Estimated plasma concentration = internal dose (μmole)/ (Vd in L/kg×kg body weight). We assumed a Vd (volume of distribution) of 0.15 L/kg for 2,4‐D, based on a weighted average from human subjects [27, 28]; we assumed a Vd of 0.602 L/kg for glyphosate based on available data from rats [21].
2.4. Peripheral Blood Mononuclear Cell (PBMC) Preparation
Fresh whole blood samples in EDTA were collected from a pool of three healthy pet dogs belonging to students, faculty, and staff at the University of Wisconsin‐Madison School of Veterinary Medicine under approved IACUC protocol V005450. PBMCs were prepared by standard density centrifugation at 400 g for 30 min in Histopaque 1077 (Sigma Aldrich Cat No. 10771). Cell pellets were washed and resuspended in 1X phosphate buffered saline, and PBMCs were resuspended in RPMI media containing 10% foetal bovine serum [29].
2.5. Genotoxicity Testing
To test for genotoxicity, we used the CometChip assay, a high‐throughput version of the traditional alkaline comet assay that reduces intra‐assay variability [30]. The comet assay detects single and double‐strand breaks in DNA to assess the genotoxicity of a chemical. All CometChip experiments were performed using a standardised protocol [31]; the assay developer, Dr. Bevin Engelward, generously provided our laboratory with our own CometChip ‘stamp’, which generates microwells in 96‐well plates that aid in even cell distribution.
Isolated PBMCs were first exposed to an established genotoxic dose of 250 μM 2,4‐D [14, 15] or 500 uM glyphosate [16, 17], or respective vehicles, for 24 h to determine time to maximal genotoxicity. This time point was used in all subsequent concentration‐dependent dosing experiments, which incorporated estimated plasma exposures determined from reverse dosimetry. In addition, we repeated experiments with the addition of pooled canine liver microsomes to assess for possible effects of bioactivation or detoxification of each herbicide [32] on in vitro genotoxicity towards PBMCs. These microsomal experiments were assessed at 2 h to retain microsome viability and were run in parallel with non‐microsomal experiments using the same pool of PBMCs.
2.6. Statistical Analysis
Urinary measurement chemical concentrations between cases and controls at each time point, and urinary chemicals between time points across all dogs, were compared using Wilcoxon matched‐pairs signed‐rank tests. An exploratory analysis was also performed of aggregate urinary herbicides among B‐cell and T‐cell lymphoma subtypes and controls, using Kruskal–Wallis testing with Dunn's multiple comparisons post hoc test.
To assess the relationship between owner‐reported herbicide use and documented exposures, we compared urinary 2,4‐D and glyphosate concentrations between owners that reported yes or no to weed control treatment in the yard, garden, or surrounding area (including aerial spraying) at each time point in the annual GRLS questionnaire, using Mann–Whitney U tests. Urinary chemical concentrations by season (winter: December–February; spring: March–May; summer: June–August, and fall: September–November) were compared using a Kruskal–Wallis test with Tukey's post‐hoc test.
To establish genotoxic thresholds, DNA damage at each herbicide concentration was compared to the vehicle using one‐way ANOVA with Dunnett's multiple comparisons tests. DNA damage with and without liver microsomes was assessed using one‐way ANOVA with Sidak multiple comparison tests. The proportion of cases and controls that reached estimated genotoxic thresholds for each herbicide at any timepoint were evaluated using Fisher's exact tests and odds ratios (OR) with 95% confidence intervals (95% CI). All analyses were performed with commercially available software (Prism 9, GraphPad Software, San Diego CA), with p < 0.05 considered significant.
3. Results
3.1. Dog Demographics
Enrolled golden retrievers were a median of 7.1 years old for both cases and controls (Table 1) and lived across the United States (Figure 1). Of cases, about 47% were B‐cell lymphoma, 37% T‐cell, and 17% uncharacterised. Urine and blood samples were collected from 30% of CML cases in winter (December–February), 30% in spring (March–May), 13% in summer (June–August), and 27% in fall (September–November). For matched controls, 23% were sampled in winter, 17% in spring, 37% in summer, and 23% in fall (Table 1). Most of the dogs lived in cold to moderate USDA hardiness zones (zones 3–8; 70% of CML cases and 87% of controls; Table 1) where herbicide use would be expected to be seasonal.
TABLE 1.
Demographic characteristics of a case–control study of golden retriever dogs with multicentric lymphoma.
| Multicentric lymphoma cases | Unaffected controls | |
|---|---|---|
| Number of dogs | 30 | 30 |
| Median age at diagnosis or matched enrollment (observed range) | 7.1 years(3.5–10.6) | 7.1 years(3.2–10.9) |
| Sex and neuter status (number of dogs) |
FS 8 FI 4 MN 12 MI 6 |
FS 8 FI 4 MN 12 MI 6 |
| Immunophenotype of lymphoma (number of dogs) |
B cell, overall, 14 T cell, overall, 11 Not characterised 5 |
Not applicable |
| Sampling season a (number of dogs) |
Winter 9 Spring 9 Summer 4 Fall 8 |
Winter 7 Spring 5 Summer 11 Fall 7 |
| USDA plant hardiness zone b (number of dogs) | Cold 6 | Cold 3 |
| Moderate 15 | Moderate 23 | |
| Warm 9 | Warm 4 |
Abbreviations: FI, female intact; FS, female spayed; MI, male intact; MN, male neutered.
