.
SPA and opiate analgesia have been found to work by activating descending pathways from brain stem to spinal cord where pain inhibition actually takes place, although a role for ascending paths has also been adduced (76). Electrically stimulating or microinjecting opiate drugs into certain brain areas activates centrifugal controls on the flow of nociceptive information through spinal cord dorsal horn neurons that in turn project both to nociceptive reflex paths and to the brain. In other words, the brain can powerfully control its own pain input. Brain stem loci supporting both SPA and opiate analgesia include the mesencephalic periaqueductal gray matter (PAG) and several nuclei in the rostral ventral medulla whose projections descend via the dorsolateral funiculus to the dorsal horn and the nucleus caudalis (see ref. 10 for review).
The demonstration of high concentrations of opiate receptors and opioid peptides in areas of the CNS involved in pain inhibition further suggest that the endogenous opioids are likely mediators of SPA. In addition to its susceptibility to antagonism by naloxone, SPA manifests tolerance with repetition and cross-tolerance with morphine (68), meeting the three criteria for opioid mediation commonly used today.
To be credible, a natural pain inhibitory system must have natural stimuli that activate it. Under emergency conditions of fight or flight, pain inhibition may be more adaptive to the organism than pain perception itself, allowing the animal to focus attention not on pain but on strategies for defense or aggression (57). Three different groups demonstrated in 1976 that, indeed, noxious and/or fear-provoking stimuli can produce stress-induced analgesia (SIA), suggesting that stress is an important trigger for activating endogenous analgesia substrates (3,38,93). SIA has subsequently been demonstrated in a wide variety of animals, including humans, and can be elicited by a wide range of stressors (see ref. 47 for reviews).
SIA and SPA have been found to have many similar characteristics, supporting the view that stress is a natural trigger for endogenous analgesia systems in the brain. For example, both SPA and SIA can be mediated by at least two distinct neurochemical systems. SPA elicited from certain brain sites (e.g., ventral PAG) satisfies the three criteria for opioid mediation described above, whereas SPA from other brain areas (e.g., dorsal PAG) must be categorized as nonopioid (22). Likewise, certain stress paradigms produce opioid SIA, whereas, by application of these same criteria, SIA from certain other stress paradigms results in nonopioid SIA (106,114). In one study we showed that opioid forms of SIA manifest cross-tolerance with opioid SPA, suggesting that opioid SPA and SIA share a common opiate receptor site (105).
It has been shown for almost all SIA paradigms that by simply altering the parameters of the same stressor it is possible to elicit either opioid or nonopioid SIA. Specifically, we have found that within certain boundaries, and holding other parameters constant, less severe stressors (e.g., weaker or briefer inescapable footshock, rotation of shorter duration, or swimming in warmer water) cause opioid-mediated SIA, and more severe stressors (e.g., stronger or longer footshock, longer duration rotation, or swimming in colder water) cause nonopioid SIA (23,62,104,106). Yet the same duration and intensity of footshock that causes nonopioid SIA if presented continuously, causes opioid SIA if presented intermittently (53,106). The SIA substrate is even more complex; there are at least two different kinds of opioid SIA, one supporting learned helplessness and requiring consciousness and an intact pituitary-adrenal axis (``hormonal'' SIA), the other not (``neural'' SIA) (106,114).
Such studies have helped to reveal the multiplicity of mechanisms and pathways that can be differentially activated as a function of the specific parameters of the stressor. Different parameters of even the same stressor cause physiological reactions that differ not only neurochemically (opioid versus nonopioid) but also in terms of anatomical, behavioral, endocrinological, immunological, and even oncological criteria. Questions have remained regarding the specific neurochemical nature, anatomical loci, and functional significance of nonopioid mechanisms of analgesia.
There are four themes being addressed by our current research that relate to the topic of this volume. Three of these seek to address one or another of the aforementioned shortcomings of opioid analgesics: (i) in an attempt to develop analgesic compounds that might not display the side effects associated with morphine-like drugs, we are at present investigating the neurochemical bases of nonopioid SIA; (ii) we are exploring the ability of a certain class of antagonists to prevent the development of tolerance to opioids; and, (iii) we are examining the genetic determinants of analgesia in an attempt to understand individual differences in the magnitude of the analgesic response.
