You are on page 1of 4

Laboratory Animal & Molecular Pathology

TOXICOLOGIC PATHOLOGY, vol 30, no 3, pp 390– 393, 2002


Copyright C 2002 by the Society of Toxicologi c Pathology

Periosteal Hyperostosis (Exostosis) in DBA/1 Male Mice


JOANN C. L. SCHUH, RANDY HALL, DINA LAMBERT, KIM HARRINGTON, KEN MOHLER, AND DAUPHINE BARONE

Immunex Corporation, 51 University Street, Seattle, Washington 98101, USA

ABSTRACT
Periosteal hyperostosis (exostosis) was identiŽ ed in 5.9% (11/188) of DBA/1 male mice 10 – 14 weeks old used for collagen-induce d arthritis
(CIA) efŽ cacy testing of immunomodulator y biologics. Mice with and without CIA in the affected limb, and also control and treated groups, were
involved, with bilateral lesions in one mouse. Hyperostosis was characterized by circumferential and raised masses of variable location, length, and
laterality, generally external to but occasionally breaching the periosteum of the metatarsals, metacarpals, tibia, femur, and humerus. Proportionally,
the hyperostotic foci consisted of cancellous and woven bone, followed by osteoid, cartilage, and Ž brous connective tissue and rarely in ammatory
cells. A displaced, presumably pathological fracture with callus formation was a concurrent lesion in only one case. Tartrate-resistant acid phosphatase -
positive cells were frequent at bony interfaces, indicating an active resorptive process. Periosteal hyperostosi s is an incidental and potentially common
Ž nding in DBA/1 mice. Underreporting may occur due to the male bias in disease expression of this CIA model, sampling bias (generally paws only),
tissue obliteration in the presence of CIA, and lack of comprehensive historical data on the backgroun d and aging lesions in this strain of mouse.
IdentiŽ cation of such confoundin g bony lesions is important to the interpretation of efŽ cacy studies, and suggests the need to further examine the
biology of bone developmen t in this strain of mouse.
Keywords. Animal model of human disease; efŽ cacy study; murine collagen-induce d arthritis; musculoskeleta l pathology.

