Mice Overexpressing Murine Oncostatin M (OSM) Exhibit Changes in Hematopoietic and Other Organs that Are Distinct from Those of Mice Overexpressing Human OSM or Bovine OSM

Materials and Methods

Construction of OSM retroviral vectors

Full-length cDNA encoding mOSM, hOSM, or bOSM was amplified by PCR using synthetic oligonucleotides that introduced an EcoRI site and a Kozak sequence at the 5′ end and a BglII site at the 3′ end. The PCR products were digested with EcoRI and BglII and subcloned into the MSCV2.1 retroviral vector (kindly supplied by Dr. R. Hawley, University of Toronto, Toronto, Ontario, Canada) (Fig. 1). Expression of the OSM insert was driven by the retroviral long terminal repeat (LTR) promoter already present in the MSCV2.1 plasmid. The constructs were sequenced using a 377 Sequencer (Applied Biosystems) to confirm the absence of mutations. The vectors contained a G418-resistance (G418^r) gene (phosphoglycerine kinase/neomycin resistance [pgk/neo] cassette) to allow selection of successful transfections.

OPEN IN VIEWER

Fig. 1

Construction of retroviral gene transfer vectors. A cDNA (750 bp) of one oncostatin M variant—either mouse (mOSM), human (hOSM), or bovine (bOSM)—was subcloned into the EcoRI/BglII site of vector MSCV2.1 just upstream of a pgk/neo cassette. Abbreviations: LTR, long terminal repeat; Ψ⁺, phi+ retroviral packaging signal.

Results

Consistent with our prior work, ²⁰ mOSM used a different signaling pathway in vitro than did either hOSM or bOSM. Treatment of cultured mouse NIH-3T3 cells with mOSM induced phosphorylation of mOSMRβ and the downstream signal transducer STAT3, but not mLIFRβ (Fig. 2). In contrast, treatment with either hOSM or bOSM led to phosphorylation of mLIFRβ and STAT3, but not mOSMRβ (Fig. 2).

OPEN IN VIEWER

Fig. 2

Immunoblot for ligand-induced receptor phosphorylation; NIH 3T3 cells. The mouse OSM-specific receptor OSMRβ was phosphorylated only after application of Amgen-fabricated (293T rmOSM) or commercially obtained (rmOSM) mouse oncostatin M, while treatment with human OSM (293T rhOSM [proprietary] or rhOSM [commercial]) or bovine OSM (rbOSM) led to phosphorylation of LIFRβ. The downstream effector STAT3, used in both the OSMRβ and LIFRβ signaling cascades, was phosphorylated by all OSM variants. The control lane designated 293T Mock did not contain an OSM cDNA.

Retroviral-mediated gene transfer of all three OSM variants was efficient and persistent. The transfection efficiency into bone marrow, as demonstrated by the in vitro granulocyte-macrophage colony-forming cell assay, varied between replicates but was equivalent for all the retroviral vectors: 65% to 80% for the control vector, 70% to 82% for mOSM, 68% to 85% for hOSM, and 74% to 84% for bOSM. Expression of vector-encoded OSM was readily detected in cells isolated from hematopoietic organs for at least 10 weeks after transplantation in mOSM-transduced animals (Fig. 3) or until death at approximately 4 weeks in animals that received hOSM or bOSM (data not shown). The level of expression in hematopoietic organs provided by the three OSM-containing vectors was similar (data not shown).

OPEN IN VIEWER

Fig. 3

Northern analysis of retroviral-mediated mOSM expression in vivo. Mice given an MSCV2.1 vector bearing mOSM expressed appreciable levels of mOSM, whereas animals given an empty MSCV vector alone did not. The numbered lanes represent total splenic RNA preparations from individual mice. The RNA was harvested 10 weeks after intravenous introduction of the retroviral vector.

