Histological Response of the Bovine Mammary Gland to lntramammary Devices’ S. C. NICKERSON, P. J. WASHBURN, and N. T. BODDIE Mastitis Research Laboratory Hill Farm Research Station
Louisiana Agricultural Experiment Station Louisiana State University Agricultural Center Route 1, Box 10 Homer 71040 ABSTRACT
Histologic response of the bovine mammary gland to presence of three intramammary device models (abraded, star, or grooved) was studied in 12 lactating cows. Uninfected quarters fitted with devices exhibited greater leukocyte infiltration into teat and gland cistern linings as well as into mammary parenchyma adjacent to the gland cistern compared with unfitted control quarters. Cytologic alterations to cistern linings in device-fitted quarters ranged from degeneration and sloughing of surface cells of the double-layered epithelium to hyperplasia, stratification, and keratin formation. In uninfected quarters, quantification of mammary parenchymal components showed no differences among models for percentages of epithelium, but percentage of lumen was lowest and stroma highest for the star intramammary device model, suggesting reduced secretory activity or mammary involution in response to these intramammary devices. Presence of bacterial infection amplified the histologic responses to all devices. Leukocyte infiitration remained greater in device-fitted quarters compared with controls but was elevated over uninfected quarters for all treatments. Likewise, hyperplasia, stratification, and keratin formation of the cistern epithelial
lining were more frequently observed in infected quarters. Among models in infected quarters, percentage of lumen was lowest and stroma highest in quarters fitted with abraded devices. In most cases, presence of infection masked any effect of devices on mammary parenchymal components. Plaque formation was observed on all models and tended to be thicker on those rehieved from infected quarters. Electron microscopic examination showed that plaque was composed of leukocytes, cell debris, keratin, and amorphous material. Results demonstrated that most intramammary device models were successful in stimulating leukocytosis into the gland, and tissues from fitted quarters displayed alterations to cisternal linings; however, quarters fitted with these devices exhib ited r e d u d secretory activity. (Key words: bovine, histological response, intramammary device) Abbreviation key: CNS = coagulase-negative staphylococci, ICB = intracisternal bead device, IMD = intramammary device, IMI = intramammary infection. INTRODUCTION
Placement of intramammary devices (IMD) into the gland cistern of the bovine mammary gland has been shown to elevate the neutrophil concentration in foremilk and stripping fractions, thereby increasing resistance to infection with mastitis pathogens (3, 8, 16, 17). The exact mode by which the IMD stimulates leuReceived Mar& 6. 1991. kocytosis is unknown, although the device is Accepted May 10, 1991. thought to cause a mild, localized irritation to lApproved by the director of the Louisiana AgricuItual Experiment Station as h4anuscript Number 90.80- the gland cistern epithelial lining, resulting in inflammation (8). It is thought that this minor 4244. 1991 J Dairy Sci 74:338+3395
3383
3384
NICKERSON E" AL.
tissue damage generates chemoattractants, resulting in an influx of neutrophils that adhere to IMD surfaces. Subsequent degranulation of these leukocytes releases proteolytic enzymes, leading to deposition of plaque, which may be chemotactic and stimulate further leukocytosis (5, 18). Although the inflammatory response to the IMD may decrease susceptibility to intramammary infection (IMI),it may also result in damage to milk-synthesizing tissues of the udder. For example, placement of abraded JMD in mammary quarters for 6 mo resulted in pathological changes to udder tissues, including hyperplasia and metaplasia of the epithelial cells and hypertrophy of the subepithelial connective tissue; pathological changes wert more marked in infected versus uninfected IMD quarters (5). Similarly, placement of smwth IMD for 3 yr resulted in epithelial hyperplasia, squamous metaplasia, and polyp formation (1). In addition, use of abraded IMD sometimes led to bleeding into milk (2, 7, 19). More recently, novel high density IMD models have been developed to reduce the amount of irritation to internal tissues of the udder, reduce bleeding and plaque formation, and keep devices near the floor of the gland cistern and away fiom secretary tissue (11). The purpose of this investigation was to determine whether the new device models were effective in stimulating lea,~I~ocytosis into the linings of the teat and gland cisterns without affecting integrity of mammary seaetory tissue. MATERIALS AND METHODS
Twelve lactating Jersey cows brom the Hill
Farm Research Station dairy herd were used. Prior to placement, days into lactation ranged and lactation from 120 to 428 (mean = M), age ranged from 1 to 7. At least two quarters of each cow were fitted with the same IMD model (one IMD per quarter); unfitted quarters served as controls. Bacteriologic status of quarters was determined prior to placement using standard procedures (4). The test devices, composed of fluorinated ethylene propylene plastic, were 2 mm in diameter and 160 mm in length, forming a loop 28 mm in diameter (11). The grooved and star models were designed with d a c e structunes to proJournal of Dairy Science Vol. 74, No. 10, 1991
mote leukocytosis without gross tissue irritation, yet with appropriate densities that would allow gravitation to the floor of the gland cistern and limit leukocyte response close to the point of bacterial entry. The grooved model had 55' of its outer circumference shaped into a shallow round-bottomed groove. The groove was designed to exert pressure on the cisternal epithelium, serving as an irritant to promote local leukocytosis. The star model had five angular ridges along its length to increase the amount of pressure on the cisternal epithelium and hawse leukocytosis. The abraded model was composed of polyethylene and had its surfaces roughened with emery board to maximize the amount of initation to the gland cistern lining. The surface of the abraded model was previously found to promote plaque formation and bleeding into milk. Thus, it was hypothesized that the grooved and star models would elicit a less pathogenic response in udder tissues than found previously with abraded M D (5). Time in place before slaughter ranged from 5 wk to 19 mo. Approximately 5 h after the last milking, cows were slaughtered and mammary secretions collected to determine infection status. Intramammary devices were also retrieved and processed for scanning electron microscopy as described by Nemanic and Pitelka (9). Plaque removed from the devices was fixed and processed for transmission electron microscopy as by Nickerson et al. (14). Teat and gland cistern tissue samples and mammary parenchyma from each fitted and control quarter were taken and processed for paraffin embedding and light microscopic examination as by Nickerson et al. (10). Parenchymal tissue samples were taken from two zones in each quarter. Zone 1 tissue (deep parenchyma) was taken from the dorso-lateral portion of the gland at a depth of 4 cm.Zone 2 tissue (cisternal parenchyma) was immediately adjacent to the gland cistern. In the two cases in which IMD were found lodged in large ducts, tissue in contact with devices was excised and processed as just described. Sections were taken on an A 0 rotary microtome and stained with hematoxylin and eosin. Histologic analysis of the teat cistern lining (two sections per quarter) and gland cistern lining (four sections per quarter) was performed to evaluate the leukocyte response as well as the reaction of
'ITSSUE RESPONSE To INTRAMAMMARY DEVICES
the epithelium to the devices. D e p of leukocytosis (concentration of neutrophils) in teat and gland cistern linings and into the secretory parenchyma immediately adjacent to the gland cistern was expressed as average score for infiltrating cells per 10 microscope fields at 250 x where 1 = very few or no neutrophils observed (e10 cells); 2 = moderate neutrophil infiltration (10 to 300 cells); and 3 = marked neutrophil infiltration (>300 cells). Ten fields were examined per slide, and four slides were examined for each of the two quarter zones for each IMD and control quarter. The cytologic reaction of the epithelium for each tissue section was classif'ied as 1 = normal tissue architecture; 2 = sloughing of the epithelial lining; 3 = epithelial hyperplasia or pseudostratification; and 4 = stratification with keratin formation. To estimate secretory activity, mammary parenchyma was analyzed by quantifying percentages of alveolar epithelium, alveolar lumen, and interalveolar connective tissue stroma using a modification of a proce dure described previously (12). Briefly, a reference grid in the microscope ocular was superimposed over the tissue, providing nine contact points. Each point was classified as epithelium, lumen, or stroma at 250 x for each microscope field and converted to a percentage. Ten fields were examined per slide, and four slides were examined for each of the two quarter zones for each IMD and control quarter. The surfaces of the various IMD were observed for plaque formation using an AMR 1200 scanning electron microscope (Advanced Metals Research Corporation, Bedford, MA) operating at 20 kV. Plaque was scraped from some devices and embedded in plastic. Ultrathin sections (60to 90 nm) of plastic embedded plaque were taken on a MT-5O00 ultramicrotome, stained with uranyl acetate and lead citrate, and examined using a Philips EM 300 electron microscope (N. V. Philips Gloeilampenfabrieken, Eindhoven, Neth.) o p erating at 60 kV. Least squares analysis of variance was used to determine significance of the various treatments (abraded, star, and grooved IMD-fittted; control quarters) on the occurrence of the variables: degree of leukocytosis into teat cistern, gland cistern, and parenchymal areas; cytologic alterations to the teat and gland cistern
3385
epithelial linings; and percentages of alveolar epithelium, alveolar lumen, and interalveolar stroma composing the mammary parenchyma. Duncan's multiple range test was used to determine differences amcmg treatment means. In addition, data from all quarters fitted with IMD were combined and compared with data from control quarters using the same analyses. RESULTS Bacterloioglc Status of Control and IMPFltted Quarters
At the time devices were placed, 14 quarters of the 12 cows were infected. Quarters fitted with the abraded model were uninfected, but coagulase-negative staphylococci (CNS) were isolated from three quarters fitted with the star model and three quarters fitted with the g~oovedmodel, and a colifom was isolated from one quarter fitted with the grooved model. Staphylococcur auteur (two quarters), Streprococcus dysgulactiae (one quarter), Streptococcus agalactiae (one quarter), Nocurdiu spp. (two quarters), and CNS (one quarter) were isolated from control quarters; remaining quarters were uninfected. Overall, 26% of IMD-fitted quarters and 41% of control quarters were infected at time of placement. At time of slaughter, when devices were removed, 13 quarters were infected, 9 of the original 14 infections persisted, 1 was cured spontaneously, and 4 new infections occurred. a-the new infections, Streptococcus uberis was isolated from 2 quarters fitted with the star model, CNS was isolated from a grooved IMD-fitted quarter, and Staph. aureus was isolated from a quarter fitted with the abraded model; no new infections occurred in control quarters. overall, 30% of IMD-fitted quarters and 29% of controls were infected at the end of the study. Gross and Mlcroscopic Histologic Observations
Gross observation of teat and gland cistern linings from IMD-fitted quarters at slaughter revealed no differences from those of control teats, even in those quarters where devices were apparently lodged in large ducts prior to removal. All IMD models were intact upon Journal of Dairy Science Vol. 74, No. 10, 1991
3386
NICKERSON ET AI.,.