INTRODUCTION lesions. Association with preexisting bone surfaces deŽ nes


Arthritis induced by priming and challenge with autolo- hyperostosis as periosteal (exostosis ), endosteal (enostosis ),
gous or heterologous collagen is a common murine model and trabecular or medullary (osteosclerosis ). Differentiation
used to characterize autoimmune arthritis and to test the of hyperostosis from primary bone tumors, trauma (primary
efŽ cacy of therapeutics for rheumatoid arthritis (5, 8). Ge- fractures), and dysplastic aging lesions may be difŽ cult (12).
netic susceptibility or resistance to collagen-induced arthri- This report represents a case series of histologically identi-
tis (CIA) in mice is related to both major histocompatibility Ž ed periosteal hyperostosis as a confounding lesion in young
complex (MHC) and non-MHC genes (1, 8, 19). Mice with male DBA/1 mice with experimental CIA used to evaluate
the MHC class II haplotypes H-2q (DBA/1, C3H, B10.Q) efŽ cacy of novel biologics for rheumatoid arthritis.
and H-2r (B10.RIII, RII/SJ), H-2w3, and H-2w17 are most
sensitive to induction of CIA (7). Even within susceptible METHODS
mouse strains, there is often marked heterogeneity of disease
The data presented in this report are derived from Ž ve sep-
incidence and severity associated with vendor source, sexes, arate efŽ cacy studies of CIA using 188 male DBA/1OlaHsd
interexperiment and interlaboratory variability, and immu- mice over a period of 3 years. Mice were obtained from
nization protocols. Ability to respond to autologous collagen,
Harlan UK (Blackthorn, Oxon), group housed in microisola-
T- and B-cell presence and responses, and cytokines (1, 4, 9, tor cages, fed rodent feed (Harland Tekland, Madison, WI)
10, 18, 20) also modulate susceptibility to, incidence of, and and water ad libitum, and acclimatized for 1 to 3 weeks
severity of CIA.
before induction of arthritis. At 6– 10 weeks of age, a stan-
Hyperostosis or osteosclerosis (12, 21) is a nonneoplastic dard collagen-induced arthritis protocol for mice was initi-
proliferation of woven and/or lamellar bone, cartilage, os- ated (4) with the following modiŽ cations. Mice were immu-
teoid, Ž broblasts, and giant cells that must be differenti-
nized with 100 l g chicken type II collagen (Sigma, Chicago)
ated from osteopetrosis, osteochondrodysplasia , and Ž bro- in complete Freund’s adjuvant (Difco, Detroit, MI) and chal-
osseous dysplasia and fracture callus (3, 12, 14, 15). lenged 21 days later with 200 l g collagen in incomplete
Increased bone or bone-cell formation (proliferative hy-
Freund’s adjuvant (Difco). Within 10 – 14 days, the majority
perostosis ) or decreased bone resorption (nonproliferative of the mice developed enlarged and red distal extremities
hyperostosis ) may be identiŽ ed depending on the age of the indicative of CIA, and affected mice were enrolled in the
treatment studies conducted using a blind code. The efŽ cacy
Address correspondenc e to: Dr. JoAnn Schuh, at her current address: treatment protocols consisted of 10 – 14 days of daily ip injec-
Applied Veterinary Pathobiolog y, 1752 Lewis Place NW, Bainbridge Island, tions of control protein of 50 – 150 l g human immunoglobu-
WA 98110-3663 , USA; e-mail: schuhj@bainbridge.net . lin G (IgG; huIgG) or comparable quantities of several novel
390 0192-6233/02$3.00 $0.00
Vol. 30, No. 3, 2002 HYPEROSTOSIS IN DBA/1 MICE 391

immunomodulators, at several test doses. Bony proliferation


was found to be unrelated to treatment; therefore, the spe-
ciŽ c immunomodulators, doses, and dosing regimens are not
described. The Ž nal group of four mice was collected after be-
ing excluded from a treatment study due to lack of clinically
evident CIA, and all were subsequently conŽ rmed to lack
histological changes typical of CIA (negative responders).
At necropsy, only joints were sampled from mice 10 – 14 wk
old. The skin was removed to the phalanges, and bilateral joint
areas were collected by incising proximal to the appropriate
joints and collecting the distal bone, muscle, and connective
tissue with minimal trimming of excess muscle mass. The
joint areas examined in all studies consisted of phalanges
to mid-radius/ulna (forepaw), phalanges to mid-metatarsals
(hindpaw ), and mid-metatarsals to distal tibia/Ž bula (tibio-
tarsal/talus ). The mid-radius/ulna to mid-humerus (elbow)
and distal tibia/Ž bula to distal femur (knee) areas were also
examined in four of the studies. The tissues were Ž xed in 10%
FIGURE 1.—Tibia, periosteum, DBA/1 mouse. Marked periosteal hyperostosis
neutral buffered formalin for 24 – 36 h, decalciŽ ed in formic typiŽ ed by osteoid, woven and lamellar bone. H&E. Bar 50 l m.
acid or ImmunoCal (Decal Corporation, Congers, NY), em-
bedded in parafŽ n, sectioned at 5 – 7 l m, and stained with
hematoxylin and eosin (H&E) and safranin O stains. The normal tissue architecture; therefore, evidence of prolifera-
slides were blind coded (except the last group of four mice) tion was only identiŽ ed concurrently with minimal CIA in a
before evaluation and scoring for in ammatory, bony, and single case. There was no association between location and
cartilaginous lesions. All Ž ndings were recorded, but only group assignment, and no evidence of laterality, with only one
typical in ammatory CIA lesions were considered further animal with bilateral lesions. Lesions were identiŽ ed equally
in evaluation of treatment efŽ cacy. Subsequently, additional in the paws and proximal long bones of the fore and hind
sections of non-CIA proliferative lesions were step-sectioned limbs.
(50-l m steps) and stained with H&E to further characterize Histologically, the characteristic proliferation consisted of
the Ž ndings, or stained for tartrate-resistant acid phosphatase well-organized and focal hypertrophy of woven and lamellar
(TRAcP) to identify osteoclastic and chondroclastic activity bone, osteoid, and cartilage (Figure 1). Compared to normal
(kit A387, Sigma). cartilage, cartilage foci in lesions were generally paler by
safranin O staining, indicating reduced glycosaminoglycans,
RESULTS and had scalloped outer margins, particularly in association
Except for a palpable mass of the left midshaft humerus in with multinucleate cells, presumably chondroclasts. Clusters
one mouse, all lesions were only identiŽ ed histologically. As of Ž broblasts, Ž brous connective tissue, and rare in amma-
summarized in Table 1, proliferative lesions of hard tissues tory cells were present near some of the cartilaginous foci.
were identiŽ ed in 1.0% (12/1150 ) of sections examined and in The proliferative mass enveloped the bone beyond preexist-
5.9% (11/164 ) of male DBA/1 mice. Affected mice included ing periosteum, with occasional and variable subperiosteal
controls (clinical CIA and no treatment), treated mice (clin- extensions. From the thickest area, the extra-periosteal hy-
ical CIA and treatment with one of several immunomodula- pertrophy gradually decreased proximal and distal until the
tory molecules and doses), and those excluded from treatment lesions blended imperceptibly with the normal periosteum.
(no clinical CIA and no treatment). Severe CIA obliterates The mass identiŽ ed grossly in one case correlated with an

