Altered Hematopoiesis, Behavior, and Sexual Function in μ Opioid Receptor–deficient Mice

The μ opioid receptor gene Oprm was disrupted in embryonic stem (ES) cells by replacing the coding exon 1 with a PGK-neo selection marker using the method of gene targeting as described (7, 8). Targeted ES cells were microinjected into C57/ BL6 mouse blastocysts, and the resultant chimeric mice were mated to wild-type female mice of the Swiss black strain to obtain germline transmission of the targeted gene. Genomic Southern blotting and nested PCR were used to confirm the presence of the disrupted gene in the progeny mice. Mice were housed at 72–74°F and 60% humidity with 12-h light/dark lighting cycle. Water and food were available ad libitum.

Unseparated marrow and spleen cells were plated respectively at 5 × 10⁴ and 5 × 10⁵ per ml in 1.0% methylcellulose culture medium with 30% vol/vol fetal bovine serum (Hyclone Labs, Inc., Logan, UT) and 1 U/ml recombinant human erythropoietin (Amgen Biologicals, Thousand Oaks, CA), 50 ng/ml recombinant murine steel factor (Immunex Corp., Seattle, WA), 5% vol/vol pokeweed mitogen mouse spleen cell– conditioned medium, and 0.1 mM hemin (Kodak Co., Rochester, NY). Colonies deriving from hematopoietic progenitors were scored after 7 d of incubation in a humidified environment at 5% CO₂ and lowered (5%) O₂ as described (9–11). Absolute numbers of progenitors per organ were calculated based on the number of viable, unseparated nucleated cells per femur or spleen, and the number of colonies scored per number of cells plated. The percentage of progenitors in S-phase was estimated by the high specific activity tritiated thymidine kill technique (9–11).

Locomotor activity measurements were recorded in a quiet room under dim light by video cameras, and the tapes were analyzed off-line. Individual mice placed in a round plastic chamber were habituated for 15 min before the activity was recorded. Horizontal activity was scored as breaking either one of the two center lines (perpendicular to each other). Vertical activity was scored as the animal raised its head and/or stood up along the side wall. Morphine (in 0.9% saline) was injected intraperitoneally at a 2.3 mg/kg, the ED₅₀ dosage of analgesia in mice (12). U62066 (dissolved in ethanol and diluted in saline) was injected intraperitoneally at a dose of 3 pmol/kg.

Mating behavior was tested essentially as described (13). Sperm from the right-side epididymis was flushed out in 5 ml of phosphate-buffered saline, and sperm number counted both by Coulter cell counter and by warmed Neubauer counting chamber (14). Sperm motility was the percentage of sperm showing forward progressing, linear motion.

Data of heterozygotes and homozygotes were analyzed for statistically significant differences from that of wild-type mice by Student's t test. An asterisk (*) indicates significant difference (P <0.05) from the wild type. All data are shown as mean ± SEM.

The μ opioid receptor gene Oprm, a single copy gene on mouse chromosome 10 (15), was disrupted in ES cells by gene targeting (16, 17) for generation of μ opioid receptor–deficient mice. Analysis of genomic DNA from progeny mice confirmed the presence of the disrupted gene (data not shown). Deficiency in the μ opioid receptor does not appear to cause major changes in mouse development, as mice of all three genotypes grew with comparable body weight (Fig. 1), without noticeable changes in neuroanatomy or histology (data not shown). When tested for opioid-mediated antinociception using the tail flick latency assay, we observed similar results to those reported by Matthes et al. (5), i.e., a lack of morphineinduced increase of tail flick latency in mice homozygous for the deficient receptor gene (Oprm^−/−; data not shown).

Morphine and other opioids are known to enhance motor activity in rodents. Locomotor activity is defined as two types of behavior: horizontal (the extent of the mouse's running around in observation cage) and vertical (the frequency of the animal's upward movement, either as rearing or standing up against the side of the cage). When observed for spontaneous horizontal activity, wild-type mice habituated over a period of time after being placed in the observation cage (Fig. 2,A). In homozygous mice (Oprm^−/−) that had not been exposed to morphine, the spontaneous horizontal activity was greatly reduced compared with that of the wild type (Fig. 2 A). This suggests that even in naive animals that have not been exposed to exogenously administered opioids, the μ opioid receptor may play a role in modulating the basal tone of spontaneous locomotor activity.