retrieval and exhibited varying degrees of plaque formation. Light microscopic examination of teat and gland cistern linings of unfitted control quarters demonstrated an intact doublelayered epithelium resting on subepithelial connective tissue containing capillaries and lymphoid cells (Figure 1). Lymphocytes were present in or adjacent to the epithelial lining of all quarters, but neutrophils were usually observed only in IMD-fitted or in infected quarters. Light microscopic examination of teat and gland cistern linings of JMD-fitted quarters showed that some areas were identical to controls; however, in many cases, tissue alterations were observed where contact between the devices and the cisternal lining probably took place. Alterations ranged from degeneration and sloughing of the luminal epithelial cell layer, leaving a single layer intact Figwe 2), and pseudostratification and hyperplasia of the epithelial lining (Figure 3). to stratification and keratinization of the epithelium (Figure 4). In areas exhibiting stratified squamous epithelium, cells in various stages of mitosis were commonly seen (Figure 5). Neutrophils were occasionally observed in alveolar lumina of the mammary parenchyma adjacent to the gland cistern as well as adjacent to and within the gland and teat cistern linings. These leukocytes were only rarely observed breaching the epithelium where they were found squeezing between luminal cells of the epithelial lining and entering cisternal areas Figure 6). Eosinophils were observed in the subepithelial stroma of the gland cistern in two quarters fitted with abraded IMD.
Microscopic examination of areas excised fmm sites where IMD were found lodged in large ducts at time of necropsy revealed marked pathological changes. Most of the epithelial lining was stratified as in Figure 4, and the subepithelial connective tissue was thickened due to fibrous proliferation and leukocyte infiltration. Leukocytes were also observed traversing the stratified squamous epithelium, and they formed clots in the lumen of some ducts with material thought to be fibrin. Examination of mammary parenchymal areas adjacent to the gland cistern in uninfected control quarters for leukocyte infiltration revealed typical secretory alveoli with few or no neutrophils as shown in Figure 7. Such tissues exhibited a large alveolar luminal area and a very limited stromal area, indicating active secretion. Similar areas in tissue sections from quarters fitted with IMD showed a moderate leukocyte infiltration into alveolar luminal areas (Figure 8). Percentage tissue area occupied by alveolar lumen was less, and that occupied by interalveolar stroma was greater, than that found in unfitted quarters. A histologic analysis of IMD quarters compared with control quarters, regardless of infection status, is in Table 1. Leukocytosis into the teat and gland cistern linings as well as into the parenchyma adjacent to the gland cistern was greater in abraded and grooved models compared with control quarters and those fitted with the star model. Significant cytologic alterations to the gland cistern linings were observed in all IMD-fitted quarters compared with controls. Greatest alterations were found in quarters fitted with the abraded
Figure 1. The epithelial lining and subepithelial conneCtive tissue typical of those found in the teat and gland cistern linings from norma& minfectcd mammary glands without. . J' devices. The double layered epithelium is composed of a t x d , cuboidal cell layer underlying a luminal colarrmpr cell layer. Tbe subepithelial connective tissue has few infiltrating lenlrocytes; lympbcytes are present at arrows. ~ 4 7 5 . Figure 2. "be cpihIiaI lining Qpkd of that found in isolated arcaq of the gland cistern of guarters fitted with inr~~manrmadevices classified as degenerating and sloughed The d a c e columnar cells (D) have become flattenad and are degamate however, the basal epithelial layer (B) remains intact. Infilfra?ing neumphils (N) are observed at arrows. x475. P i 3. The epithelial lining and subepithelial connective tissue typical of those found in some gland cisternal areas of quarters fitted with inhamammm devices classified as pseudostratified or hypaplastic. The epithelium appears multilaymd, and some layers pmbude into the cistern lumen (L.).InNtrating mtrophils are observed in the epithelium at arrows. x475. Pigum 4. Epithelial lining typical of that found in some areas of quarters fitted with the abraded intramammary devices classified as stratified sqnemous epithelium. Neutrophils are present in the subepithelial connective t i w e (arrows), and the IenLocytcs have formed a clot with fibrin m the cistern lumen. ~200. J o d of Dairy Science Vol. 74, No. 10, 1991
TISSUE RESPONSE TO INTRAMAMMARY DEVICES
3387
J o d of Dairy Science Vol. 74. No. 10, 1991
3388
NICIEWSON ET AL.
model. Analysis of mammary parenchymal components demonstrated that, although percentages of alveolar epithelium (both m e s combined) were similar in IMD-fitted and control quarters, IMD quarters (all models combined) exhibited significantly less alveolar luminal area and significantly more interalveolar stroma than controls (compare Figures 7 and 8). This was mainly attributed to quarters fitted with the star model. Observations were similar for Zone 1 tissues. Results of the analysis of Zone 2 tissues followed a similar trend; however, quarters fitted with the abraded model exhibited significantly less alveolar lumen and significantly more stroma than controls. In uninfected quarters, degree of leukocytosis as well as alterations to the epithelial linings were significantly greater in IMD quarters (all models combined) (Table 2). Among models, leukocytosis was lowest in quarters fitted with the star model in both teat and gland cistern areas. Cytologic alterations to the gland cistern lining were greatest in abraded IMD-fitted quarters. The IMD-fitted quarters (all models combined) exhibited less luminal area (P < .05) and more stroma (P < .05) in Zone 1 and Zone 2 tissues than tissue from unfitted controls. Among models, quarters fitted with the star model exhibited the lowest luminal areas and greatest stromal area overall (both zones combined) and for Zone 1 parenchyma. In infected quarters, presence of inflammation generally obscured any effect of the IMD. In most cases, degree of leukocyte inliltration and score for cytologic alterations were elevated over values of uninfected quarters (Table 3). Degree of leukocytosis into the teat cistern lining and cytologic alterations to the gland cistern lining were higher (P < .05) in abraded quarters compared with other quarters.
Leukocytosis into the gland cistern parenchyma was greater in quarters fitted with the grooved model compared with controls. Percentage lumen was lowest and stroma highest in abraded quarters compared with all other treatments; this was the case overall (both zones combined) and for Zone 1 parenchyma. Percentage epithelium in Zone 1 parenchyma of abraded IMD quarters was significantly lower than all other treatments. Data analysis demonstrated no effect of stage of lactation on the histologic differences observed between control and IMD quarters (data not shown). Overall, degree of leukocytosis and alterations to the cisternal linings were not different between cows 4 8 0 d in milk and those >305 d in m i k However, analysis of parenchymal components showed that cows >305 d in milk exhibited less alveolar luminal area and more interalveolar stroma than cows 4 8 0 d in milk. This would be expected because tissue from cows >305 d in milk would be more involuted than tissue from cows earlier in lactation. Likewise, length of time that IMD were in place and parity had little effect on differences observed among treatments. Tissues from quarters exposed to IMD R mo exhibited slightly greater leukocytosis and alterations to the cisternal linings than those from quarters exposed >5 mo. Tissues from quarters exposed to IMD 6 mo also exhibited less lumen and more stroma; however, three of the four cows in this group were >305 d in milk, and tissue was probably undergoing involution. Electron microscope examination of the devices showed that plaque was present to some degree on all models but tended to be thicker on devices retrieved from infected quarters (Figures 9 and 10). Plaque was composed mainly of intact leukocytes but also
Figure 5. Stratified epithelium from a quarter fitted with an abraded inlmtmammmy device model; note neutrophils (N) and mitotic figures
0. ~750.