The hematologic profile induced by mOSM diverged significantly from those elicited by hOSM or bOSM. As early as 2 weeks after transduction, mice overexpressing hOSM or bOSM exhibited significant elevations in circulating numbers of platelets (Fig. 4) and neutrophils, while erythrocyte, lymphocyte, and eosinophil numbers were significantly reduced (Table 1). Circulating monocyte numbers were also substantially elevated in mice that received a nonmurine OSM, although the large interindividual variation prevented demonstration of a significant difference. This constellation of changes persisted throughout the study in mice that overexpressed hOSM or bOSM. When measured in mice overexpressing mOSM, these same hematologic parameters were comparable with those of control animals (Table 1).

OPEN IN VIEWER

Table 1

Hematologic profile elicited in mice by mOSM diverges from those produced by nonmurine OSM variants. ∗ .

Transgene	CM	mOSM	bOSM	hOSM
Red blood cell (×10−6/μl)	9.00 ± 0.08	9.45 ± 0.08	†7.25 ± 0.52	†7.20 ± 0.34
Platelet (×10−6/μl)	945 ± 12	904 ± 24	†1648 ± 110	†1348 ± 54
White blood cell (×10−3/μl)	7.55 ± 0.83	6.30 ± 0.76	7.21 ± 0.89	7.05 ± 0.51
Lymphocyte (×10−3/μl)	6.11 ± 0.77	4.99 ± 0.62	†1.20 ± 0.21	†1.30 ± 0.33
Neutrophil (×10−3/μl)	0.71 ± 0.07	0.79 ± 0.08	†4.17 ± 1.02	†3.32 ± 0.98
Monocyte (×10−3/μl)	0.21 ± 0.05	0.23 ± 0.04	1.49 ± 0.80	2.03 ± 1.11
Eosinophil (×10−3/μl)	0.06 ± 0.02	0.05 ± 0.01	†0.02 ± 0.00	†0.02 ± 0.00
Basophil (×10−3/μl)	0.03 ± 0.00	0.03 ± 0.00	0.05 ± 0.01	0.06 ± 0.02

∗

Blood cell counts represent group means ± SEM from samples acquired 4 weeks after transduction with a retroviral vector containing no gene insert (CM [control]) or mOSM, bOSM, or hOSM (n = 5 per group).

†

Significant difference from the CM group, P ≤ .05.

OPEN IN VIEWER

Fig. 4

Comparison of circulating platelet numbers; mouse whole blood. Overexpression of hOSM or bOSM, but not mOSM, enhanced platelet production relative to the platelet numbers measured in control mice (CM). Values could not be determined for hOSM or bOSM overexpressing mice at later time points because all animals in these two groups had died. Values represent group means ± SEM.

Anatomic alterations were induced in multiple organs by all three OSM variants, but the structural phenotype incited by mOSM was distinct from those produced by hOSM and bOSM. Mice overexpressing hOSM or bOSM died at 4 to 5 weeks after transplantation, whereas animals overexpressing mOSM survived to at least 52 weeks. Body weights of mice overexpressing hOSM or bOSM decreased significantly by approximately 25% after transplantation (Fig. 5), whereas those of animals overexpressing mOSM were similar to control mice. Interestingly, all OSM-expressing animals had significantly higher absolute and relative spleen weights and lower absolute and relative thymus weights compared to those of mice that received the control vector (Table 2). A plethora of microscopic lesions were apparent in tissues of mice transduced with OSM, particularly in lymphoid organs (Table 3). Lesions in animals overexpressing either bOSM or hOSM were almost always equivalent in both pattern and degree; if present, changes in mOSM-transduced mice were uniformly of lesser degree. Additional OSM-associated alterations were noted in nonlymphoid organs, and again the bOSM and hOSM constructs yielded a similar phenotype that was distinct from that induced by mOSM (Table 4). Using a blinded paradigm for histopathologic analysis, sections of most organs could be separated into two (control and mOSM, nonmurine OSM) or three (control, mOSM, non-murine OSM) phenotypes with essentially 100% accuracy using HE-stained sections (Tables 3 and 4).