Winter: December–February; Spring: March–May; Summer: June–August; Fall: September–November.
Cold: zones 3–5; Moderate: zones 6–8; Warm: zones 9–11.
FIGURE 1.
(A) Distribution of 30 golden retrievers with multicentric lymphoma, recruited from the Golden Retriever Lifetime Study for a nested case–control study. Shading indicates the number of individual dogs per state. (B) Distribution of 30 matched control golden retrievers without multicentric lymphoma, recruited from the Golden Retriever Lifetime Study for a nested case–control study. Shading indicates the number of individuals of dogs per state.
3.2. Urinary Chemical Measurements
In the year prior to diagnosis, urinary concentrations of 2,4‐D were not significantly higher in CML cases (median: 18.0 ng/mg creat) when compared to matched controls (median: 17.4 ng/mg creat; p = 0.69; Figure 2). There were also no differences in urinary 2,4‐D concentrations at the time of diagnosis between cases (median: 18.6 ng/mg creat) and controls (median: 16.4 ng/mg creat; p = 0.18; Figure 2).
FIGURE 2.
Urinary 2,4‐D concentrations in golden retrievers with multicentric lymphoma and matched controls at two time points: 1 year prior to diagnosis or matched control time point (T‐1y) and at diagnosis or matched control time point (T0).
For glyphosate, urinary concentrations were not significantly higher in CML cases in the year prior to diagnosis (median: 3.60 ng/mg creat) when compared to matched controls (median: 4.08 ng/mg creat; p = 0.73; Figure 3). Further, urinary glyphosate concentrations were not different at the time of diagnosis in cases (median: 3.76 ng/mg creat) and controls (median: 4.34 ng/mg creat; p = 0.98; Figure 3).
FIGURE 3.
Urinary glyphosate concentrations in golden retrievers with multicentric lymphoma and matched controls at two time points: 1 year prior to diagnosis or matched control time point (T‐1y) and at diagnosis or matched control time point (T0).
Notably, urinary herbicide concentrations were stable between measured time points across all dogs, with no significant differences between median urinary 2,4‐D at T‐1y (17.0 ng/mg creat) compared to T0 (16.8 ng/mg creat; p = 0.54), or between median urinary glyphosate at T‐1y (3.9 ng/mg creat) and T0 (4.0 ng/mg creat; p = 0.47; Figures 2 and 3).
When we analysed total molar aggregate herbicides (in nmol/mg creat) in the urine of golden retrievers by lymphoma subtype, there were no significant differences at either T‐1y or T0 among cases of T‐cell lymphoma, B‐cell lymphoma, and controls, although subgroups were small (Supplemental Figure S1).
3.3. Estimated Plasma Herbicide Concentrations and In Vitro Genotoxicity
Using reverse dosimetry from urine concentrations, estimated plasma 2,4‐D exposures across both time points ranged from 0.006 to 0.287 μM in cases (median: 0.046 μM) and 0.005 to 0.105 μM in controls (median: 0.046 μM). Estimated plasma glyphosate exposures ranged from < 0.0005 to 0.085 μM in cases (median: 0.003 μM) and 0.0003 to 0.029 μM in controls (median: 0.003 μM).
In vitro time course experiments established an optimal time point of 6 h for DNA damage assessments (data not shown). There was significant DNA damage towards canine PBMCs compared to vehicle at 2,4‐D concentrations of 0.10 μM and higher (Figure 5A). We saw no consistent enhancement or decrease in genotoxicity with the addition of canine liver microsomes (Supplemental Figure S2). Estimated plasma 2,4‐D concentrations reached the genotoxic threshold of 0.10 μM at any time point in 4 of 30 dogs with CML (13.3%) and 2 of 30 controls (6%; OR: 2.15, 95% CI: 0.46–11.91; p = 0.67).
FIGURE 5.
(A) Genotoxicity of 2,4‐D (measured as single and double strand DNA breaks) in fresh canine peripheral blood mononuclear cells (PBMC's). Cells were incubated with 2,4‐D for 6 h. Error bars represent standard deviations. *** Significantly different from vehicle, p < 0.001. (B) Genotoxicity of glyphosate (measured as single and double strand DNA breaks) in fresh canine PBMCs. Cells were incubated with glyphosate for 6 h. Error bars represent standard deviations. *p = 0.04, ***p < 0.0001; both compared to vehicle.
Glyphosate was also genotoxic towards canine PBMCs at glyphosate concentrations of 0.10 μM and higher (Figure 5B), with no consistent effect of canine liver microsomes on glyphosate genotoxicity in vitro (Supplemental Figure S3). Estimated genotoxic plasma glyphosate concentrations of 0.10 μM were not reached in any of the 30 cases or 30 controls across either time point.