Figure 6. Portion of gland cistern lming of quarter fitted with an abraded ni?-y device model showing neutrophils (arrows) in various stages of migration across the epithelial layer. x750. Figure 7. Parenchymaltissue typical of that found in Zone 1 and Zone 2 of most uninfected int~amatnmarydevice fitted quarters c l a s s i i i as n o d . Such tissues exhibited a large alveolar luminal area Q and a very l i i t e d stromal (S) area, indicating copious secretory activiv. ~ 2 0 0 . Figure 8. Parenchymal tissue found in Zone 2 of a quarter fitted with an abraded intramammary device model. Percentage tissue area occupied by ahrtolar lumen is less, and that occupied by interalveolar stroma is more, than that in Figure 7. Note several leukocytes present in Lhe alveolar lamioa (arrows). x200.
J o d of Dairy Science Vol. 74, No. 10. 1991
TISSUE RESPONSE To
INTRAMAMMARY DEVICES
3389
I
Journal of Dairy Science Vol. 74, No. 10, 1991
3390
NICKERSON ET AL..
contained cell debris, keratin, and amorphous infection in abraded IMD quarters was threematerial (Figure 11). Where plaque was mini- fold that of controls (37 vs. 12%). At time of mal or absent, leukocytes were often observed dissection in the present study, no gross tissue attached to the clean plastic surfaces of some effects of the devices were observed, even in IMD models (Figure 12). two quarters in which IMD were lodged in large ducts and had to be removed with forDISCUSSION ceps. This is in contrast to findings of Collins Data on the microbiologic status of IMD- et al. (5), who observed petichiae dong the fitted and control qumers at time of slaughter walls of the gland cistern and larger ducts of showed that prevalence of natural infection some IMD-fitted quarters. Histological responses to IMD were similar was similar in IMD and control quarters (30 to those reported earlier (5); however, pathovs. 29%). New infections oc& more frelogical changes were not as marked. In the quently in IMD quarters than in controls (6 vs. 0), but, overall, the total number was small. present study, thickening of the subepithelial Similarly, Corlett et al. (6)found no difference connective tissue of the gland cistern lining in prevalence of natural infection among con- and fibroblast proliferation were not observed trol and smooth IMD quarters in three com- except at IMD sites in large ducts of infected mercial herds. In contrast, Collins et al. (5) quarters, although leukocyte infiltration and found that, 6 mo after placement, prevalence of epithelial hyperplasia and keratinization were
TABLE 1. Effect of -ni devices on tissue characteristics of the teat and gland cistern linings aad mammary parenchyma regardless of quarter infection stam. IntTamammary
Tissue Compo~nt
Abraded (n = 6)
device model
(n = 13)
orooved (n = 8)
Star
(n = 26)
Control (n = 19)
All
Teat cistern lining Leukocytosis1
2.17'
AlteratioosZ
1.oob
1.46b 1so.
1.81* 1.46"
1.69 1.43*
lSlb 1.12"
2.67' 3.33.
1.69b 1.%b
2.27" 2.15b
2.05* 2.19,
1.2gb 1.m
2.w
1.3Sb
223'
1.81*
1.2gb
Gland cistern lining LeuLocytosis'
Alterations2 Leukocytosis into gland c i s t a n parenchymal ~ammaryparenchyma3 (both zones combimd) Epithelium
Lumen Slroma
Zone 1 parenchyma3 Epithelium Lumen stroma Zone 2 parenchyma3 Epithelium Lumen
Strana
3 1.07" 28.24' 40.69*'
30.w 21.& 47.42.
30.26" 30.93' 38.81b
30.51 26.66* 42.83'
3 1.36' 32.66' 35.98b
28.98' 40.83' 30.19b
30.96' 20.73b 48.31'
31.01' 33.38' 35.61b
30.75 28.72* 40.53+
31.%' 34.67' 33.37b
33.15' 15.65b 5120"
30.25'b 23.27* 46.48h
29.ab 28.43' 42.09*
30.35 24.53* 45.12*
30.78ab 30.6ga 38.54b
4b*cMcans without a superscript in common differ (P < .OS). 'Data expressed as average score for infiltrating cells where 1 = very few or no neutrophils observed; 2 = moderate neutrophil infiltration;and 3 = marked neutrophil infitration. 2Data expressed as average score where 1 = OOmLal tissue architecture; 2 = sloughing of the epithelial lining;3 = epithelial hyperplasia or pseudostraMicatioq 4 = stratiftcation with keratin formation. 3Data expressed as percentage of each pareochymal component. *DifZercd from control (P < .OS). Journal of Dairy Science Vol. 74, No. 10, 1991
3391
TISSUE RESPONSE TO INTRAMAMMARY DEVICES
TABLE 2. Effect of intramammary devices on tissue characteristics of the teal and gland cistern linings and mammary parenchyma of uainfecte4i quarters. Intramammary device model
Tissue component Teat cistern lining Leukocytosis' Alterations2 Gland cistern lining LeukOCytOSiSl
Alterations' Leukocytosis into gland cistern parenchymal Mammary parenchyma3 (both zones combined) Epithelium Lumen Stroma Zone 1 parenchyma3 Epithelium Lumen stroma Zone 2 parenchyma3 Epithelium Lumen Strollla
(n = 8)
Grooved (n = 7)
All (n = 19)
Control (n = 12)
1.75' 1 .oob
1.06b 1.13b
1.61' 1.56
1.39* 1.29*
1 .wb 1 .oob
2.56 3 .W
1.44b 2.00b
2.11' 2.06b
1.87* 2.13*
1 1.04=
2.W
1. o b
2.1 1'
1.63*
1.08b
32.298 37.29' 30.4Zb
30.82' 18Nb 50.24'
30.34' 32.87' 36.79b
30.76 28.23. 41.01*
30.47a 36.51' 33.d
32.64' 56.67' l0.6p
31.50.
15.5@ 53.00'
30.Ma 34.03b 35.31b
31.23 29.91* 38.86*
31 .25' 38.21b
30.108 22.5ObC 4 7 . e
30.W 3 1.68& 38.32bc
30.27 26.51* 43.22*
29.72' 34.86' 35.42'
Abraded (n = 4)
3 1.94' 17.92' 50.14'
Star
.w
30.54b
4b*cMeans without a superscript in common differ (P c .05). 'Data expressed as average score for infiltrating cells where 1 = very few or no neutrophils observed; 2 = moderate neutrophil infiltration; and 3 = marked neultrophjl infiltration. 2Data expressed as average score where 1 = normal tissue architecture; 2 = sloughing of the epithelial liniOg; 3 = epithelial hyperplasia or pseudostratifkation; 4 = stratification with keratin formation. b a t a apressed as percentage of each parenchymal component. *Differed from control (P < .05).
noted. Histological examination of tissues from cows fitted with IMD over a 3-yr period revealed epithelial hyperplasia, squamous metaplasia, and polyp formation (1). Leukocytosis as well as alterations to the epithelial linings were greater in infected IMD quarters than uninfected IMD quarters as previously demonstrated (5). Eosinophilic infiltration into the subepithelial tissues of the gland cistern was commonly observed in two quarters fitted with abraded devices as observed by Collins et al. (5). The areas exhibiting sloughing and pseudostratification were very similar to those observed in teat cisterns of nonlactating quarters fitted with intracisternal bead devices (ICB) (13). Similarly, neutrophil infitration into the cisternal linings of IMD quarters was identical to that observed in ICB-fitted quarters, and, as
observed with the ICB, IMD-fitted quarters exhibited greater leukocytosis into the mammary parenchyma adjacent to the gland cistern. In contrast to the IMD study, use of ICB did not result in reductions in luminal area and increases in stroma, although the latter study was performed in nonlactating cows. Huston and Heald (8) studied the morphologic features of the gland cistern linings of six firstcalf heifers 3 wk after placement of smooth IMD. They found that IMD-fitted quarters exhibited a significant change toward a single layer of epithelial cells compared with a double layer in control quarters. This phenomenon was also observed in quarters fitted with IMD in the present study, although hyperplasia was also noted. It is possible that the single-layered epithelium observed by Huston and Heald (8) represented sloughing of the Journal of Dairy Science Vol. 74, No. 10, 1991
3392
NICKERSON ET AL.
Journal of Dairy Science Vol. 74, No. 10, 1991
3393
TISSUE RESPONSE TO INTRAMAMMARY DEVICES
TABLE 3. Effect of inftrammmq devices on tissue characteristics of the gland and teat cistern linings and mammary parenchyma of infected quarters.