TABLE 1.—Frequency of confoundin g periosteal hyperostosis in male DBA/1 mice used in collagen-induce d arthritis (CIA) efŽ cacy studies for immunomodulator y
molecules.

Number of mice Number of sectionsa


Total Affected Affected
Study Examined Affected Control Treated Examined Controls Treated Location (group b )
CIA 1 40 2 1/10 1/30 320 1/80 1/240 Metatarsal (T), femur (C)
CIA 2 30 3 1/15 2/15 238 1/120 2/118 Metatarsal (T), metatarsal (C), metacarpal (T)
CIA 3 32 2 0/9 2/23 244 0/68 2/176 Humerus (T), tibia (T)
CIA 4 82 3 1/10 2/72 316 1/39 2/277 Tibia, humerus (T), metacarpal (T)
CIA 5 4 1 1/4 Nonec 32 2/32 None Tibia (C-bilateral)
Total 188 11 4/48 7/140 1150 5/339 7/811
Percent 5.9 8.3 5.0 1.5 0.9
a
Generally, eight sections were collected per mouse except for CIA 4 (four sections of paws only). Not all sections could be evaluated histologically due to problems with orientation and
quality.
b (T)
indicates hyperostosis in a mouse treated with one of several immunomodulator y molecules after CIA induction; (C) indicates hyperostosis in a mouse in a control group after CIA
induction.
c Mice collected did not have clinical CIA and were excluded from subsequent treatment trials.
392 SCHUH ET AL TOXICOLOGIC PATHOLOGY