3.4. Herbicides by Owner Reported Use and Season
Owners reported using herbicides in the previous year prior to sampling in 50% of cases and 63% of control dogs (p = 0.43). Canine urinary 2,4‐D and glyphosate concentrations did not differ by owner‐reported household herbicide use at either time point (Table 2). Urinary concentrations of 2,4‐D and glyphosate also did not differ by season of sampling, either measured in aggregate (Figure 4) or by individual herbicides (Supplemental Table 1).
TABLE 2.
Canine urinary herbicide concentrations by owner‐reported household herbicide use at two time points sampled approximately 1 year apart (T‐1y and T0). Urinary herbicides are reported as medians with observed ranges.
| Sampling | Herbicide users | Non‐herbicide users | p |
|---|---|---|---|
|
Urinary 2,4‐D T‐1y |
18.3 ng/mg creat (10.5–37.8 ng/mg creat) |
17.4 ng/mg creat (10.5–33.6 ng/mg creat) |
0.38 |
|
Urinary 2,4‐D T0 |
18.4 ng/mg creat (7.4–138.4 ng/mg creat) |
16.5 ng/mg creat (7.3–42.6 ng/mg creat) |
0.32 |
| Urinary glyphosate T‐1y |
3.6 ng/mg creat (0.8–38.4 ng/mg creat) |
5.4 ng/mg creat (0.8–80.7 ng/mg creat) |
0.21 |
| Urinary glyphosate T0 |
3.9 ng/mg creat (0.9–26.7 ng/mg creat) |
4.1 ng/mg creat (0.4–24.8 ng/mg creat) |
0.69 |
FIGURE 4.
Urinary herbicides (molar aggregate of 2,4‐D and glyphosate) by season of sampling in golden retriever dogs.
The GRLS collects longitudinal data on CML outcomes for all enrolled dogs. Only one control dog from our study population was subsequently diagnosed with CML, a full 5.3 years after T0 samples were collected. Given the time elapsed between sampling time and CML outcome in this dog, and the uncertainties about lag times between chemical exposures and cancer outcomes in dogs, we retained this control dog in our primary analyses. However, we also performed a secondary analysis of all outcomes censoring this dog and its matched case, and P values were stable (data available upon request).
4. Discussion
2,4‐D (a phenoxy herbicide) and glyphosate (an organophosphate) represent two distinct classes of herbicides, and we found both in the urine of virtually all pet golden retrievers that we sampled. We found uncorrected urine concentrations of 2,4‐D ranging from 5.8 to 260 ng/mL (data not shown), which is consistent with reports in healthy pet dogs of various breeds exposed to 2,4‐D from the environment (mean 24 ng/mL) [22, 33]. However, one study found corrected median urinary 2,4‐D concentrations of 50 ng/mg creat in pet dogs within 48 h after 2,4‐D lawn treatments [34], which is larger than all corrected urinary 2,4‐D concentrations in our healthy controls (ranging from 7.4 to 35.2 ng/mg creat across both time points). The urinary half‐life of 2,4‐D in dogs is about 4 days [35]; consistent with this, the highest urinary concentrations of 2,4‐D in dogs were previously detected within 2 days of known herbicide application, with a marked decline within 16 days post‐exposure [33]. Unfortunately, we did not have data on the timing of the most recent 2,4‐D applications in our golden retriever population. Subsequent studies should consider using pooled urine samples from multiple days throughout the year, which were not available from the GRLS cohort.
As for glyphosate, we found median urinary concentrations in our golden retriever population of 3.6–4.3 ng/mg creat, which is comparable to mean urinary concentrations of 10.6 ng/mg creat recently reported in 30 healthy pet and shelter dogs [22]. That study also reported a range of uncorrected urinary glyphosate concentrations in those dogs of 0.4–49.1 ng/mL [22]. For comparison, we found uncorrected median glyphosate urinary concentrations of 9.2 ng/mL (range: 1.1–290 ng/mL) across all golden retrievers (data not shown). Relatively higher urinary glyphosate concentrations in some dogs in our population could be because fewer dogs in the previous study were pet dogs, with possibly lower exposure to treated lawns or public parks.
Urinary concentrations of 2,4‐D and glyphosate were not different between golden retrievers with CML and unaffected controls. This is in contrast to a positive ecological study, which found that ‘moderate’ herbicide use, based on farm surveys by census blocks, was associated with the diagnosis of lymphoma in dogs of various breeds in the UK [8]. We did not have similar integrated annual herbicide and agricultural census surveys by zip code in the United States for comparison. Instead, we directly measured herbicide concentrations in the urine as a surrogate for systemic exposure, with data from two time points. However, these conflicting results demonstrate the need for better integrated measures of direct chemical exposures over time in dogs with and without CML.
The prevalence of B‐cell or T‐cell CML immunophenotypes can vary by geographic region [36], which could suggest distinct environmental risk factors. Therefore, we performed an exploratory analysis of urinary herbicides by immunophenotype subgroup. While we did not detect a difference in aggregate urinary herbicides between B‐cell and T‐cell CML subtypes, these subgroups were small. Environmental risk factors for NHL can differ by immunophenotype in people [37], and follow‐up studies are needed to evaluate chemical exposures in more immunophenotyped CML cases and matched controls from the GRLS cohort. Urine concentrations of 2,4‐D and glyphosate did not change significantly over sampling times approximately 1 year apart across all dogs. This could reflect consistent daily routines for many pet dogs. Surprisingly, urinary 2,4‐D and glyphosate concentrations were not lower in dogs whose owners reported no herbicide use. Outdoor herbicide exposures in these dogs could be due to airborne particle drift, seepage of chemicals from neighbouring lawns, or contact during walks through parks or other lawns. There may also be problems with owner recall or awareness of chemical applications on their property.