Inf-y Tissue component Teat cistern lining Leukocytosis' Alterations2 Gland cistern lining Leukocytosis' AI terations2 Luekocytosis into gland cistern parenchyma' Mammary parenchyma3 (both zones combined) Epithelium LUmell stroma Zone 1 parenchyma3 Epithelium Lumen Stroma
Zone 2 parenchyma3 Epithelium Lumen stroma
device model
(n = 5)
Grooved (n = 1)
Au (n = 8)
Control (n = 5)
3.w 1 .cob
2.lOb 2.w
2.25b 1 .38ab
2.25 1.70
2.25b 1.3Iab
3.w 4.w
2.10bc 1.90b
2.63ab 2.38b
2 .w 2.3W
1.79 1.18'
2.wb
1
2.5@
2.W
1 .63b
Abraded (n = 2)
Star
.e
28.61' 10.14b 61.25'
30.42' 25.62a 43.%b
30.05' 25.83' 44.12b
30.19 23.85 45 .%
32.36' 28.33' 39.31b
21 .67b 9.17b 69.16'
30.33' 27.Wa 42.67b
31.94" 3 1.67" 36.39b
29.90 26.61 43.49
32.74' 30.7Sa 36.51b
35.56" 11.11a 53.33'
30.55h 23.74. 45.71'
28.15b 20.w 51.85'
30.19 21.33 48.48
3 1 .98ab 25.91' 42.11'
4b+cMcanswithout a superscript in common differ (P < .05). 'Data expressed as average m r e for intiitrating cells where 1 = very few or no neutrophils observed; 2 = moderate neutrophil infiltration; and 3 = marked neutrophil infiltration. 2Data expressed as average score where 1 = normal tissue architecture; 2 = sloughing of the epithelial lining; 3 = epithelial hypezplasia or pseudostratificatiow 4 = stratification with keratin formation. 'Data expressed as percentage of each parenchymal component. *Differed from control (P < .OS).
luminal epithelial layer after 3 wk. Continued presence of the devices may have resulted in a tissue reaction to the sloughing, leading to hyperplasia of the epithelium. In the present study, significant leukocytosis and alterations to the teat cistern epithelium were also noted. At time of slaughter, all but two IMD were found free floating in the gland cistern; two were found lodged in large ducts. However, during lactation, the grooved IMD of one quarter of one cow was typically palpated in the teat cistern prior to milking and had to be forced up into the gland cistern prior to machine milking. It is possible that IMD models in other quarters periodically descended to the teat cistern, eliciting a tissue reaction. The IMD quarters tended to exhibit less alveolar luminal area and more stroma, indicat-
ing reduced secretory potential or greater involution compared with controls. Involution was greatest in infected abraded IMD quarters. Although quarter milk production was not measured in this study, another investigation found that quarters fitted with abraded IMD produced less milk (7), and it was speculated that this was a result of the pathological changes observed (5). Paape and Weinland (19) also found that quarters fitted with abraded IMD tended to produce less milk compared with unfitted controls. Five of the cows in the present study were in late lactation (over 305 d), and this could have contributed to greater involution in IMD quarters. Paape et al. (20) found that cows fitted with abraded IMD had shorter lactations than controls and suggested that an increase in milk SCC in treated quarters could have contributed to an earlier involution of the mammary secretory tissue. Journal of Dairy Science Vol. 74, No. 10. 1991
3394
NICKERSON ET AI.,.
REFERENCES Tissue wne within a quarter had marked effect on percentages of parenchymal tissue 1 Barnurn, D., M.Hazlett, R.Johnston, and J. Costerton. components for some IMD models. In Zone 1 1985. B i o f i i on bovine inl@auymaydevices and their effect on the gland and rmk Page 81 in Roc. of uninfected quarters, those fitted with the 35th Annu. Mtg. Can. SOC. Mimbiol., Dalhousie abraded model exhibited significantly more luUniv., Halifax, NS, Can. men and less stroma than other treatments 2 Bright, S. A., W. D. Schultze, M.J. Paape,and R. R. Peters. 1987. Effect of smooth or abraded intramam(Table 2). In Zone 2 of uninfected quarters, mary devices on the outcome of experimental cbalthose fitted with abraded Ih4D exhibited less lenge infection of bovine ma~mnary qumers with lumen and more stroma than the other treatEscherichia coli. Am. J. Vet. Res. 48:1290. ments. This suggests that the area adjacent to 3Brooks, B. E., and D. A. Barnum. 1982. The use of polyethylene intmmammary device in protection of the gland cistern in infected abraded IMD the lactating bovine udder against experimental Metquarters (Zone 2) was more involuted, whereas tion with Staphylococcus aureus or Streptococcus the deep parenchyma (Zone 1) either became agalacriae. Can. J . Comp. Med. 46267. more secretory to compensate for reduced ac4Brown, R. W., D. A. Bamum, D. E. Jasper, J. S. McDonald, and W. D. Shultze. 1981. Page 16 in tivity of Zone 2 or was actually becoming h%icrobiological procedures for use in the diagnosis of static as milk accumulated in alveolar lumina bovine mastitis. 2nd ed. Natl. Mastitis Counc., Inc., and displaced the interalveolar stromal tissue. Arlington, VA. 5Collins, R. A., G. Oldham, P. G. Francis, and N. It was thought that if IMD were detrimental to Craven. 1989. Pathological changes following implansecretory activity, the mammary parenchymal tation of intr~wnauuwarydevices (MD) and immunozone adjacent to the IMD site (gland cistern, logical mediator release by cells on recovered IMDs. Zone 2) would exhibit less secretory activity Res. Vet. Sci. M253. 6CorleK N. J., R. R. Peters, M. J. Paape, and W. D. than the deep parenchyma (Zone 1). However, Shultze. 1984. Effect of n im-v device on new with the exception of the abraded model, this infection rate, milk yield, and milk somatic cell counts generally was not the case. in Maryland dairy herds. J. Dairy Sci. 672571. The plaque observed on the surface of IMD 7Francis, P. G. 1985. preliminary results of a within cow trial relating to the effects of a polyethylene was very similar to that observed on ICB (14) htramammary device on bovine maslitis. E e l . Milchand on smooth molded polyethylene IMD (15); wirtsch. Fomchungsber. 37473. plaque tended to be thicker on IMD retrieved 8 Huston, G. E..and C. W. Heald. 1983. Effect of the device on milk infection status, yield, from infected quarters as observed previously and somauc cell count and on the morphological (15). Bacteria were rarely observed on the features of the lactiferons sinus of the bovine udder. surface of IMD from infected quarters. Am. J. Vet. Res. M1856. In conclusion, the three IMD models tested 9 Nemanic, M.K.,and D. R Pitelka. 1971. A scanning electron microscopy study of the lactating mamumy led to increases in leukocyte infiltration and to gland. J. Cell. Biol. 48:410. cytologic alterations in the teat and gland cislONickemon, S. C., P.A. Baker, and P. Trinidad. 1989. tern epithelial linings of the mammary gland. Local immunostimulation of the bovine mammary Quarters fitted with IMD appeared more ingland with interleukin-2. J. Dairy Sci. 72:1764. 11 Nickerson, S. C., R. L. Boddie, W. E. Owens, and J. voluted than unfitted controls, exhibiting less L.Watts. 1990. Effects of novel inlmmammq device alveolar luminal area and more interalveolar models on incidence of intmmammary infection after stroma. Involution tended to be greatest in experimental challenge with mastitis pathogens. J. Dairy Sci. 73~2774. infected quarters fitted with IMD.Plaque was 12 Nickerson, S. C., and C. W. Heald. 1981. Histopathopresent on all models and was thicker on IMD logic response of the bovine mammary gland to experretrieved from infected quarters. There were no imentally induced Staphylococcus uureus infection. effects of any of the IMD models tested on the Am. J. Vet. Res. 42:1351. 13Nickerson, S. C., W.J. Tboqson, W. M.Kortum, new IMI rate during the trial. Thus, the inand N. T. Boddie. 1987. Histological response of creased leukocytosis into fitted glands was not bovine mammary tissue to an intracisternal bead protective against mastitis, and the mammary device. J. Dairy Sci. 70687. tissue exhibited a pathologic response to IMD. 14Nickerson, S. C., W. J. Thompson, S. P. Oliver, and
R. M. Akm. 1988. Effects of intracisternal bead
ACKNOWLEDGMENTS
The authors thank W. M. Kortum for providing the IMD and E C. Huff for typing of this manuscript. Journal of Dairy Science Vol. 74, No. 10. 1991
devices on lacteal secretion components,plaque formation, and bacterial infection during the nonlactating period. Am. J. Vet. Res. 491205. 15 Paape,M.J., N. J. Corlett, and W. M.Kortum. 1987. Examination of intmmmmary devices from infected and uninfected naammary quarters by scanning electron microscopy. J. Dairy Sci. 701045.
TISSUE RESPONSE TO INTRAMAMMARY DEVICES 16 Paape, M.J., W. D. Schultze, N. J. Corlett, and 8 . T. Weinland. 1988. Wect of abraded ni1-y device on outcome in lactating cows after challenge exposure with Streptococcus ubcris. Am. J. Vet. Res. 49:790. 17Paape, M. J., W. D. Schultze, A. J. Guidry, W. M. Kortum, and B. T. Weinland. 1981. Effect of an intramammary polyethylene device on the concentration of leukocytes and immunoglobulins in milk and on the leukocyte response to Escherichia coli endotoxin and challenge exposure with Smphylococcus uureus. Am. J. Vel. Res. 42374. 18 Paape, M.J.. W. D. Schultze. R. R Peters, and N. J.
3395
Corlett. 1984. Effects of inlramauummy devices on milk somatic cells, milk yield and new infection rate. Page 148 in Proc. 23rd Annu. Mtg., Natl. Mastitis Couoc.. Inc., Arlington, VA. 19Wape, M I., and B. T. Weinland. 1988. Effect of abraded intramammary device on milk yield, tissue damage, and cellular composition. J. Dairy Sci. 71: 250. ZOPaape, M. J., G. Ziv, R. H. Miller, and W. D. Schultze. 1986. Update on the use of intmumrmmy devices in the control of mastitis. Page 87 in Proc. 25th Annu. M@. Natl. Mastitis COU~C.. hc., Arlington. VA.