ations to bone may also occur in B6C3F1 mice treated with


4-hexylresorcinol, bisphosphonates , and estrogens (2, 12).
This case series of bony and cartilaginous proliferation ap-
peared to be most compatible with trauma or hyperostosis.
The presence of hyperostosis in control mice, including those
with and without proven CIA, eliminated treatment as a possi-
ble factor. Completely naive mice (i.e., those not undergoing
CIA induction) were not available for examination, making
it difŽ cult to eliminate the adjuvants or collagen used in the
CIA protocol as possible factors. However, four mice (CIA
5 group) were speciŽ cally collected to provide a sample of
mice that were negative responders to collagen immuniza-
FIGURE 2.—Humerus, DBA/1 mouse. Periosteal hyperostosis (arrows) at the
tion. That a hyperostotic lesion was found in only one site
distal humerus near the olecranon merges with a nonunion fracture (FX) and
callus formation. H&E. Bar 500 l m.
in one mouse in the absence of clinically or histologically
evident CIA suggests that the immunization reagents were
unlikely inducers of these tissue changes. Collagen alone in-
overriding and displaced fracture of the proximal humerus duces autoimmune arthritis (5), but there are no published
with periosteal hypertrophy merging into the fracture callus reports of Freund’s adjuvants alone inducing any nonin am-
(Figure 2). Step sections of the other samples did not iden- matory changes to bone or cartilage. Typically, intradermally
tify any additional fractures or evidence of preexisting or administered Freund’s adjuvants cause exuberant and persis-
concurrent lesions. In all cases, osteoblasts and osteoclasts tent granulomas and granulomatous in ammation localized
were frequent in the regions of lamellar but not woven bone. to the site of injection. Extension of the in ammation and
TRAcP staining was found in both mono- and multinucleate erosion of adjacent bone can be a debilitating adverse event
cells in association with scalloping of the lamellar bone and at the classic CIA induction site at the head of the tail in
cartilage (Figure 3). mice (Schuh, unpublishe d data). While conceivable, prolif-
eration in the absence of in ammation would be difŽ cult
to attribute to, and uncharacteristic of the known effects of
DISCUSSION the reagents used for CIA. Other differential diagnoses were
In the mouse, differential diagnoses for spontaneous eliminated due to the predominance of proliferative bone,
bony and cartilaginous reactions include fractures, mus- lack of in ammation and lysis, young age, focal nature of
culoskeletal trauma, in ammation, Ž brous-osseous dyspla- the Ž ndings, and speciŽ c-pathogen-free source colony for
sia (12), idiopathic hyperostosis (12, 16) or osteosclerosis these animals. Although sporadic periosteal trauma or frac-
(21), and retroviral-induced osteopetrosis (17). Congenital tures cannot be fully eliminated as a cause of the prolifer-
proliferative lesions of bone and cartilage include diffuse ation, the frequency is higher than would be expected for
spinal skeletal hyperostosis in the tiptoe-walking Yoshimura an incidental Ž nding. Furthermore, similar lesions were not
mouse (16), osteochondrodysplasi a in mice transgenic for identiŽ ed in many other strains of mice used in our facil-
interferon (IFN) c (15), and osteopetrosis in several dele- ity over this period. DBA/1 mice, particularly males, have
tional mutations (3, 11). Pharmacologically induced alter- high open-Ž eld and locomotor activity, and accidental frac-
tures might occur when these temperamentally hyperactive
mice are forcibly removed from cage surfaces for examina-
tion. However, if fractures were the underlying cause of the
proliferative lesions in this case series, lesions would be ex-
pected to occur almost exclusively in the predominate sites
of the tibia and Ž bula (12). Step sectioning did not reveal
any underlying fractures or other causes for the prolifera-
tion. Therefore, the Ž ndings were considered most compat-
ible with spontaneous periosteal hyperostosis as previously
described in aging B6C3F1 and ICR mice (2, 12). Despite
variations in the size and location of the hyperostosis, bone
predominated over cartilage and Ž brous connective tissue.
Multinucleate and mononucleate cells were frequent Ž ndings
in association with resorptive pockets along bone and carti-
lage, suggesting that these lesions were formed by a prolif-
erative (increased bone and resorption) rather than a nonpro-
liferative (reduced bone resorption) mechanism (12). In one
case where a fracture was identiŽ ed grossly, hyperostosis was
found for a considerable distance distal to but merging with
FIGURE 3.—Tibia, DBA/1 mouse. Tartrate resistant acid phosphatas e the nonunion fracture and callus formation. This suggested
(TRAcP)-positive multinucleate osteoclasts and chondroclast s (arrows) and that the hyperostosis was concurrent with a pathological frac-
mononucleat e precursors (arrowheads) were frequent and associated with active ture. However, fracture callus is a special form of hyperostosis
bone and cartilage remodeling. TRAcP histochemistry. Bar 50 l m. (6, 12), and this case highlights the difŽ culty of separating
Vol. 30, No. 3, 2002 HYPEROSTOSIS IN DBA/1 MICE 393