Urine concentrations of 2,4‐D and glyphosate also did not differ by season, which was surprising. This might reflect other sources of 2,4‐D and glyphosate besides seasonal lawn applications. Non‐seasonal herbicide exposures could result from persistent chemical residues in the outdoor environment (including crops, soil, ground water, and air) [38, 39], surface contamination of fruits and vegetables [40], or even from dog food, in which glyphosate has been detected [41]. Unfortunately, we were not able to include analyses of pet food herbicide concentrations in the current study. Follow‐up studies should assess the relationship between herbicide residues in fed foods and urinary herbicides in dogs. Additionally, although cases and controls were matched for seasonal months of sampling, they were not matched for local plant hardiness zones, and this additional variable could have influenced the seasonality of herbicide exposures across the population.
Our most important finding is that both 2,4‐D and glyphosate are lymphoid genotoxins in dogs, with a threshold of 0.10 μM for each herbicide. 2,4‐D has been associated with chromosomal aberrations in workers exposed to 2,4‐D, although plasma 2,4‐D concentrations were not measured [42]. In vitro, 2,4‐D does not cause DNA strand breaks at 1–10 μM in human PBMCs from non‐smokers. However, DNA damage is seen at 10 μM 2,4‐D exposures in PBMCs from human smokers [15], and the interaction between second‐hand smoke and 2,4‐D exposures deserves further exploration in pet dogs. The genotoxicity of 2,4‐D could be mediated, at least in part, by oxidative stress [43]. Therefore, follow‐up studies should include the addition of the bacterial base excision repair enzyme Fpg in the comet assay, which can detect oxidative DNA damage prior to overt DNA strand break [44].
As for glyphosate, we also found a genotoxic threshold of 0.10 μM in canine PBMCs. Canine PBMCs may be much more sensitive to glyphosate than human PBMCs, which have a reported genotoxic threshold of 250 μM glyphosate [17]. However, there were some methodological differences between our canine and the previous human study (e.g., 24‐h incubation times for human cells versus 6 h for our canine cells), and we did not run human PBMCs in parallel with our canine cells for comparison. Glyphosate genotoxicity is likely mediated by direct oxidative DNA damage, seen in human PBMCs at 250 μM exposures [17]. Oxidative DNA damage in canine PBMCs could also be further assessed by including Fpg in future comet assay experiments with glyphosate.
Based on our reverse dosimetry assumptions, we found that about 13% of 30 golden retrievers with CML and 7% of 30 matched controls had estimated genotoxic plasma concentrations of 2,4‐D. A post‐hoc sample size calculation estimates that 391 cases and 391 controls would be needed to show this observed difference, if repeatable, to be significant. In any case, it is concerning that herbicide exposures are ubiquitous in pet golden retrievers, and about 10% of our dog population overall might have had genotoxic 2,4‐D exposures.
Our study has important limitations, starting with a limited sample size. We did use a single dog breed, which eliminated between‐breed risk as a co‐variate. However, we only included urine samples at two time points over 1 year; the latency period from chemical exposure to CML development is unclear and might not have been captured in our 1‐year sample window. In human NHL, latency periods are as short as 1.5 years for occupational exposures to benzene [45], and as long as 20 years for other carcinogenic chemicals [46], but no latency period data are available for NHL and herbicides. We also had to make some assumptions in our reverse dosimetry plasma estimations. While we accounted for each dog's individual body weight, we had to extrapolate volume of distribution data for both herbicides from human and rodent studies. It would be ideal to directly measure plasma herbicide concentrations in these dogs from the GRLS cohort as a follow‐up. In addition, we only used one type of genotoxicity, DNA strand breaks, and subsequent studies could also include tests for oxidative stress, blood micronucleus test, and chromosomal abberations [47, 48]. Finally, one matched control dog recently developed CML, more than 5 years after the T0 samples were collected. Given the uncertainties between lag times for chemical exposures and CML outcomes in dogs, we included this dog in primary analyses. We also performed secondary analyses that excluded this dog and its matched case for all outcomes; p‐values were stable and our conclusions were not altered.
Overall, we found that virtually all pet golden retrievers studied had evidence of 2,4‐D and glyphosate exposures. Potential sources of these persistent, non‐seasonal urinary herbicide concentrations should be explored and could include drinking water, fresh produce, commercial dog foods, or persistent outdoor residues. We also demonstrated that both herbicides are genotoxic to canine lymphoid cells at concentrations as low as 0.10 μM and that an estimated 10% of our 60 dogs overall had potentially genotoxic 2,4‐D exposures. Follow‐up studies should include additional measures of in vitro genotoxicity for both 2,4‐D and glyphosate in canine lymphoid cells, assess canine plasma herbicide exposures directly, and explore non‐seasonal sources of highly prevalent 2,4‐D and glyphosate exposures in dogs.