OPEN IN VIEWER

Table 2

Body and organ weight changes induced by mOSM are less pronounced than those produced by nonmurine OSM variants. ∗

Transgene	CM	mOSM	bOSM	hOSM
Total body weight (TBW)	23.8 ± 0.4	23.2 ± 0.7	17.4 ± 0.2† ‡	16.9 ± 0.7† ‡
Spleen weight	0.098 ± 0.004	0.154 ± 0.018†	0.195 ± 0.012† ‡	0.165 ± 0.005†
Spleen as % of TBW	0.41 ± 0.02	0.67 ± 0.02†	1.12 ± 0.05† ‡	0.98 ± 0.03† ‡
Thymus weight	0.042 ± 0.002	0.023 ± 0.002†	0.008 ± 0.001† ‡	0.008 ± 0.001† ‡
Thymus as % of TBW	0.18 ± 0.01	0.10 ± 0.01†	0.05 ± 0.00† ‡	0.05 ± 0.00† ‡

∗

Weights (in grams) represent group means ± SEM from samples acquired 4 weeks after transduction with a retroviral vector containing no gene insert (CM [control]) or mOSM, bOSM, or hOSM (n = 5 per group).

†

Denotes a significant difference from the CM group, P ≤ .05.

‡

Denotes a significant difference between groups receiving a nonmurine OSM and those given mOSM, P ≤ .05.

OPEN IN VIEWER

Table 3

Anatomic reorganization of lymphoid organs induced by mOSM exhibits a different pattern from the structural phenotypes incited by nonmurine OSM variants. ∗

Transgene	CM	mOSM	bOSM	hOSM
Lymph node morphology (Fig. 6)
Thin cortex, loosely organized medulla	5
Thick cortex containing numerous small lymphocytes†		5
Thick medulla containing plasma cells and granulocytes†			5	5
Spleen morphology (Fig. 7)
Abundant white pulp, prominent B and T cell zones	5
Reduced white pulp, especially T cell zone†		5
No white pulp† and florid extramedullary hematopoiesis			5	5
Thymus morphology (Fig. 8)
Thick cortex, myriad lymphocytes (continuous sheets)	5
Thin cortex, fewer lymphocytes (in intersecting clusters)†		5
No cortex, attenuated medulla†			5	5

∗

Tissues were examined 4 weeks after transduction with a retroviral vector containing no gene insert (CM [control]) or mOSM, bOSM, or hOSM (n = 5 per group). Values represent the number of animals within a group with a given organ morphology as assigned using a “blinded” paradigm for histopathology analysis.

†

Denote changes considered to be primary OSM-mediated effects.

OPEN IN VIEWER

Table 4

Nonlymphoid organs have differing architectures in mice that overexpress either mOSM or a nonmurine OSM variant. ∗

Transgene	CM	mOSM	bOSM	hOSM
Bone morphology
Normal diaphyseal cortices and marrow cavity	5	5
Marked osteosclerosis and marrow hypoplasia			5	5
Pancreas morphology
Normal exocrine tissue (dense acini, many granules)	5	5
Atrophic exocrine tissue (loose acini, few granules)			5	5
White fat morphology
Normal adipocytes	5	5
Atrophied adipocytes, diffuse macrophage infiltration			5	5
Adrenal gland morphology
Thin inner cortical zone	5	5
Fourfold thickening of the inner cortical zone			5	5
Cervix morphology
Hyperplastic mucosa consisting of mucus-producing cells	5	5
Thin mucosa, perivascular inflammatory cell infiltrates			5	5
Stomach (Glandular) morphology
Normal mucosa and serosa	5	3
Minimal lymphocyte infiltrate of mucosa		2	1
Minimal mucosal mucinosis, serosal lymphocyte infiltrate			4	5

∗

Tissues were examined 4 weeks after transduction with retroviral vector containing no gene insert (CM [control]) or mOSM, bOSM, or hOSM (n = 5 per group). Values represent the number of animals within a group with a given organ morphology as assigned using a “blinded” paradigm for histopathologic analysis. None of these changes were considered to be primary OSM-mediated effects.