Journal of Dairy Science Vol. 74, No. 10. 1991
Louisiana Agricultural Experiment Station Louisiana State University Agricultural Center Route 1, Box 10 Homer 71040 ABSTRACT
Histologic response of the bovine mammary gland to presence of three intramammary device models (abraded, star, or grooved) was studied in 12 lactating cows. Uninfected quarters fitted with devices exhibited greater leukocyte infiltration into teat and gland cistern linings as well as into mammary parenchyma adjacent to the gland cistern compared with unfitted control quarters. Cytologic alterations to cistern linings in device-fitted quarters ranged from degeneration and sloughing of surface cells of the double-layered epithelium to hyperplasia, stratification, and keratin formation. In uninfected quarters, quantification of mammary parenchymal components showed no differences among models for percentages of epithelium, but percentage of lumen was lowest and stroma highest for the star intramammary device model, suggesting reduced secretory activity or mammary involution in response to these intramammary devices. Presence of bacterial infection amplified the histologic responses to all devices. Leukocyte infiitration remained greater in device-fitted quarters compared with controls but was elevated over uninfected quarters for all treatments. Likewise, hyperplasia, stratification, and keratin formation of the cistern epithelial
lining were more frequently observed in infected quarters. Among models in infected quarters, percentage of lumen was lowest and stroma highest in quarters fitted with abraded devices. In most cases, presence of infection masked any effect of devices on mammary parenchymal components. Plaque formation was observed on all models and tended to be thicker on those rehieved from infected quarters. Electron microscopic examination showed that plaque was composed of leukocytes, cell debris, keratin, and amorphous material. Results demonstrated that most intramammary device models were successful in stimulating leukocytosis into the gland, and tissues from fitted quarters displayed alterations to cisternal linings; however, quarters fitted with these devices exhib ited r e d u d secretory activity. (Key words: bovine, histological response, intramammary device) Abbreviation key: CNS = coagulase-negative staphylococci, ICB = intracisternal bead device, IMD = intramammary device, IMI = intramammary infection. INTRODUCTION
Placement of intramammary devices (IMD) into the gland cistern of the bovine mammary gland has been shown to elevate the neutrophil concentration in foremilk and stripping fractions, thereby increasing resistance to infection with mastitis pathogens (3, 8, 16, 17). The exact mode by which the IMD stimulates leuReceived Mar& 6. 1991. kocytosis is unknown, although the device is Accepted May 10, 1991. thought to cause a mild, localized irritation to lApproved by the director of the Louisiana AgricuItual Experiment Station as h4anuscript Number 90.80- the gland cistern epithelial lining, resulting in inflammation (8). It is thought that this minor 4244. 1991 J Dairy Sci 74:338+3395
3383
3384
NICKERSON E" AL.
tissue damage generates chemoattractants, resulting in an influx of neutrophils that adhere to IMD surfaces. Subsequent degranulation of these leukocytes releases proteolytic enzymes, leading to deposition of plaque, which may be chemotactic and stimulate further leukocytosis (5, 18). Although the inflammatory response to the IMD may decrease susceptibility to intramammary infection (IMI),it may also result in damage to milk-synthesizing tissues of the udder. For example, placement of abraded JMD in mammary quarters for 6 mo resulted in pathological changes to udder tissues, including hyperplasia and metaplasia of the epithelial cells and hypertrophy of the subepithelial connective tissue; pathological changes wert more marked in infected versus uninfected IMD quarters (5). Similarly, placement of smwth IMD for 3 yr resulted in epithelial hyperplasia, squamous metaplasia, and polyp formation (1). In addition, use of abraded IMD sometimes led to bleeding into milk (2, 7, 19). More recently, novel high density IMD models have been developed to reduce the amount of irritation to internal tissues of the udder, reduce bleeding and plaque formation, and keep devices near the floor of the gland cistern and away fiom secretary tissue (11). The purpose of this investigation was to determine whether the new device models were effective in stimulating lea,~I~ocytosis into the linings of the teat and gland cisterns without affecting integrity of mammary seaetory tissue. MATERIALS AND METHODS
Twelve lactating Jersey cows brom the Hill
Farm Research Station dairy herd were used. Prior to placement, days into lactation ranged and lactation from 120 to 428 (mean = M), age ranged from 1 to 7. At least two quarters of each cow were fitted with the same IMD model (one IMD per quarter); unfitted quarters served as controls. Bacteriologic status of quarters was determined prior to placement using standard procedures (4). The test devices, composed of fluorinated ethylene propylene plastic, were 2 mm in diameter and 160 mm in length, forming a loop 28 mm in diameter (11). The grooved and star models were designed with d a c e structunes to proJournal of Dairy Science Vol. 74, No. 10, 1991
mote leukocytosis without gross tissue irritation, yet with appropriate densities that would allow gravitation to the floor of the gland cistern and limit leukocyte response close to the point of bacterial entry. The grooved model had 55' of its outer circumference shaped into a shallow round-bottomed groove. The groove was designed to exert pressure on the cisternal epithelium, serving as an irritant to promote local leukocytosis. The star model had five angular ridges along its length to increase the amount of pressure on the cisternal epithelium and hawse leukocytosis. The abraded model was composed of polyethylene and had its surfaces roughened with emery board to maximize the amount of initation to the gland cistern lining. The surface of the abraded model was previously found to promote plaque formation and bleeding into milk. Thus, it was hypothesized that the grooved and star models would elicit a less pathogenic response in udder tissues than found previously with abraded M D (5). Time in place before slaughter ranged from 5 wk to 19 mo. Approximately 5 h after the last milking, cows were slaughtered and mammary secretions collected to determine infection status. Intramammary devices were also retrieved and processed for scanning electron microscopy as described by Nemanic and Pitelka (9). Plaque removed from the devices was fixed and processed for transmission electron microscopy as by Nickerson et al. (14). Teat and gland cistern tissue samples and mammary parenchyma from each fitted and control quarter were taken and processed for paraffin embedding and light microscopic examination as by Nickerson et al. (10). Parenchymal tissue samples were taken from two zones in each quarter. Zone 1 tissue (deep parenchyma) was taken from the dorso-lateral portion of the gland at a depth of 4 cm.Zone 2 tissue (cisternal parenchyma) was immediately adjacent to the gland cistern. In the two cases in which IMD were found lodged in large ducts, tissue in contact with devices was excised and processed as just described. Sections were taken on an A 0 rotary microtome and stained with hematoxylin and eosin. Histologic analysis of the teat cistern lining (two sections per quarter) and gland cistern lining (four sections per quarter) was performed to evaluate the leukocyte response as well as the reaction of
'ITSSUE RESPONSE To INTRAMAMMARY DEVICES
the epithelium to the devices. D e p of leukocytosis (concentration of neutrophils) in teat and gland cistern linings and into the secretory parenchyma immediately adjacent to the gland cistern was expressed as average score for infiltrating cells per 10 microscope fields at 250 x where 1 = very few or no neutrophils observed (e10 cells); 2 = moderate neutrophil infiltration (10 to 300 cells); and 3 = marked neutrophil infiltration (>300 cells). Ten fields were examined per slide, and four slides were examined for each of the two quarter zones for each IMD and control quarter. The cytologic reaction of the epithelium for each tissue section was classif'ied as 1 = normal tissue architecture; 2 = sloughing of the epithelial lining; 3 = epithelial hyperplasia or pseudostratification; and 4 = stratification with keratin formation. To estimate secretory activity, mammary parenchyma was analyzed by quantifying percentages of alveolar epithelium, alveolar lumen, and interalveolar connective tissue stroma using a modification of a proce dure described previously (12). Briefly, a reference grid in the microscope ocular was superimposed over the tissue, providing nine contact points. Each point was classified as epithelium, lumen, or stroma at 250 x for each microscope field and converted to a percentage. Ten fields were examined per slide, and four slides were examined for each of the two quarter zones for each IMD and control quarter. The surfaces of the various IMD were observed for plaque formation using an AMR 1200 scanning electron microscope (Advanced Metals Research Corporation, Bedford, MA) operating at 20 kV. Plaque was scraped from some devices and embedded in plastic. Ultrathin sections (60to 90 nm) of plastic embedded plaque were taken on a MT-5O00 ultramicrotome, stained with uranyl acetate and lead citrate, and examined using a Philips EM 300 electron microscope (N. V. Philips Gloeilampenfabrieken, Eindhoven, Neth.) o p erating at 60 kV. Least squares analysis of variance was used to determine significance of the various treatments (abraded, star, and grooved IMD-fittted; control quarters) on the occurrence of the variables: degree of leukocytosis into teat cistern, gland cistern, and parenchymal areas; cytologic alterations to the teat and gland cistern
3385
epithelial linings; and percentages of alveolar epithelium, alveolar lumen, and interalveolar stroma composing the mammary parenchyma. Duncan's multiple range test was used to determine differences amcmg treatment means. In addition, data from all quarters fitted with IMD were combined and compared with data from control quarters using the same analyses. RESULTS Bacterloioglc Status of Control and IMPFltted Quarters
At the time devices were placed, 14 quarters of the 12 cows were infected. Quarters fitted with the abraded model were uninfected, but coagulase-negative staphylococci (CNS) were isolated from three quarters fitted with the star model and three quarters fitted with the g~oovedmodel, and a colifom was isolated from one quarter fitted with the grooved model. Staphylococcur auteur (two quarters), Streprococcus dysgulactiae (one quarter), Streptococcus agalactiae (one quarter), Nocurdiu spp. (two quarters), and CNS (one quarter) were isolated from control quarters; remaining quarters were uninfected. Overall, 26% of IMD-fitted quarters and 41% of control quarters were infected at time of placement. At time of slaughter, when devices were removed, 13 quarters were infected, 9 of the original 14 infections persisted, 1 was cured spontaneously, and 4 new infections occurred. a-the new infections, Streptococcus uberis was isolated from 2 quarters fitted with the star model, CNS was isolated from a grooved IMD-fitted quarter, and Staph. aureus was isolated from a quarter fitted with the abraded model; no new infections occurred in control quarters. overall, 30% of IMD-fitted quarters and 29% of controls were infected at the end of the study. Gross and Mlcroscopic Histologic Observations
Gross observation of teat and gland cistern linings from IMD-fitted quarters at slaughter revealed no differences from those of control teats, even in those quarters where devices were apparently lodged in large ducts prior to removal. All IMD models were intact upon Journal of Dairy Science Vol. 74, No. 10, 1991
3386
NICKERSON ET AI.,.