Ethics Statement
All work with animals at the authors' institution was performed under approval IACUC protocol V005450. Banked samples provided by the Golden Retriever Lifetime Study were collected using a protocol approved by the Morris Animal Foundation's Animal Welfare Advisory Board.
Conflicts of Interest
The authors declare no conflicts of interest.
Supporting information
Supplemental Figure S1. Aggregate molar concentrations of the herbicides 2,4‐D and glyphosate in the urine of healthy golden retrievers (n = 30), and those lymphoma immunophenotyped as B‐cell (n = 14) or T‐cell (n = 11), at two time points; T‐1y is 1 year prior to diagnosis or matched time point; T0 is the time of diagnosis or matched time point. p > 0.35 across the three groups at each time point.
Supplemental Figure S2. Genotoxicity of increasing concentrations 2,4‐D towards canine PBMCs in the presence or absence of canine liver microsomes. 2‐h incubation. Error bars represent standard deviations. ** p = 0.002 between presence or absence of canine liver microsomes at a single concentration.
Supplemental Figure S3. Genotoxicity of increasing concentrations of glyphosate towards canine PBMCs in the presence or absence of canine liver microsomes. Error bars represent standard deviations. 2‐h incubation.
Supplemental Table 1. Canine urinary 2,4‐D and glyphosate concentrations by season of sampling, across two time points approximately 1 year apart (T‐1y and T0), in 60 pet golden retriever dogs. Urinary herbicides are reported as medians with observed ranges.
Acknowledgements
We would like to thank Dr. Bevin Engelward and Dr. Joshua Corrigan at the Massachusetts Institute for Technology for assistance and materials for the CometChip assay. We also thank Ms. Amy Elbe for providing blood samples from healthy dogs for the genotoxicity assays and Hannah Peterson MS, for assistance with some in vitro experiments. Finally, we would like to thank Morris Animal Foundation, its staff members and all participants in the Golden Retriever Lifetime Study, including the dog owners, their golden retrievers and the Study veterinarians who made this work possible.
A.T. was supported by TL1TR002375 within the Institute for Clinical and Translational Research at the University of Wisconsin‐Madison, supported by U54TR002373. This research was supported by a Research Forward grant from the Office of the Vice Chancellor for Research and Graduate Education (OVCRGE) at the University of Wisconsin‐Madison.
The Golden Retriever Lifetime Study and the samples used in the current study were made possible through financial support provided by the Morris Family Foundation, the Mark & Bette Morris Family Foundation, VCA, the V Foundation, Blue Buffalo Company, Petco Love, Zoetis, Antech Inc., Elanco, the Purina Institute, Orvis, the Golden Retriever Foundation, the Hadley and Marion Stuart Foundation, Mars Veterinary, generous private donors and the Flint Animal Cancer Center at Colorado State University. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Funding: This work was supported by University of Wisconsin‐Madison, Office of the Vice Chancellor for Research and Graduate Education.
Data Availability Statement
The data that support the findings of this study are available from the corresponding author upon reasonable request.
References
- 1. Zandvliet M., “Canine Lymphoma: A Review,” Veterinary Quarterly 36, no. 2 (2016): 76–104, 10.1080/01652176.2016.1152633. [DOI] [PubMed] [Google Scholar]
- 2. Seelig D. M., Avery A. C., Ehrhart E. J., and Linden M. A., “The Comparative Diagnostic Features of Canine and Human Lymphoma,” Veterinary Sciences 3, no. 2 (2016): 11, 10.3390/vetsci3020011. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Teske E., “Canine Malignant Lymphoma: A Review and Comparison With Human Non‐Hodgkin's Lymphoma,” Veterinary Quarterly 16, no. 4 (1994): 209–219, 10.1080/01652176.1994.9694451. [DOI] [PubMed] [Google Scholar]
- 4. Pastor M., Chalvet‐Monfray K., Marchal T., et al., “Genetic and Environmental Risk Indicators in Canine Non‐Hodgkin's Lymphomas: Breed Associations and Geographic Distribution of 608 Cases Diagnosed Throughout France Over 1 Year,” Journal of Veterinary Internal Medicine 23, no. 2 (2009): 301–310, 10.1111/j.1939-1676.2008.0255.x. [DOI] [PubMed] [Google Scholar]