OPEN IN VIEWER

Fig. 5

Comparison of total body weight. Overexpression of hOSM or bOSM, but not mOSM, led to significant body weight loss (denoted by asterisks) relative to time-matched control mice (CM) by 4 weeks after transplantation. Values could not be derived for hOSM or bOSM overexpressing animals at 30 weeks because all mice in these two groups had died. Values represent group means ± SEM.

All three OSM variants induced substantial anatomic reorganization of the various lymphocyte compartments within lymphoid organs, and again the phenotype associated with mOSM diverged significantly from those provoked by hOSM or bOSM. For example, lymph nodes of animals overexpressing any of the three OSM variants were markedly enlarged. However, in mOSM-transduced mice the cortex and paracortex were markedly expanded by coalescing cords and nodules consisting of T cells and, to a lesser extent, B cells. In both bOSM- and hOSM-expressing mice the medulla was more prominent, though the cellular composition of the expanded regions varied substantially from individual to individual; the most numerous components were plasma cells and granulocytes (Fig. 6, inset) and—in bOSM animals only—T lymphocytes (Fig. 6). Spleens of all mice overexpressing any OSM variant were greatly enlarged by extensive extramedullary hematopoiesis, though the extent was less in mOSM mice. In the white pulp, the periarteriolar lymphatic sheath (T-cell domain) was attenuated in all OSM-expressing animals, whereas the peripheral white pulp (B-cell domain) appeared to be minimally or unaffected in mOSM mice but was absent in bOSM and hOSM animals (Fig. 7). The thymic cortex was reduced in thickness for all OSM-transduced animals, though the degree was modest in mOSM animals and pronounced in bOSM and hOSM mice (Fig. 8). The thymic cortices of mice given bOSM or hOSM but not mOSM also had many fewer T cells (Fig. 8). Scattered B lymphocytes were evident in the thymuses of all OSM-overexpressing animals; the B cells were widely dispersed in mOSM mice but concentrated near the periphery in animals expressing a nonmurine OSM (Fig. 8). The alterations in white cell populations were considered to be primary effects of OSM overexpression as they were induced by both the mouse and the nonmurine OSM variants. The florid extramedullary hematopoiesis in animals that received nonmurine OSM was interpreted as a tertiary consequence engendered by bone marrow ablation (see the following paragraph).

OPEN IN VIEWER

Fig. 6

Lymph node, mouse overexpressing oncostatin M. Overexpression of OSM substantially altered lymph node morphology relative to control mice (CM) by 4 weeks after transplantation, but the pattern induced by hOSM or bovine bOSM differed from that elicited by mOSM. The lymph node of CM was partitioned distinctly into two zones, a thin outer rim of densely packed B lymphocytes and a larger, loosely aggregated central region containing T cells. Lymph node cortices and paracortices of mOSM expressors were markedly thickened by coalescing cords and nodules of lymphocytes comprising both B and T cells. In contrast, enlarged lymph nodes of bOSM- and hOSM-transgenic had a more prominent medulla filled with T cells intermingled with (in inset) plasma cells and granulocytes; the medullary cell composition varied considerably among individuals within both these groups. Stains: HE (top), anti-B220 (B cell; middle), anti-CD3 (T cell; bottom [serial sections relative to anti-B220 panels]).