A fourth major aim of our current research effort is to assess the deleterious effects of pain and stress on health-related end points. Specifically, we are examining the influence of pain and stress on immunological (natural killer cell activity) and oncological (tumor metastasis) criteria. Our findings in this area underscore the urgency of improving methods to alleviate the suffering of cancer patients, as it is becoming increasingly clear that pain is more than just a symptom but can itself be an aggressive disease, a significant pathogen that, if ignored, can have life-threatening consequences.
Recent studies have demonstrated the involvement of excitatory amino acids (EAAs; glutamate and aspartate) in endogenous pain inhibition (2,43). The EAA receptor subtype known as the N-methyl-D-aspartate (NMDA) receptor has been specifically implicated in SIA; NMDA receptor antagonists have been found to attenuate morphine analgesia (43,58), as well as opioid SIA produced by conspecific defeat (100) and restraint (58).
Whereas relatively high doses of NMDA blockers were used in studies demonstrating antagonism of opioid analgesia, we find that a very low dose of the specific, noncompetitive NMDA receptor antagonist, MK-801 (dizocilpine; 0.075 mg/kg, i.p.), produces a selective attenuation of nonopioid SIA while not reducing morphine analgesia. This conclusion is based on two converging lines of evidence (62,63).
First, we have reported that MK-801 completely blocks swim SIA ([SSIA]; 3 min swimming in 20oC water) in the opiate receptor-deficient mouse strain CXBK. These mice, developed by recombinant inbreeding techniques and found serendipitously to display marked deficits in opiate binding, exhibit only nonopioid SSIA. MK-801 partially blocks SSIA in the opiate receptor-rich CXBH strain, as does naloxone. MK-801 and naloxone delivered together completely block SSIA in CXBH mice (63).
Second, in a recently published study (62),
we compared the effects of the same low dose of MK-801 on SSIA in mice
elicited by a 3-min swim at three different water temperatures. Consistent
with earlier reports (20,25,79,104), we find
that, holding other stress parameters constant, opioid mechanisms of analgesia
seem to be triggered by less severe stress, in this case warmer water temperature
(32oC), whereas nonopioid mechanisms are
triggered by more severe stress, in this case lower water temperature (15oC).
SSIA from an intermediate temperature (20oC)
causes a mixed opioid/nonopioid analgesia, one partially blocked by naloxone.
MK-801 powerfully blocked the nonopioid 15oC
SSIA, had no significant effect on the opioid 32oC
form, and in combination with naloxone completely blocked the mixed opioid/nonopioid
20oC SSIA
(Figure 1)
.
In both of these studies, 0.075 mg/kg (i.p.) MK-801 was found to be ineffective in attenuating analgesia produced by 10 mg/kg (i.p.) morphine sulfate. These experiments were performed using the hot-plate test as the algesiometric assay; we also have evidence that 0.075 mg/kg MK-801 blocks nonopioid SSIA in the formalin test (111), a model of tonic, inescapable pain. Thus, we believe we have provided powerful evidence that the effect of MK-801 at this low dose is largely specific to nonopioid forms of analgesia, suggesting a role for the NMDA receptor in such analgesic mechanisms. An EAA-containing pathway connecting two brain structures crucially involved in descending pain inhibition=mthe PAG and the nucleus raphe magnus=mhas been identified by anatomical (2,116) and electrophysiological techniques (116). Other evidence suggests that the neurotransmitters serotonin (45,91,92) and histamine (35) might also be involved in nonopioid analgesic circuitry.
Although, as noted above, MK-801 appears to antagonize nonopioid analgesia selectively, we have most recently observed that this drug powerfully blocks the analgesic effect of the selective k opioid agonist, U-50,488, without attenuating morphine analgesia in rats (48). That analgesia produced by U-50,488 is opioid in nature is demonstrated by its naloxone reversibility (112). However, clear differences between k and m analgesia have been noted (see refs. 41,71 for reviews), and k agonists have attracted considerable attention as novel (non-m) analgesic compounds due to their relative lack of side effects compared to m agonists (28). The demonstration that k analgesia as well has an NMDA component underscores the relevance of investigating NMDA mechanisms of analgesia.