- 5. Kimura K. C., de Almeida Zanini D., Nishiya A. T., and Dagli M. L., “Domestic Animals as Sentinels for Environmental Carcinogenic Agents,” BMC Proceedings 7, no. S2 (2013): K13, 10.1186/1753-6561-7-s2-k13. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Hayes H. M., Tarone R. E., Cantor K. P., Jessen C. R., McCurnin D. M., and Richardson R. C., “Case‐Control Study of Canine Malignant Lymphoma: Positive Association With Dog owner's Use of 2,4‐Dichlorophenoxyacetic Acid Herbicides,” Journal of the National Cancer Institute 83, no. 17 (1991): 1226–1231. [DOI] [PubMed] [Google Scholar]
- 7. Hayes H. M., Tarone R. E., and Cantor K. P., “On the Association Between Canine Malignant Lymphoma and Opportunity for Exposure to 2,4‐Dichlorophenoxyacetic Acid. Multicenter Study,” Environmental Research 70, no. 2 (1995): 119–125, 10.1006/enrs.1995.1056. [DOI] [PubMed] [Google Scholar]
- 8. Schofield I., Stevens K. B., Pittaway C., et al., “Geographic Distribution and Environmental Risk Factors of Lymphoma in Dogs Under Primary‐Care in the UK,” Journal of Small Animal Practice 60, no. 12 (2019): 746–754, 10.1111/jsap.13075. [DOI] [PubMed] [Google Scholar]
- 9. Weisenburger D. D., “Environmental Epidemiology of Non‐Hodgkin's Lymphoma in Eastern Nebraska,” American Journal of Industrial Medicine 18, no. 3 (1990): 303–305, 10.1002/ajim.4700180310. [DOI] [PubMed] [Google Scholar]
- 10. de Roos A. J., Zahm S. H., Cantor K. P., et al., “Integrative Assessment of Multiple Pesticides as Risk Factors for Non‐Hodgkin's Lymphoma Among Men,” Occupational and Environmental Medicine 60, no. 9 (2003): E11, 10.1136/oem.60.9.e11. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Eriksson M., Hardell L., Carlberg M., and Akerman M., “Pesticide Exposure as Risk Factor for Non‐Hodgkin Lymphoma Including Histopathological Subgroup Analysis,” International Journal of Cancer 123, no. 7 (2008): 1657–1663, 10.1002/ijc.23589. [DOI] [PubMed] [Google Scholar]
- 12. Takashima‐Uebelhoer B. B., Barber L. G., Zagarins S. E., et al., “Household Chemical Exposures and the Risk of Canine Malignant Lymphoma, a Model for Human Non‐Hodgkin's Lymphoma,” Environmental Research 112 (2012): 171–176, 10.1016/j.envres.2011.12.003. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Hayes H. M., Tarone R. E., Cantor K. P., Jessen C. R., McCurnin D. M., and Richardson R. C., “Case‐Control Study of Canine Malignant Lymphoma: Positive Association With Dog owner's Use of 2, 4‐Dichlorophenoxyacetic Acid Herbicides,” JNCI: Journal of the National Cancer Institute 83, no. 17 (1991): 1226–1231, 10.1093/jnci/83.17.1226. [DOI] [PubMed] [Google Scholar]
- 14. Soloneski S., González N. V., Reigosa M. A., and Larramendy M. L., “Herbicide 2,4‐Dichlorophenoxyacetic Acid (2,4‐D)‐induced Cytogenetic Damage in Human Lymphocytes In Vitro in Presence of Erythrocytes,” Cell Biology International 31, no. 11 (2007): 1316–1322, 10.1016/j.cellbi.2007.05.003. [DOI] [PubMed] [Google Scholar]
- 15. Sandal S. and Yilmaz B., “Genotoxic Effects of Chlorpyrifos, Cypermethrin, Endosulfan and 2,4‐D on Human Peripheral Lymphocytes Cultured From Smokers and Nonsmokers,” Environmental Toxicology 26, no. 5 (2011): 433–442, 10.1002/tox.20569. [DOI] [PubMed] [Google Scholar]
- 16. Nagy K., Argaw Tessema R., Szasz I., Smeirat T., Al Rajo A., and Adam B., “Micronucleus Formation Induced by Glyphosate and Glyphosate‐Based Herbicides in Human Peripheral White Blood Cells,” Frontiers in Public Health 9 (2021): 639143, 10.3389/fpubh.2021.639143. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Wozniak E., Sicinska P., Michalowicz J., et al., “The Mechanism of DNA Damage Induced by Roundup 360 PLUS, Glyphosate and AMPA in Human Peripheral Blood Mononuclear Cells ‐ Genotoxic Risk Assessement,” Food and Chemical Toxicology 120 (2018): 510–522, 10.1016/j.fct.2018.07.035. [DOI] [PubMed] [Google Scholar]
- 18. Labadie J., Swafford B., DePena M., Tietje K., Page R., and Patterson‐Kane J., “Cohort Profile: The Golden Retriever Lifetime Study (GRLS),” PLoS One 17, no. 6 (2022): e0269425. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Guy M. K., Page R. L., Jensen W. A., et al., “The Golden Retriever Lifetime Study: Establishing an Observational Cohort Study With Translational Relevance for Human Health,” Philosophical Transactions of the Royal Society, B: Biological Sciences 370, no. 1673 (2015): 20140230, 10.1098/rstb.2014.0230. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. Craun K., Luethcke K. R., Shafer M., et al., “Environmental Chemical Exposures in the Urine of Dogs and People Sharing the Same Households,” Journal of Clinical and Translational Science 5, no. 1 (2020): e54, 10.1017/cts.2020.548. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Anadón A., Martínez‐Larrañaga M. R., Martínez M. A., et al., “Toxicokinetics of Glyphosate and Its Metabolite Aminomethyl Phosphonic Acid in Rats,” Toxicology Letters 190, no. 1 (2009): 91–95, 10.1016/j.toxlet.2009.07.008. [DOI] [PubMed] [Google Scholar]