When morphine was injected, wild-type mice exhibited increased locomotor activity (Fig. 2,B), a phenomenon commonly observed in mice (18). The homozygous mice, on the other hand, were not affected by morphine injection, and their spontaneous horizontal activity stayed relatively constant over the course of the test (Fig. 2,B). This difference between wild-type and homozygous mice is significant both for the 10-min segments during recording (Fig. 2,B, left), and over the entire recording time of 60 min (Fig. 2,B, right). Interestingly, heterozygous mice displayed intermediate levels of horizontal activity that did not change after morphine injection (Fig. 2 B).

When the vertical activity was examined, there was no significant difference among three genotypes of mice (Fig. 2,C). Upon morphine injection, the vertical activity in wildtype mice was greatly reduced (Fig. 2,D), in agreement with previous studies (18). In homozygous mice, morphine did not change the level of vertical activity during the test period (Fig. 2,D), indicating a lack of μ opioid receptor for mediating the morphine effect. Just as for horizontal activity, vertical activity in heterozygous mice was not affected by morphine (Fig. 2 D), suggesting that one normal allele of the μ opioid receptor gene is not sufficient to overcome the deficit of the other mutated allele in this respect.

These results are consistent with previous observations that in normal (wild-type) mice, morphine increases horizontal locomotor activity and reduces vertical locomotor activity (18). Our data demonstrate that the μ opioid receptor plays a central role in morphine modulation of locomotor activity, since in both homozygous and heterozygous mice, the morphine effects were largely abolished.

Animal studies support the role of opioids in modulating the host defense system (19, 20). For example, both β-endorphin and enkephalins, endogenous opioids that have high affinity for μ and δ opioid receptors, affect mature cells of the hematopoietic system, i.e., monocytes, macrophages, and neutrophils (19). However, no studies have thus far linked the μ opioid receptor system to blood cell production, especially for progenitor cells of myeloid origin that give rise to granulocytes, monocytes/macrophages, erythrocytes, and platelets. Hematopoiesis, the production of blood cells, is regulated by cytokines, which include members of the chemokine family (21–24). Chemokine receptors, for example the interleukin-8 receptor, are members of the G protein-coupled receptor family like the μ opioid receptor, and the μ opioid receptor displays moderate levels of sequence homology to both the interleukin-8 receptor and C-C chemokine receptors (25). Given these structural parallels and implied functional interactions, it was of interest to examine the μ opioid receptor–deficient mice for potential changes in hematopoietic processes.

To explore a potential role of μ opioid receptors in hematopoiesis, myeloid blood cell production in mice was examined using bone marrow, spleen, and blood. We found that the absolute numbers of femoral bone marrow granulocytemacrophage (CFU-GM), erythoid (BFU-E), and multipotential (CFU-GEMM) progenitor cells in bone marrow were significantly higher in homozygotes (Oprm^−/−) compared to wild-type mice (Table 1), whereas there were no apparent differences in nucleated cellularity in the bone marrow and spleen (Table 1) or in the leukocytes, erythrocytes, and platelet counts in the blood (data not shown). Also, although myeloid progenitors in the bone marrow and spleen were in a slowly cycling state in wild-type mice, these progenitors were in rapid cell cycle in homozygous mice as determined by the percentage of progenitors in S-phase (Table 1). The cycling status of myeloid progenitors in bone marrow and spleen of heterozygous mice (Oprm^+/−) was intermediate between that of wild type (Oprm^+/+) and homozygotes (Oprm^−/−). Also, the absolute numbers of CFU-GM, BFU-E, and CFU-GEMM in spleen were increased in homozygotes compared to wild-type mice, although these differences were not significant.