Our findings, as is typical of the literature, were based on data obtained from adult male subjects. Interestingly, we have recently observed (74) that nonopioid SIA in female mice, although equipotent to that of males, is not NMDA-mediated. MK-801, at the same or even higher dose, is completely ineffective in attenuating 15oC SSIA (and the nonopioid component of 20oC SSIA) in females. However, ovariectomized female mice display the male pattern of sensitivity to MK-801 antagonism, which is abolished by estrogen replacement therapy. In male mice, whether intact or castrated, estrogen administration does not alter the sensitivity of nonopioid SSIA to MK-801 antagonism (74). Such findings lead us to believe that female mice possess an estrogen-dependent SIA mechanism specific to their gender, of as yet unknown neurochemical identity. The demonstration of sexually dimorphic mechanisms of endogenous pain inhibition raises the possibility that novel analgesic drugs developed on the basis of single-sex studies will be ineffective in the opposite sex. At the very least, the discovery of important qualitative sex differences in analgesic mechanisms highlights the necessity to consider gender as a critical factor in both basic and clinical research.
Because the NMDA antagonist MK-801 blocks nonopioid analgesia, it may be that EAA agonists will have clinical utility as analgesics. For several reasons, however, it may prove difficult to develop pharmaceuticals based on this premise. First, there is a large literature documenting the neurotoxic effects of EAA agonists, including glutamate itself (see refs. 75,94 for reviews). In fact, MK-801 and other NMDA antagonists are currently undergoing clinical trials for the treatment of neurotoxicity associated with ischemic, hypoxic, and traumatic insult (120). Second, NMDA receptors have also been shown to play a role in ascending mechanisms of pain transmission. Following spinal administration, NMDA agonists produce nociception (1), and antagonists cause analgesia (21,87). Moreover, the NMDA receptor appears to mediate the phenomena of ``wind-up'' and ``central sensitization'' seen in the spinal cord dorsal horn after repetitive C-fiber activation or peripheral injury (29,31,117). Due to the ubiquitous nature of EAA receptors in the central nervous system (27), it is not at all surprising that NMDA receptors are involved in pain perception as well as pain inhibition. Thus, whereas the use of EAA receptor ligands as analgesics may not be currently feasible, the study of EAA involvement in mechanisms of pain and analgesia is likely to be crucial to a full understanding of these processes.
, four single daily injections of morphine (15 mg/kg, i.p.) administered
concurrently with MK-801 (0.05 mg/kg, i.p.) significantly attenuated tolerance
to a test dose of morphine on the 5th day (60).
In this same study, the broad-spectrum EAA antagonist kynurenic acid, which
blocks both NMDA and non-NMDA (kainate and AMPA/quisqualate) receptors,
actually enhanced morphine analgesia during the tolerance induction phase,
but still reduced tolerance to a test dose of morphine on the 5th day.
The phenomenon of tolerance may be at least partially due to associative factors, including state-dependent learning (59,99). Learning an association between environmental cues and the administration of morphine may result in the gradual development of antagonistic or compensatory responses to the anticipated effects of morphine, thus reducing the drug's effectiveness (7). It is conceivable, therefore, that the attenuation of morphine tolerance by MK-801 is due to its blocking associative learning; in fact, the NMDA receptor has been shown to be crucially involved in several forms of associative learning, including long-term potentiation (26) and Pavlovian fear conditioning (49).
We have recently reported that MK-801 can block opiate tolerance even when given 2 hours after each of the daily morphine injections during the 4-day tolerance induction period (61). By delivering the MK-801 after morphine analgesia has largely dissipated, the possibility that its effects can be attributed to state-dependent learning is reduced. In addition, we have most recently found that tolerance is blocked when morphine and MK-801 are given together once in sustained-release preparations (15). This paradigm is one claimed to minimize the opportunity for associative learning to occur (see ref. 110). Even when the test dose of morphine is given as many as 56 days later, much less tolerance is seen if the original dose of morphine was accompanied by MK-801. Thus, the NMDA receptor seems to play a significant role in nonassociative factors underlying the development of long-term opiate tolerance. In this same study (15), naloxone-precipitated withdrawal symptoms (teeth chattering, ptosis) were observed. If MK-801 had been given with morphine, however, teeth-chattering behavior was found to be attenuated, a finding that has been replicated and extended by another laboratory (40). The nonspecific EAA receptor antagonist kynurenic acid has also been demonstrated to attenuate morphine withdrawal symptoms (88).