cryptogenic proliferation and active reparative processes in 2. Chhabra RS, Huff JE, Haseman J, Hall A, Baskin G, Cowan M (1988).
bone. Inhibition of some spontaneou s tumors by 4-hexylresorcino l in F344/N rats
Despite their use in numerous immunological studies and and B6C3F1 mice. Fundam Appl Toxicol 11: 685 – 690.
3. Dougall WC, Glaccum M, Charrier K, Rohrback K, Brasel K, DeSmedt
as a CIA susceptible strain, Ž ndings of background or inci-
T, Daro E, Smith J, Tometsko ME, Maliszewski CR, Armstrong A, Shen
dental lesions in tissues have not been described in DBA/1 V, Bain S, Cosman D, Anderson D, Morrissey PJ, Peschon JJ, Schuh JCL
mice. Testing for efŽ cacy of biologics often requires the use (2000). RANK is essential for osteoclast and lymph node development .
of speciŽ c strains of mice for which there is considerable im- Genes Dev 13: 2412 – 2424.
munological but little pathophysiologica l characterization. A 4. Durie FH, Fava R, Foy TM, Aruffo A, Ledbetter J, Aruffo A (1993). Pre-
nearly 6% incidence of spontaneous periosteal hyperostosis vention of collagen-induce d arthritis with an antibody to gp39, the ligand
in DBA/1 males suggests the need to fully evaluate the bone for CD40. Science 261: 1328 – 1330.
biology and pathology in this strain. This is supported by a re- 5. Durie FH, Fava R, Noelle R (1994). Collagen-induced arthritis as a model
cent study on genetic variability of ectopic osteoinduction in of rheumatoid arthritis. Clin Immunol Immunopatho l 1: 11 – 18.
inbred and outbred mice (13). Differences in bone mass and 6. Ferguson CM, Miclau T, Hu D, Alpern E, Helms JA (1998). Common
molecular pathways in skeletal morphogenesi s and repair. Ann NY Acad Sci
morphology after an osteoinductive implant were not associ-
857: 33 – 42.
ated with MHC haplotype, expression of bone morphogenic 7. Gustafsson K, Karlsson M, Andersson L, Holmdahl R (1990). Structures on
proteins, or spontaneous physical activity. High responders the I – A molecule predisposing for susceptibility to type II collagen-induce d
that produced bone, cartilage, and bone marrow were inde- autoimmune arthritis. Eur J Immunol 20: 2127 – 2131.
pendently derived strains such as the DBA/2, RFM, C57BL/6, 8. Holmdahl R, Andersson M, Goldsmidt TJ, Jansson L, Karlsson M,
and ADR mice. Poor responders, deŽ ned by predominate Malmstrom V, Mo J (1989). Collagen-induced arthritis as an experimental
cartilage production, included mice with common derivation model of rheumatoid arthritis. Immunogenetics , pathogenesi s and autoim-
from Bagg albino stock (BALB/c, A, CBA, and C3H). Ge- munity. Acta Pathol Microbiol Immunol Scand (APMIS) 97: 575 – 584.
netically based hyperactive osteoinduction after periosteal 9. Holmdahl R, Jansson L, Andersson M, Larsson E (1988). Immunogenetic s
trauma may be the basis for hyperostosis in DBA/1 mice. of type II collagen, autoimmunity and susceptibility to collagen arthritis.
The true incidence of spontaneous hyperostosis previously Immunology 65: 305 – 310.
10. Johansson AC, Vestberg M, Holmdahl R (2000). Non-major histocompati -
described in aging B6C3F1 and ICR mice (12, 21) and in our bility complex dependent variations in lymphocyte activity between inbred