- 22. Karthikraj R. and Kannan K., “Widespread Occurrence of Glyphosate in Urine From Pet Dogs and Cats in New York State, USA,” Science of the Total Environment 659 (2019): 790–795, 10.1016/j.scitotenv.2018.12.454. [DOI] [PubMed] [Google Scholar]
- 23. Barnekow D. E., Hamburg A. W., Puvanesarajah V., and Guo M., “Metabolism of 2,4‐Dichlorophenoxyacetic Acid in Laying Hens and Lactating Goats,” Journal of Agricultural and Food Chemistry 49, no. 1 (2001): 156–163, 10.1021/jf000119r. [DOI] [PubMed] [Google Scholar]
- 24. Niemann L., Sieke C., Pfeil R., and Solecki R., “A Critical Review of Glyphosate Findings in Human Urine Samples and Comparison With the Exposure of Operators and Consumers,” Journal für Verbraucherschutz und Lebensmittelsicherheit 10, no. 1 (2015): 3–12. [Google Scholar]
- 25. Balderrama‐Carmona A. P., Valenzuela‐Rincón M., Zamora‐Álvarez L. A., et al., “Herbicide Biomonitoring in Agricultural Workers in Valle del Mayo, Sonora Mexico,” Environmental Science and Pollution Research 27, no. 23 (2020): 28480–28489, 10.1007/s11356-019-07087-6. [DOI] [PubMed] [Google Scholar]
- 26. Osborne C. A., “The Ins and Outs of Polyuria and Polydipsia,” dvm360 (2001), https://www.dvm360.com/view/ins‐and‐outs‐polyuria‐and‐polydipsia. [Google Scholar]
- 27. Sauerhoff M. W., Braun W. H., Blau G. E., and Gehring P. J., “The Fate of 2,4‐Dichlorophenoxyacetic Acid (2,4‐D) Following Oral Administration to Man,” Toxicology 8, no. 1 (1977): 3–11, 10.1016/0300-483x(77)90018-x. [DOI] [PubMed] [Google Scholar]
- 28. Kohli J. D., Khanna R. N., Gupta B. N., Dhar M. M., Tandon J. S., and Sircar K. P., “Absorption and Excretion of 2,4‐Dichlorophenoxyacetic Acid in Man,” Xenobiotica 4, no. 2 (1974): 97–100. [DOI] [PubMed] [Google Scholar]
- 29. Peterson H. M., Manley C. I., Trepanier L. A., and Pritchard J. C., “Genotoxicity From Metronidazole Detected In Vitro, but Not In Vivo, in Healthy Dogs in a Randomized Clinical Trial,” American Journal of Veterinary Research 84, no. 1 (2022): 6, 10.2460/ajvr.22.07.0112. [DOI] [PubMed] [Google Scholar]
- 30. Sykora P., Witt K. L., Revanna P., et al., “Next Generation High Throughput DNA Damage Detection Platform for Genotoxic Compound Screening,” Scientific Reports 8, no. 1 (2018): 2771, 10.1038/s41598-018-20995-w. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31. Ge J., Prasongtanakij S., Wood D. K., et al., “CometChip: A High‐Throughput 96‐Well Platform for Measuring DNA Damage in Microarrayed Human Cells,” Journal of Visualized Experiments 92, no. 92 (2014): e50607, 10.3791/50607. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32. Shi J., Xie C., Liu H., et al., “Metabolism and Bioactivation of Fluorochloridone, a Novel Selective Herbicide, In Vivo and In Vitro,” Environmental Science & Technology 50, no. 17 (2016): 9652–9660, 10.1021/acs.est.6b02113. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33. Reynolds P. M., Reif J. S., Ramsdell H. S., and Tessari J. D., “Canine Exposure to Herbicide‐Treated Lawns and Urinary Excretion of 2,4‐Dichlorophenoxyacetic Acid,” Cancer Epidemiology, Biomarkers & Prevention 3, no. 3 (1994): 233–237. [PubMed] [Google Scholar]
- 34. Knapp D. W., Peer W. A., Conteh A., et al., “Detection of Herbicides in the Urine of Pet Dogs Following Home Lawn Chemical Application. Research Support, Non‐U.S. Gov't,” Science of the Total Environment 456–457 (2013): 34–41, 10.1016/j.scitotenv.2013.03.019. [DOI] [PubMed] [Google Scholar]
- 35. van Ravenzwaay B., Hardwick T. D., Needham D., Pethen S., and Lappin G. J., “Comparative Metabolism of 2,4‐Dichlorophenoxyacetic Acid (2,4‐D) in Rat and Dog,” Xenobiotica 33, no. 8 (2003): 805–821, 10.1080/0049825031000135405. [DOI] [PubMed] [Google Scholar]
- 36. Ruple A., Avery A. C., and Morley P. S., “Differences in the Geographic Distribution of Lymphoma Subtypes in Golden Retrievers in the USA,” (1476–5829 (Electronic)). [DOI] [PubMed]