OPEN IN VIEWER

Fig. 7

Spleen; mouse overexpressing oncostatin M. Overexpression of OSM markedly changed splenic morphology relative to control mice (CM) by 4 weeks after transplantation, but the phenotype produced by hOSM or bOSM diverged from that incited by mOSM. The spleen of CM was partitioned distinctly into two zones, red pulp and white pulp, and the latter region was also divisible into a pale inner area (the periarteriolar lymphatic sheath, or PALS; T cells) and a dark outer rim (the peripheral white pulp, or PWP; B cells). The gross splenic enlargement seen in all OSM-transgenic mice resulted from extensive extramedullary hematopoiesis. The PALS was attenuated in mice expressing any OSM variant, while the PWP was absent in animals overexpressing bOSM or hOSM but not mOSM. Stains: HE (top), anti-B220 (B cell; middle), anti-CD3 (T cell; bottom [serial sections relative to anti-B220 panels]).

OPEN IN VIEWER

Fig. 8

Thymus; mouse overexpressing oncostatin M. Overexpression of OSM noticeably distorted thymic morphology relative to control mice (CM) by 4 weeks after transplantation, but the configuration elicited by hOSM or bOSM departed from that induced by mOSM. The thymus of CM was partitioned distinctly into two zones, a thick outer rim of densely packed lymphocytes and a loosely aggregated core. The cortex contained myriad T cells. Thymic cortices of mOSM expressers were modestly reduced in thickness and less densely packed with T cells, while organs of mice with bOSM or hOSM lacked cortical tissue and contained almost no T cells. All mice that overexpressed an OSM had B cells in the thymic cortex; the wide dispersion of these cells noted in mOSM mice differed distinctly from their peripheral concentration in animals expressing a nonmurine OSM. Stains: HE (top), anti-CD3 (T cell; middle), anti-B220 (B cell; bottom).

Interestingly, overexpression of hOSM or bOSM resulted in effacement of the bone marrow. The initial change, evident in a few animals, appeared to be multifocal production of fibrous connective tissue consistent with osteosclerosis (data not shown). Fibrotic foci contained little or no bone. Most animals appeared to have more advanced changes, characterized by merging of connective tissue islands and formation of bony trabeculae within the fibrotic foci (Fig. 9). Numerous cathepsin K-labeled osteoclasts lined trabeculae, while endosteal surfaces in diaphyses of control and mOSM-expressing mice had no osteoclasts. The absence of osteosclerosis and osteoclast proliferation in mOSM-transduced animals suggests that this effect was a secondary response specific to nonmurine OSM rather than a primary OSM-mediated event. Additional secondary alterations in several nonlymphoid organs were evident in all or almost all mice given a nonmurine OSM but occurred in none of the animals treated with mOSM (Table 4).

OPEN IN VIEWER

Fig. 9

Tibial diaphysis; mouse overexpressing oncostatin M. Mice expressing bOSM or hOSM, but not mOSM, developed extensive fibrosis and/or osteosclerosis of the bone marrow. Intersecting trabeculae of woven bone essentially effaced bone marrow. Numerous cathepsin K–positive osteoclasts (arrowheads) lined the trabeculae, while endosteal surfaces in bone shafts of control and mOSM-expressing mice had no osteoclasts. Stains: HE (left column) and anti-cathepsin K (osteoclast marker; right column).

Discussion

OSM and OSMRβ (the OSM-binding β-chain subunit of the heterodimeric receptor) are closely related, respectively, to LIF and its preferred β-chain, LIFRβ. It has been postulated based on experiments with recombinant hOSM or bOSM applied to in vitro and/or in vivo mouse systems ^{5, 6, 22, 40} that OSM and LIF share many biological functions due to the structural similarity of their β-subunits and their sharing the same signal-transducing molecule. For example, OSM is reported to substantially enhance megakaryocyte production and bone growth while depleting thymic T cells based on data derived from bOSM-transgenic mice. ^{22, 40} Interestingly, mice injected with or made transgenic for LIF also develop thymic atrophy and have increased platelet counts, as well as splenomegaly and weight loss. ^{24, 36} This concordance of biological response between two closely related cytokines raises the possibility that crosstalk mediated by a trans-species interaction of nonmurine OSM with mouse LIFRβ is responsible for the LIF-like physiologic reactions that have been attributed to OSM. Indeed, previous in vitro work in our laboratory using purified ligands and mouse NIH 3T3 cells supported this premise ²⁰ : hOSM signals via mLIFRβ but not mOSMRβ, whereas mOSM activates mOSMRβ but not mLIFRβ. Our current in vivo data extend this prior work and show that in vivo phenotypes in mice induced by mOSM and nonmurine OSMs diverge. Thus, prior studies of nonmurine OSMs in mouse systems may have unintentionally probed LIF signaling rather than authentic OSM biology.