retrieval and exhibited varying degrees of plaque formation. Light microscopic examination of teat and gland cistern linings of unfitted control quarters demonstrated an intact doublelayered epithelium resting on subepithelial connective tissue containing capillaries and lymphoid cells (Figure 1). Lymphocytes were present in or adjacent to the epithelial lining of all quarters, but neutrophils were usually observed only in IMD-fitted or in infected quarters. Light microscopic examination of teat and gland cistern linings of JMD-fitted quarters showed that some areas were identical to controls; however, in many cases, tissue alterations were observed where contact between the devices and the cisternal lining probably took place. Alterations ranged from degeneration and sloughing of the luminal epithelial cell layer, leaving a single layer intact Figwe 2), and pseudostratification and hyperplasia of the epithelial lining (Figure 3). to stratification and keratinization of the epithelium (Figure 4). In areas exhibiting stratified squamous epithelium, cells in various stages of mitosis were commonly seen (Figure 5). Neutrophils were occasionally observed in alveolar lumina of the mammary parenchyma adjacent to the gland cistern as well as adjacent to and within the gland and teat cistern linings. These leukocytes were only rarely observed breaching the epithelium where they were found squeezing between luminal cells of the epithelial lining and entering cisternal areas Figure 6). Eosinophils were observed in the subepithelial stroma of the gland cistern in two quarters fitted with abraded IMD.
Microscopic examination of areas excised fmm sites where IMD were found lodged in large ducts at time of necropsy revealed marked pathological changes. Most of the epithelial lining was stratified as in Figure 4, and the subepithelial connective tissue was thickened due to fibrous proliferation and leukocyte infiltration. Leukocytes were also observed traversing the stratified squamous epithelium, and they formed clots in the lumen of some ducts with material thought to be fibrin. Examination of mammary parenchymal areas adjacent to the gland cistern in uninfected control quarters for leukocyte infiltration revealed typical secretory alveoli with few or no neutrophils as shown in Figure 7. Such tissues exhibited a large alveolar luminal area and a very limited stromal area, indicating active secretion. Similar areas in tissue sections from quarters fitted with IMD showed a moderate leukocyte infiltration into alveolar luminal areas (Figure 8). Percentage tissue area occupied by alveolar lumen was less, and that occupied by interalveolar stroma was greater, than that found in unfitted quarters. A histologic analysis of IMD quarters compared with control quarters, regardless of infection status, is in Table 1. Leukocytosis into the teat and gland cistern linings as well as into the parenchyma adjacent to the gland cistern was greater in abraded and grooved models compared with control quarters and those fitted with the star model. Significant cytologic alterations to the gland cistern linings were observed in all IMD-fitted quarters compared with controls. Greatest alterations were found in quarters fitted with the abraded
Figure 1. The epithelial lining and subepithelial conneCtive tissue typical of those found in the teat and gland cistern linings from norma& minfectcd mammary glands without. . J' devices. The double layered epithelium is composed of a t x d , cuboidal cell layer underlying a luminal colarrmpr cell layer. Tbe subepithelial connective tissue has few infiltrating lenlrocytes; lympbcytes are present at arrows. ~ 4 7 5 . Figure 2. "be cpihIiaI lining Qpkd of that found in isolated arcaq of the gland cistern of guarters fitted with inr~~manrmadevices classified as degenerating and sloughed The d a c e columnar cells (D) have become flattenad and are degamate however, the basal epithelial layer (B) remains intact. Infilfra?ing neumphils (N) are observed at arrows. x475. P i 3. The epithelial lining and subepithelial connective tissue typical of those found in some gland cisternal areas of quarters fitted with inhamammm devices classified as pseudostratified or hypaplastic. The epithelium appears multilaymd, and some layers pmbude into the cistern lumen (L.).InNtrating mtrophils are observed in the epithelium at arrows. x475. Pigum 4. Epithelial lining typical of that found in some areas of quarters fitted with the abraded intramammary devices classified as stratified sqnemous epithelium. Neutrophils are present in the subepithelial connective t i w e (arrows), and the IenLocytcs have formed a clot with fibrin m the cistern lumen. ~200. J o d of Dairy Science Vol. 74, No. 10, 1991
TISSUE RESPONSE TO INTRAMAMMARY DEVICES
3387
J o d of Dairy Science Vol. 74. No. 10, 1991
3388
NICIEWSON ET AL.
model. Analysis of mammary parenchymal components demonstrated that, although percentages of alveolar epithelium (both m e s combined) were similar in IMD-fitted and control quarters, IMD quarters (all models combined) exhibited significantly less alveolar luminal area and significantly more interalveolar stroma than controls (compare Figures 7 and 8). This was mainly attributed to quarters fitted with the star model. Observations were similar for Zone 1 tissues. Results of the analysis of Zone 2 tissues followed a similar trend; however, quarters fitted with the abraded model exhibited significantly less alveolar lumen and significantly more stroma than controls. In uninfected quarters, degree of leukocytosis as well as alterations to the epithelial linings were significantly greater in IMD quarters (all models combined) (Table 2). Among models, leukocytosis was lowest in quarters fitted with the star model in both teat and gland cistern areas. Cytologic alterations to the gland cistern lining were greatest in abraded IMD-fitted quarters. The IMD-fitted quarters (all models combined) exhibited less luminal area (P < .05) and more stroma (P < .05) in Zone 1 and Zone 2 tissues than tissue from unfitted controls. Among models, quarters fitted with the star model exhibited the lowest luminal areas and greatest stromal area overall (both zones combined) and for Zone 1 parenchyma. In infected quarters, presence of inflammation generally obscured any effect of the IMD. In most cases, degree of leukocyte inliltration and score for cytologic alterations were elevated over values of uninfected quarters (Table 3). Degree of leukocytosis into the teat cistern lining and cytologic alterations to the gland cistern lining were higher (P < .05) in abraded quarters compared with other quarters.
Leukocytosis into the gland cistern parenchyma was greater in quarters fitted with the grooved model compared with controls. Percentage lumen was lowest and stroma highest in abraded quarters compared with all other treatments; this was the case overall (both zones combined) and for Zone 1 parenchyma. Percentage epithelium in Zone 1 parenchyma of abraded IMD quarters was significantly lower than all other treatments. Data analysis demonstrated no effect of stage of lactation on the histologic differences observed between control and IMD quarters (data not shown). Overall, degree of leukocytosis and alterations to the cisternal linings were not different between cows 4 8 0 d in milk and those >305 d in m i k However, analysis of parenchymal components showed that cows >305 d in milk exhibited less alveolar luminal area and more interalveolar stroma than cows 4 8 0 d in milk. This would be expected because tissue from cows >305 d in milk would be more involuted than tissue from cows earlier in lactation. Likewise, length of time that IMD were in place and parity had little effect on differences observed among treatments. Tissues from quarters exposed to IMD R mo exhibited slightly greater leukocytosis and alterations to the cisternal linings than those from quarters exposed >5 mo. Tissues from quarters exposed to IMD 6 mo also exhibited less lumen and more stroma; however, three of the four cows in this group were >305 d in milk, and tissue was probably undergoing involution. Electron microscope examination of the devices showed that plaque was present to some degree on all models but tended to be thicker on devices retrieved from infected quarters (Figures 9 and 10). Plaque was composed mainly of intact leukocytes but also
Figure 5. Stratified epithelium from a quarter fitted with an abraded inlmtmammmy device model; note neutrophils (N) and mitotic figures
0. ~750.
Figure 6. Portion of gland cistern lming of quarter fitted with an abraded ni?-y device model showing neutrophils (arrows) in various stages of migration across the epithelial layer. x750. Figure 7. Parenchymaltissue typical of that found in Zone 1 and Zone 2 of most uninfected int~amatnmarydevice fitted quarters c l a s s i i i as n o d . Such tissues exhibited a large alveolar luminal area Q and a very l i i t e d stromal (S) area, indicating copious secretory activiv. ~ 2 0 0 . Figure 8. Parenchymal tissue found in Zone 2 of a quarter fitted with an abraded intramammary device model. Percentage tissue area occupied by ahrtolar lumen is less, and that occupied by interalveolar stroma is more, than that in Figure 7. Note several leukocytes present in Lhe alveolar lamioa (arrows). x200.