- 37. De Roos A. J., Fritschi L., Ward M. H., et al., “Herbicide Use in Farming and Other Jobs in Relation to Non‐Hodgkin's Lymphoma (NHL) Risk,” Occupational and Environmental Medicine 79, no. 12 (2022): 795–806, 10.1136/oemed-2022-108371. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38. Ali S., Ullah M. I., Sajjad A., Shakeel Q., and Hussain A., Environmental and Health Effects of Pesticide Residues (Springer International Publishing, 2021), 311–336. [Google Scholar]
- 39. Tadeo J. L., Sanchez‐Brunete C., and Wilson I. D., Herbicides | Gas Chromatography (Academic Press, 2000), 2984–2991. [Google Scholar]
- 40. Chen M.‐X., Cao Z.‐Y., Jiang Y., and Zhu Z.‐W., “Direct Determination of Glyphosate and Its Major Metabolite, Aminomethylphosphonic Acid, in Fruits and Vegetables by Mixed‐Mode Hydrophilic Interaction/Weak Anion‐Exchange Liquid Chromatography Coupled With Electrospray Tandem Mass Spectrometry,” Journal of Chromatography A 1272 (2013): 90–99. [DOI] [PubMed] [Google Scholar]
- 41. Zhao J., Pacenka S., Wu J., et al., “Detection of Glyphosate Residues in Companion Animal Feeds,” Environmental Pollution (Barking, Essex: 1987) 243, no. Pt B (2018): 1113–1118, 10.1016/j.envpol.2018.08.100. [DOI] [PubMed] [Google Scholar]
- 42. Kaioumova D. F. and Khabutdinova L., “Cytogenetic Characteristics of Herbicide Production Workers in Ufa,” Chemosphere 37, no. 9–12 (1998): 1755–1759, 10.1016/s0045-6535(98)00240-9. [DOI] [PubMed] [Google Scholar]
- 43. Mesnage R., Brandsma I., Moelijker N., Zhang G., and Antoniou M. N., “Genotoxicity Evaluation of 2,4‐D, Dicamba and Glyphosate Alone or in Combination With Cell Reporter Assays for DNA Damage, Oxidative Stress and Unfolded Protein Response,” Food and Chemical Toxicology 157 (2021): 112601, 10.1016/j.fct.2021.112601. [DOI] [PubMed] [Google Scholar]
- 44. Muruzabal D., Collins A., and Azqueta A., “The Enzyme‐Modified Comet Assay: Past, Present and Future,” Food and Chemical Toxicology 147 (2021): 111865, 10.1016/j.fct.2020.111865. [DOI] [PubMed] [Google Scholar]
- 45. Straube S., Westphal G. A., and Hallier E., “Comment on: Implications of Latency Period Between Benzene Exposure and Development of Leukemia‐a Synopsis of Literature,” Chemico‐Biological Interactions 2 (2010): 248–249, author reply 247. [DOI] [PubMed] [Google Scholar]
- 46. Weisenburger D. D., “A Review and Update With Perspective of Evidence That the Herbicide Glyphosate (Roundup) is a Cause of Non‐Hodgkin Lymphoma,” Clinical Lymphoma, Myeloma & Leukemia 21, no. 9 (2021): 621–630, 10.1016/j.clml.2021.04.009. [DOI] [PubMed] [Google Scholar]
- 47. Bolognesi C., “Genotoxicity of Pesticides: A Review of Human Biomonitoring Studies,” Mutation Research 543, no. 3 (2003): 251–272, 10.1016/s1383-5742(03)00015-2. [DOI] [PubMed] [Google Scholar]
- 48. Lioi M. B., Scarfi M. R., Santoro A., et al., “Cytogenetic Damage and Induction of Pro‐Oxidant State in Human Lymphocytes Exposed In Vitro to Gliphosate, Vinclozolin, Atrazine, and DPX‐E9636,” Environmental and Molecular Mutagenesis 32, no. 1 (1998): 39–46. [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Supplemental Figure S1. Aggregate molar concentrations of the herbicides 2,4‐D and glyphosate in the urine of healthy golden retrievers (n = 30), and those lymphoma immunophenotyped as B‐cell (n = 14) or T‐cell (n = 11), at two time points; T‐1y is 1 year prior to diagnosis or matched time point; T0 is the time of diagnosis or matched time point. p > 0.35 across the three groups at each time point.
Supplemental Figure S2. Genotoxicity of increasing concentrations 2,4‐D towards canine PBMCs in the presence or absence of canine liver microsomes. 2‐h incubation. Error bars represent standard deviations. ** p = 0.002 between presence or absence of canine liver microsomes at a single concentration.
Supplemental Figure S3. Genotoxicity of increasing concentrations of glyphosate towards canine PBMCs in the presence or absence of canine liver microsomes. Error bars represent standard deviations. 2‐h incubation.
Supplemental Table 1. Canine urinary 2,4‐D and glyphosate concentrations by season of sampling, across two time points approximately 1 year apart (T‐1y and T0), in 60 pet golden retriever dogs. Urinary herbicides are reported as medians with observed ranges.
Data Availability Statement
The data that support the findings of this study are available from the corresponding author upon reasonable request.