DBA/1 mice remains unknown. Incidental bone abnormali- mouse strains susceptible to various autoimmune diseases. Scand J Immunol
ties are difŽ cult to document in mice due to the lack of gross 52: 21 – 29.
Ž ndings to aid identiŽ cation of affected tissues. Due to the 11. Karsenty G (1998). Genetics of skeletogenesis . Dev Genet 22: 301 – 323.
nature of CIA studies with DBA/1 mice, males are almost 12. Long PH, Leininger JR (1999). Bones, joints, and synovia. In: Pathology
exclusively used, and sampling artifact occurs as most in- of the Mouse, Maronpot RR, Boorman GA, Gaul BW (eds). Cache River
vestigators only evaluate the paws for correlation to clinical Press, Vienna, pp 645– 678.
scores. The true incidence of hyperostosis in DBA/1 mice 13. Marusic A, Katavic V, Grcevic D, Lukic IK (1999). Genetic variability of
could only be determined by evaluation of the entire skeleton new bone induction in mice. Bone 25: 25 – 32.
14. Murray AB, Schmidt J, Rieke L (1986). Retrovirus-induce d osteopetrosis
in males and females. In ammatory models such as CIA in
in mice. Am J Pathol 124: 319 – 323.
DBA/1 mice are also incompatible with proper evaluation of 15. Nii A, Reynolds DA, Young HA, Ward JM (1997). Osteochondrodysplasi a
spontaneous bony changes due to the destructive nature of occurring in transgenic mice expressing interferon-c . Vet Pathol 34: 431 –
CIA that obscures incidental Ž ndings except in unaffected 441.
or minimally in amed joint areas. Irrespective of the true 16. Okawa A, Goto S, Moriya H (1999). Calcitonin simultaneously regulates
incidence, recognition of periosteal hyperostosis as a con- both periosteal hyperostosi s and trabecular osteopenia in the spinal hyper-
founding lesion in DBA/1 mice with CIA is important to ostotic mouse (twy/twy) in vivo. Calcif Tissue Int 64: 239 – 247.
interpretation of efŽ cacy studies used for biologics intended 17. Schmidt J, Casser-Bette M, Murray AB, Luz A, Er e V (1987).
as modulators of rheumatoid arthritis. Additional evaluation Retrovirus-induce d osteopetrosis in mice. Effects of viral infection on
of DBA/1 mice may also aid in the characterization of genetic osteogenic differentiation in skeletoblast cell cultures. Am J Pathol 129:
503 – 510.
differences in bone biology.
18. Svensson L, Jirholt J, Holmdahl R, Jansson L (1998). B cell-deŽ cient mice
do not develop type II collagen-induce d arthritis (CIA). Clin Exp Immunol
ACKNOWLEDGMENTS 111: 521 – 526.
Technical assistance for these studies was provided by 19. Watson WC, Townes AS (1985). Genetic susceptibility to murine collagen II
Kathy Rohrbach, Bob Miller, Jon Jones, Jeffery Williams, autoimmune arthritis. Proposed relationship to the IgG2 autoantibod y sub-
and Ray Koelling. class response, complement C5, major histocompatibility complex (MHC)
and non-MHC loci. J Exp Med 162: 1878 – 1891.
20. Wei X, Leung BP, Arthur HM, McInnes IB, Liew FY (2001). Reduced
R EFERENCES incidence and severity of collagen-induce d arthritis in mice lacking IL-18.
1. Campbell IK, Hamilton JA, Wicks IP (2000). Collagen-induce d arthritis in J Immunol 166: 517 – 521.
C57BL/6 (H-2b) mice: New insights into an important disease model of 21. Yamasaki K, Itakura C (1993). Osteosclerosi s in aged ICR mice. Jpn J Vet
rheumatoi d arthritis. Eur J Immunol 30: 1568 – 1578. Sci 47: 799 – 802.