J o d of Dairy Science Vol. 74, No. 10. 1991
TISSUE RESPONSE To
INTRAMAMMARY DEVICES
3389
I
Journal of Dairy Science Vol. 74, No. 10, 1991
3390
NICKERSON ET AL..
contained cell debris, keratin, and amorphous infection in abraded IMD quarters was threematerial (Figure 11). Where plaque was mini- fold that of controls (37 vs. 12%). At time of mal or absent, leukocytes were often observed dissection in the present study, no gross tissue attached to the clean plastic surfaces of some effects of the devices were observed, even in IMD models (Figure 12). two quarters in which IMD were lodged in large ducts and had to be removed with forDISCUSSION ceps. This is in contrast to findings of Collins Data on the microbiologic status of IMD- et al. (5), who observed petichiae dong the fitted and control qumers at time of slaughter walls of the gland cistern and larger ducts of showed that prevalence of natural infection some IMD-fitted quarters. Histological responses to IMD were similar was similar in IMD and control quarters (30 to those reported earlier (5); however, pathovs. 29%). New infections oc& more frelogical changes were not as marked. In the quently in IMD quarters than in controls (6 vs. 0), but, overall, the total number was small. present study, thickening of the subepithelial Similarly, Corlett et al. (6)found no difference connective tissue of the gland cistern lining in prevalence of natural infection among con- and fibroblast proliferation were not observed trol and smooth IMD quarters in three com- except at IMD sites in large ducts of infected mercial herds. In contrast, Collins et al. (5) quarters, although leukocyte infiltration and found that, 6 mo after placement, prevalence of epithelial hyperplasia and keratinization were
TABLE 1. Effect of -ni devices on tissue characteristics of the teat and gland cistern linings aad mammary parenchyma regardless of quarter infection stam. IntTamammary
Tissue Compo~nt
Abraded (n = 6)
device model
(n = 13)
orooved (n = 8)
Star
(n = 26)
Control (n = 19)
All
Teat cistern lining Leukocytosis1
2.17'
AlteratioosZ
1.oob
1.46b 1so.
1.81* 1.46"
1.69 1.43*
lSlb 1.12"
2.67' 3.33.
1.69b 1.%b
2.27" 2.15b
2.05* 2.19,
1.2gb 1.m
2.w
1.3Sb
223'
1.81*
1.2gb
Gland cistern lining LeuLocytosis'
Alterations2 Leukocytosis into gland c i s t a n parenchymal ~ammaryparenchyma3 (both zones combimd) Epithelium
Lumen Slroma
Zone 1 parenchyma3 Epithelium Lumen stroma Zone 2 parenchyma3 Epithelium Lumen
Strana
3 1.07" 28.24' 40.69*'
30.w 21.& 47.42.
30.26" 30.93' 38.81b
30.51 26.66* 42.83'
3 1.36' 32.66' 35.98b
28.98' 40.83' 30.19b
30.96' 20.73b 48.31'
31.01' 33.38' 35.61b
30.75 28.72* 40.53+
31.%' 34.67' 33.37b
33.15' 15.65b 5120"
30.25'b 23.27* 46.48h
29.ab 28.43' 42.09*
30.35 24.53* 45.12*
30.78ab 30.6ga 38.54b
4b*cMcans without a superscript in common differ (P < .OS). 'Data expressed as average score for infiltrating cells where 1 = very few or no neutrophils observed; 2 = moderate neutrophil infiltration;and 3 = marked neutrophil infitration. 2Data expressed as average score where 1 = OOmLal tissue architecture; 2 = sloughing of the epithelial lining;3 = epithelial hyperplasia or pseudostraMicatioq 4 = stratiftcation with keratin formation. 3Data expressed as percentage of each pareochymal component. *DifZercd from control (P < .OS). Journal of Dairy Science Vol. 74, No. 10, 1991
3391
TISSUE RESPONSE TO INTRAMAMMARY DEVICES
TABLE 2. Effect of intramammary devices on tissue characteristics of the teal and gland cistern linings and mammary parenchyma of uainfecte4i quarters. Intramammary device model
Tissue component Teat cistern lining Leukocytosis' Alterations2 Gland cistern lining LeukOCytOSiSl
Alterations' Leukocytosis into gland cistern parenchymal Mammary parenchyma3 (both zones combined) Epithelium Lumen Stroma Zone 1 parenchyma3 Epithelium Lumen stroma Zone 2 parenchyma3 Epithelium Lumen Strollla
(n = 8)
Grooved (n = 7)
All (n = 19)
Control (n = 12)
1.75' 1 .oob
1.06b 1.13b
1.61' 1.56
1.39* 1.29*
1 .wb 1 .oob
2.56 3 .W
1.44b 2.00b
2.11' 2.06b
1.87* 2.13*
1 1.04=
2.W
1. o b
2.1 1'
1.63*
1.08b
32.298 37.29' 30.4Zb
30.82' 18Nb 50.24'
30.34' 32.87' 36.79b
30.76 28.23. 41.01*
30.47a 36.51' 33.d
32.64' 56.67' l0.6p
31.50.
15.5@ 53.00'
30.Ma 34.03b 35.31b
31.23 29.91* 38.86*
31 .25' 38.21b
30.108 22.5ObC 4 7 . e
30.W 3 1.68& 38.32bc
30.27 26.51* 43.22*
29.72' 34.86' 35.42'
Abraded (n = 4)
3 1.94' 17.92' 50.14'
Star
.w
30.54b
4b*cMeans without a superscript in common differ (P c .05). 'Data expressed as average score for infiltrating cells where 1 = very few or no neutrophils observed; 2 = moderate neutrophil infiltration; and 3 = marked neultrophjl infiltration. 2Data expressed as average score where 1 = normal tissue architecture; 2 = sloughing of the epithelial liniOg; 3 = epithelial hyperplasia or pseudostratifkation; 4 = stratification with keratin formation. b a t a apressed as percentage of each parenchymal component. *Differed from control (P < .05).
noted. Histological examination of tissues from cows fitted with IMD over a 3-yr period revealed epithelial hyperplasia, squamous metaplasia, and polyp formation (1). Leukocytosis as well as alterations to the epithelial linings were greater in infected IMD quarters than uninfected IMD quarters as previously demonstrated (5). Eosinophilic infiltration into the subepithelial tissues of the gland cistern was commonly observed in two quarters fitted with abraded devices as observed by Collins et al. (5). The areas exhibiting sloughing and pseudostratification were very similar to those observed in teat cisterns of nonlactating quarters fitted with intracisternal bead devices (ICB) (13). Similarly, neutrophil infitration into the cisternal linings of IMD quarters was identical to that observed in ICB-fitted quarters, and, as
observed with the ICB, IMD-fitted quarters exhibited greater leukocytosis into the mammary parenchyma adjacent to the gland cistern. In contrast to the IMD study, use of ICB did not result in reductions in luminal area and increases in stroma, although the latter study was performed in nonlactating cows. Huston and Heald (8) studied the morphologic features of the gland cistern linings of six firstcalf heifers 3 wk after placement of smooth IMD. They found that IMD-fitted quarters exhibited a significant change toward a single layer of epithelial cells compared with a double layer in control quarters. This phenomenon was also observed in quarters fitted with IMD in the present study, although hyperplasia was also noted. It is possible that the single-layered epithelium observed by Huston and Heald (8) represented sloughing of the Journal of Dairy Science Vol. 74, No. 10, 1991
3392
NICKERSON ET AL.
Journal of Dairy Science Vol. 74, No. 10, 1991
3393
TISSUE RESPONSE TO INTRAMAMMARY DEVICES
TABLE 3. Effect of inftrammmq devices on tissue characteristics of the gland and teat cistern linings and mammary parenchyma of infected quarters.