Cells of the myeloid lineage, such as monocytes, macrophages, and neutrophils, are among the body's first line of defense against infections, and serve important functions such as antigen presentation, phagocytosis of foreign materials, and production/release of stimulatory and suppressing cytokines active on myeloid and lymphoid cell production and function (21–24). Endogenous opioids in circulation could potentially activate the μ opioid receptor on these cells, which may modulate hematopoiesis. Our finding of enhanced production of the myeloid progenitors in μ opioid receptor–deficient mice thus provides strong evidence of this receptor's involvement in blood cell production, most likely as a negative regulatory influence. The exact link remains to be determined, and may be directly or indirectly mediated at the level of hematopoietic stem and progenitor cells. The enhanced proliferation status and number of myeloid progenitors in bone marrow and spleen without a subsequent increase in mature blood cell number suggests involvement of additional cell–cytokine mediated events between premature and mature blood cell regulation.

During breeding, we noticed that it took homozygous male mice (Oprm^−/−) a significantly longer time to impregnate a female than wild-type male mice (Table 2), and the offspring litter size by homozygous males was smaller than that by wild type (Table 2). Two factors could contribute to such a change: a reduction in mating activity and a change in reproductive physiology. To examine whether the μ opioid receptor plays a role in the sexual behavior in mice, we measured the mating activity in male mice. When an estrous wild-type female mouse was placed in a cage with a male mouse, the wild-type male would start to investigate the female and, within a few min, initiate sexual activity by mounting the female mouse. The homozygous males (Oprm^−/−), on the other hand, showed a significantly lengthened period before they initiated sexual activity with the female mouse (Table 2). Also, whereas the majority of the mounting attempts by a wild type male were appropriate (mounting from the rear of the female), the proportion of appropriate mounting by a homozygous male was much reduced (Table 2). In addition, the number of intromissions during the test session was significantly reduced in homozygous males as compared with the wild type. These results indicate that sexual activity is reduced in the homozygous male mice compared with that in the wild type. The reduction of sexual activity in homozygous mice does not appear to be caused by reduced locomotor activity, because during mating tests there was no significant difference among the wild-type, heterozygous, and homozygous male mice in the total contact time, determined as the accumulative time when the male was investigating the female mouse, including mounting and sniffing of body and genital areas (Table 2).

When physiological parameters of the male reproductive system were examined, they also appeared to be altered by disruption of the μ opioid receptor gene. As shown in Table 2, even when a homozygous male initiated intromission, there was little ejaculation (a lack of seminal plugs) during the 2-h test. This is not due to a complete inability to ejaculate, as homozygous males were able to form seminal plugs when left overnight in the cage with female mice (Table 2). However, the percentage of overnight plugs was reduced compared with that of the wild-type male mice.

The male reproductive system in homozygous male mice appears to be functional, as the prostate and seminal vesicle weights were of similar values among different genotypes (Table 2). However, sperm motility and sperm counts in homozygous mice were significantly reduced compared to those in wild-type mice (Table 2). These data suggest that the development of sperm may be affected by disruption of the μ opioid receptor gene even though the gross anatomy of the male reproductive system in homozygous male mice is not. Taken together, these results indicate that the μ opioid receptor may play an important role in the sexual behavior and reproductive physiology in mice. Further studies are needed to address whether the μ opioid receptor acts directly on the reproductive organs or functions by modulating a central mechanism.

We thank Drs. Judith Strong and Michael Vasko for comments on the manuscript, and Tom Hendrickson, Andrea Caperell-Grant, and Paula Stevenson for technical assistance.

This work was supported in part by the National Institutes of Health (NIH) grants DA09116 and DA09444 to L. Yu; HL56416, HL54037, and a project in P01 HL53586 to H.E. Broxmeyer; and DK38185 to J.A. Tischfield. L. Yu is the recipient of a Research Career Development Award (NS01557) from the NIH. S. Aronica was supported by the NIH training program T32DK07519 to H.E. Broxmeyer. S.J. Engle and A. Mestek were supported by the NIH training grant T32HD07373 to the Department of Medical and Molecular Genetics.