In the present study, retroviral-mediated overexpression of mOSM, hOSM, or bOSM resulted in a prominent phenotype distinguished by several marked changes. Three alterations—lymph node enlargement, splenic white pulp atrophy, and thymic atrophy—were considered to be primary consequences of OSM overexpression as these sites were affected in mice transduced with any OSM (Table 3). However, the biological effects of mOSM in lymphoid organs clearly differed from those of non-murine OSM as shown by partially to completely distinct patterns of structural alterations (Table 3; Figs. 6–8). Interestingly, the anatomic changes elicited by bOSM and hOSM did not exactly correlate, as the prominent T lymphocyte population observed in the expanded lymph node medulla of bOSM-overexpressing mice was not as evident in the hOSM animals (Fig. 6). The exact molecular pathways responsible for these discrepancies among the various OSM is uncertain, especially given a previous report showing that mOSM, hOSM, and bOSM mediated extrathymic, thymus-independent maturation of T lymphocyte precursors in lymph nodes to a similar degree. ⁵ Our present data do support a major finding of the prior report, as B lymphocytes are scattered throughout the thymic cortex of OSM-overexpressing mice but not control animals (Fig. 8). However, the anatomic effects of these three OSM variants diverge between the previous study ⁵ and our experiment (e.g., increased [prior] versus decreased [present] thymic weights in bOSM- and hOSM-overexpressing mice). The differences between the prior ⁵ and current data likely reflect subtle disparities in the gene construct (driven by the Lck [leukocyte-specific protein tyrosine kinase] promoter in the prior study versus the retroviral LTR promoter in the current study) and engineering methods (pronuclear injection in the prior study versus retroviral transfer into adults in the current study). We consider the most suitable explanation for the dissimilarities to be that our hypothesis is correct: murine OSM induces one or more signaling cascades that are distinct from those mediated by nonmurine OSM, whether the system is evaluated in vitro (Fig. 2 and our prior report ²⁰ ) or in vivo (Tables 1, 3, 4). Further studies will be required to ascertain the specific pathway(s) involved in mediating these variations in vivo.

Other marked changes limited to mice receiving bOSM or hOSM were interpreted as secondary effects. These changes included weight loss, pancreatic and white fat atrophy (consistent with cachexia), thrombocytosis, osteosclerosis, splenomegaly (from extramedullary hematopoiesis, likely secondary to osteosclerotic bone marrow ablation), adrenocortical atrophy, cervical mucosa attenuation and inflammation, and gastric mucinosis and inflammation. With the exception of the adrenal, cervical, and gastric modifications, these features equate to those described previously in mice after transgenic overexpression of bOSM ^{22, 40} or LIF. ^{24, 36} Additional work will be necessary to define what role, if any, LIF stimulation by nonmurine OSM plays in the pathogenesis of the adrenal, cervical, and gastric alterations.