Inf-y Tissue component Teat cistern lining Leukocytosis' Alterations2 Gland cistern lining Leukocytosis' AI terations2 Luekocytosis into gland cistern parenchyma' Mammary parenchyma3 (both zones combined) Epithelium LUmell stroma Zone 1 parenchyma3 Epithelium Lumen Stroma
Zone 2 parenchyma3 Epithelium Lumen stroma
device model
(n = 5)
Grooved (n = 1)
Au (n = 8)
Control (n = 5)
3.w 1 .cob
2.lOb 2.w
2.25b 1 .38ab
2.25 1.70
2.25b 1.3Iab
3.w 4.w
2.10bc 1.90b
2.63ab 2.38b
2 .w 2.3W
1.79 1.18'
2.wb
1
2.5@
2.W
1 .63b
Abraded (n = 2)
Star
.e
28.61' 10.14b 61.25'
30.42' 25.62a 43.%b
30.05' 25.83' 44.12b
30.19 23.85 45 .%
32.36' 28.33' 39.31b
21 .67b 9.17b 69.16'
30.33' 27.Wa 42.67b
31.94" 3 1.67" 36.39b
29.90 26.61 43.49
32.74' 30.7Sa 36.51b
35.56" 11.11a 53.33'
30.55h 23.74. 45.71'
28.15b 20.w 51.85'
30.19 21.33 48.48
3 1 .98ab 25.91' 42.11'
4b+cMcanswithout a superscript in common differ (P < .05). 'Data expressed as average m r e for intiitrating cells where 1 = very few or no neutrophils observed; 2 = moderate neutrophil infiltration; and 3 = marked neutrophil infiltration. 2Data expressed as average score where 1 = normal tissue architecture; 2 = sloughing of the epithelial lining; 3 = epithelial hypezplasia or pseudostratificatiow 4 = stratification with keratin formation. 'Data expressed as percentage of each parenchymal component. *Differed from control (P < .OS).
luminal epithelial layer after 3 wk. Continued presence of the devices may have resulted in a tissue reaction to the sloughing, leading to hyperplasia of the epithelium. In the present study, significant leukocytosis and alterations to the teat cistern epithelium were also noted. At time of slaughter, all but two IMD were found free floating in the gland cistern; two were found lodged in large ducts. However, during lactation, the grooved IMD of one quarter of one cow was typically palpated in the teat cistern prior to milking and had to be forced up into the gland cistern prior to machine milking. It is possible that IMD models in other quarters periodically descended to the teat cistern, eliciting a tissue reaction. The IMD quarters tended to exhibit less alveolar luminal area and more stroma, indicat-
ing reduced secretory potential or greater involution compared with controls. Involution was greatest in infected abraded IMD quarters. Although quarter milk production was not measured in this study, another investigation found that quarters fitted with abraded IMD produced less milk (7), and it was speculated that this was a result of the pathological changes observed (5). Paape and Weinland (19) also found that quarters fitted with abraded IMD tended to produce less milk compared with unfitted controls. Five of the cows in the present study were in late lactation (over 305 d), and this could have contributed to greater involution in IMD quarters. Paape et al. (20) found that cows fitted with abraded IMD had shorter lactations than controls and suggested that an increase in milk SCC in treated quarters could have contributed to an earlier involution of the mammary secretory tissue. Journal of Dairy Science Vol. 74, No. 10. 1991
3394
NICKERSON ET AI.,.
REFERENCES Tissue wne within a quarter had marked effect on percentages of parenchymal tissue 1 Barnurn, D., M.Hazlett, R.Johnston, and J. Costerton. components for some IMD models. In Zone 1 1985. B i o f i i on bovine inl@auymaydevices and their effect on the gland and rmk Page 81 in Roc. of uninfected quarters, those fitted with the 35th Annu. Mtg. Can. SOC. Mimbiol., Dalhousie abraded model exhibited significantly more luUniv., Halifax, NS, Can. men and less stroma than other treatments 2 Bright, S. A., W. D. Schultze, M.J. Paape,and R. R. Peters. 1987. Effect of smooth or abraded intramam(Table 2). In Zone 2 of uninfected quarters, mary devices on the outcome of experimental cbalthose fitted with abraded Ih4D exhibited less lenge infection of bovine ma~mnary qumers with lumen and more stroma than the other treatEscherichia coli. Am. J. Vet. Res. 48:1290. ments. This suggests that the area adjacent to 3Brooks, B. E., and D. A. Barnum. 1982. The use of polyethylene intmmammary device in protection of the gland cistern in infected abraded IMD the lactating bovine udder against experimental Metquarters (Zone 2) was more involuted, whereas tion with Staphylococcus aureus or Streptococcus the deep parenchyma (Zone 1) either became agalacriae. Can. J . Comp. Med. 46267. more secretory to compensate for reduced ac4Brown, R. W., D. A. Bamum, D. E. Jasper, J. S. McDonald, and W. D. Shultze. 1981. Page 16 in tivity of Zone 2 or was actually becoming h%icrobiological procedures for use in the diagnosis of static as milk accumulated in alveolar lumina bovine mastitis. 2nd ed. Natl. Mastitis Counc., Inc., and displaced the interalveolar stromal tissue. Arlington, VA. 5Collins, R. A., G. Oldham, P. G. Francis, and N. It was thought that if IMD were detrimental to Craven. 1989. Pathological changes following implansecretory activity, the mammary parenchymal tation of intr~wnauuwarydevices (MD) and immunozone adjacent to the IMD site (gland cistern, logical mediator release by cells on recovered IMDs. Zone 2) would exhibit less secretory activity Res. Vet. Sci. M253. 6CorleK N. J., R. R. Peters, M. J. Paape, and W. D. than the deep parenchyma (Zone 1). However, Shultze. 1984. Effect of n im-v device on new with the exception of the abraded model, this infection rate, milk yield, and milk somatic cell counts generally was not the case. in Maryland dairy herds. J. Dairy Sci. 672571. The plaque observed on the surface of IMD 7Francis, P. G. 1985. preliminary results of a within cow trial relating to the effects of a polyethylene was very similar to that observed on ICB (14) htramammary device on bovine maslitis. E e l . Milchand on smooth molded polyethylene IMD (15); wirtsch. Fomchungsber. 37473. plaque tended to be thicker on IMD retrieved 8 Huston, G. E..and C. W. Heald. 1983. Effect of the device on milk infection status, yield, from infected quarters as observed previously and somauc cell count and on the morphological (15). Bacteria were rarely observed on the features of the lactiferons sinus of the bovine udder. surface of IMD from infected quarters. Am. J. Vet. Res. M1856. In conclusion, the three IMD models tested 9 Nemanic, M.K.,and D. R Pitelka. 1971. A scanning electron microscopy study of the lactating mamumy led to increases in leukocyte infiltration and to gland. J. Cell. Biol. 48:410. cytologic alterations in the teat and gland cislONickemon, S. C., P.A. Baker, and P. Trinidad. 1989. tern epithelial linings of the mammary gland. Local immunostimulation of the bovine mammary Quarters fitted with IMD appeared more ingland with interleukin-2. J. Dairy Sci. 72:1764. 11 Nickerson, S. C., R. L. Boddie, W. E. Owens, and J. voluted than unfitted controls, exhibiting less L.Watts. 1990. Effects of novel inlmmammq device alveolar luminal area and more interalveolar models on incidence of intmmammary infection after stroma. Involution tended to be greatest in experimental challenge with mastitis pathogens. J. Dairy Sci. 73~2774. infected quarters fitted with IMD.Plaque was 12 Nickerson, S. C., and C. W. Heald. 1981. Histopathopresent on all models and was thicker on IMD logic response of the bovine mammary gland to experretrieved from infected quarters. There were no imentally induced Staphylococcus uureus infection. effects of any of the IMD models tested on the Am. J. Vet. Res. 42:1351. 13Nickerson, S. C., W.J. Tboqson, W. M.Kortum, new IMI rate during the trial. Thus, the inand N. T. Boddie. 1987. Histological response of creased leukocytosis into fitted glands was not bovine mammary tissue to an intracisternal bead protective against mastitis, and the mammary device. J. Dairy Sci. 70687. tissue exhibited a pathologic response to IMD. 14Nickerson, S. C., W. J. Thompson, S. P. Oliver, and
R. M. Akm. 1988. Effects of intracisternal bead
ACKNOWLEDGMENTS
The authors thank W. M. Kortum for providing the IMD and E C. Huff for typing of this manuscript. Journal of Dairy Science Vol. 74, No. 10. 1991
devices on lacteal secretion components,plaque formation, and bacterial infection during the nonlactating period. Am. J. Vet. Res. 491205. 15 Paape,M.J., N. J. Corlett, and W. M.Kortum. 1987. Examination of intmmmmary devices from infected and uninfected naammary quarters by scanning electron microscopy. J. Dairy Sci. 701045.
TISSUE RESPONSE TO INTRAMAMMARY DEVICES 16 Paape, M.J., W. D. Schultze, N. J. Corlett, and 8 . T. Weinland. 1988. Wect of abraded ni1-y device on outcome in lactating cows after challenge exposure with Streptococcus ubcris. Am. J. Vet. Res. 49:790. 17Paape, M. J., W. D. Schultze, A. J. Guidry, W. M. Kortum, and B. T. Weinland. 1981. Effect of an intramammary polyethylene device on the concentration of leukocytes and immunoglobulins in milk and on the leukocyte response to Escherichia coli endotoxin and challenge exposure with Smphylococcus uureus. Am. J. Vel. Res. 42374. 18 Paape, M.J.. W. D. Schultze. R. R Peters, and N. J.
3395
Corlett. 1984. Effects of inlramauummy devices on milk somatic cells, milk yield and new infection rate. Page 148 in Proc. 23rd Annu. Mtg., Natl. Mastitis Couoc.. Inc., Arlington, VA. 19Wape, M I., and B. T. Weinland. 1988. Effect of abraded intramammary device on milk yield, tissue damage, and cellular composition. J. Dairy Sci. 71: 250. ZOPaape, M. J., G. Ziv, R. H. Miller, and W. D. Schultze. 1986. Update on the use of intmumrmmy devices in the control of mastitis. Page 87 in Proc. 25th Annu. M@. Natl. Mastitis COU~C.. hc., Arlington. VA.
Journal of Dairy Science Vol. 74, No. 10. 1991