As noted previously, mice that overexpressed mOSM also developed a pronounced phenotype. However, the lymphoid changes in the mOSM-transduced animals were distinct from those noted in mice that received a nonmurine OSM. Overexpression of mOSM resulted in atrophy of T cell depots in splenic white pulp and thymus; the other anatomic lesions characteristic of the phenotype induced by nonmurine OSMs in these organs were not engendered by mOSM. Mice that overexpressed mOSM also exhibited thickening of the lymph node cortex and paracortex by whorls of small lymphocytes (both B and T cells), a finding that was absent in animals receiving vectors bearing bOSM or hOSM (Fig. 6). Using our novel data to reappraise previously published literature, we conclude that the genuine physiologic function of appropriate OSM signaling in vivo (mOSM/mOSMRβ) primarily involves governance of lymphocyte subsets in secondary lymphoid organs, a finding in agreement with prior reports that chronic OSM overexpression modulates extrathymic T-cell development. ^{2, 3} In particular, the T lymphocyte lineage appears to be the principal target for OSM control as indicated by prominent changes in T cell-specific domains: atrophy of the periarteriolar region of the splenic white pulp (Fig. 7) and thymic cortex (Fig. 8). This interpretation is supported by the pronounced increase in T cells within the cortex of enlarged lymph nodes of mice that overexpress mOSM (Fig. 6). Our current work did not determine the mechanism(s) by which OSM might fulfill this task. Feasible explanations for these morphologic changes are that activating the OSMRβ pathway affects T lymphocyte production in the thymus and/or subsequent homing to other lymphoid organs. However, the distorted distribution of B cells in the lymph node cortex might also indicate a function for OSM signaling in B lymphocyte biology. Further work will be needed to fully elucidate mechanisms by which OSM/OSMRβ mediate these effects.

The substantial alterations in lymphoid organ structure elicited by OSM overexpression raised several other potential realms for future inquiry. Other leukocyte populations (e.g., dendritic cells, macrophages) and other cell populations might reasonably be anticipated to exhibit altered functions, distributions, and/or kinetics in the presence of excess OSM. For example, mice transgenic for bOSM under the control of the Lck promoter (Lck-bOSM) have increased numbers of high endothelial venules (HEV), ²¹ vascular elements that are particularly rich in OSMRβ. ¹⁷ The resemblance at low magnification between the lymph node anatomy of our bOSM-transfected mice and the Lck-bOSM transgenic animals from the prior experiment ²¹ suggest the possibility that HEV are enhanced in lymph nodes of some or all of our OSM-overexpressing mice. Again, the authenticity of such potential changes fell outside the main scope of our current work and therefore remains a fertile area for further study.

Importantly, our current data strongly imply that other roles imputed to OSM, such as regulating body metabolism (total weight, fat deposition) and platelet production, arise from nonmurine OSM binding to LIFRβ. This conclusion is supported by the absence of body weight loss and the atrophy of adipose and exocrine pancreatic tissues (all hallmarks of LIF-induced cachexia) in mOSM-transfected animals versus the presence of marked body weight loss accompanied by fat and pancreatic atrophy in mice overexpressing bOSM or hOSM. The considerable hOSM- and bOSM-induced thickening of the inner adrenal cortex (a major site of cortisol production) and universal shift of the cervical mucosa morphology to the metestrus conformation (suggestive of repressed estrus cycling) likely reflect sequelae of cachexia-induced stress; an immediate explanation for the mucosal mucinosis and lymphocytic infiltration of the glandular stomach was not apparent. Our present results leave open the possibility for a few overlapping duties for LIF and OSM, as overexpression of all three OSMs led to splenomegaly associated with extramedullary hematopoiesis. However, the nature of the splenic change elicited by mOSM likely differs from that incited by bOSM or hOSM. This interpretation is reasonable as the engorgement resulting from mOSM was substantially less than that afforded by the other two molecules and was not associated with bone marrow ablation (a marked finding in mice overexpressing bOSM or hOSM). Regardless, the pronounced divergence between the murine phenotype induced by mOSM and those produced by bOSM or hOSM clearly implies that the predominant consequences of homologous OSM signaling (mOSM/mOSMRβ) are readily discernible from those elicited by nonhomologous OSM signaling. Taken together, these data affirm that prior studies of nonmurine OSMs in mouse systems may have explored LIF biology rather than real OSM physiology and highlight the need for conformation of homologous ligand and receptor interactions as the definitive indicator of cytokine biology and phenotypic characterization in genetically engineered mice.