In our study described above (15), we found that MK-801 potentiated morphine analgesia, as did kynurenic acid in an earlier experiment (60). Such findings offer the hope of developing drugs that can be safely used with opiates when given for pain control over many weeks or months (e.g., in the management of cancer pain). These drugs should not only prevent the development of opiate tolerance but may even enhance opiate analgesia while blocking tolerance, permitting still lower opiate doses to suffice.
Others have demonstrated the ability of cholecystokinin antagonists to prevent the development of opiate tolerance (32,46,82,103,113), and these antagonists as well are sometimes found to potentiate morphine analgesia (32,113). There have been reports attesting to the involvement of k opiate receptors (101,119), arginine vasopressin (118), norepinephrine (102), serotonin (36), and GABA (107) systems in tolerance development as well. Although investigators have been unable to document gross alterations in opioid receptor binding properties following chronic opiate administration, recent molecular evidence has suggested that tolerance may also be due in part to a ``functional decoupling'' of opioid receptors from their associated G proteins (see ref. 110 for review). As Trujillo and Akil (110) suggest, it is possible that this functional decoupling may be a result of an NMDA-mediated increase in intracellular calcium concentration, and in fact increases in intracellular calcium have been demonstrated in animals receiving chronic opiate administration (37,115).
There are a number of models available for use in studying the genetic determinants of a behavioral trait such as SIA. These models include the analysis of inbred strains, recombinant inbred strains, defined mutants, and selectively bred lines (14,98). In our laboratory we have been employing two such genetic models, on which we have conducted a series of studies on SPA, SIA, and morphine analgesia.
The most common approach to the genetic analysis of behavioral and drug-related traits is the characterization of inbred mouse strains (14,98). The use of such strains, however, involves the time-consuming process of breeding test crosses. A newer technique developed by Bailey (6) uses so-called recombinant-inbred (RI) lines, in which the F hybrid generation (F x F) from a cross between inbred strains (F) is ``fixed'' by repeated (>20 generations) sib matings. The first RI series, developed from a cross between C57BL/6 and BALB/c mice (the CXB series), has been extensively characterized with respect to opioid responses (14,34). Of all seven CXB RI strains, the extreme responders to a wide range of opioid phenomena are the CXBH (high) and CXBK (very low) lines, and these lines are now often studied in isolation. The CXBH and CXBK lines have been found to differ markedly in brain opiate receptor density, with CXBKs being lower and CXBHs higher than normal on this dimension (8). Likely due to this opiate receptor deficiency, CXBK mice have been found to display abnormally low morphine (8,77,80), acupuncture (85), and D-amino acid analgesia (24). Miczek and colleagues (70) have demonstrated that CXBKs display relatively low opioid-mediated conspecific ``defeat'' analgesia.
In our work with these RI strains, we have found that CXBK mice, while displaying normal levels of nonopioid SIA (4 min continuous footshock), do not display any opioid SIA (1 min continuous or 10 min intermittent footshock) whatsoever (78). We also have evidence that the SIA exhibited by CXBK mice following 3 min forced swimming in 20oC water, when present, is entirely of a nonopioid character and completely blocked by MK-801 (63,64). In addition, we have demonstrated that the threshold of SPA does not increase following naloxone pretreatment in CXBK mice as it does in the CXBH and the parental (C57BL/6 and BALB/c) strains (65). It can be concluded that, for SPA as well as for SIA, CXBK animals display only nonopioid analgesia.
Selective breeding, also known as artificial selection, is the most direct approach to studying the genetic determinants of a trait (33). Instead of examining serendipitous genetic rearrangements in inbred strains affecting the trait of interest (or, in the case of RI strains, producing and then examining them), by selective breeding one can specifically alter the frequency of those genes directly influencing that trait. It is important to note that genes are not being created, destroyed, or altered; the only effect of selection is to change the gene frequencies in a population (33,90). Beginning with Tolman's ``maze-bright/maze-dull'' rats (108), a number of artificial selection experiments in behavioral genetics have been successfully accomplished using rodents. Of specific relevance to analgesia are three selective breeding programs still in existence: the high/low autotomy rat lines of Devor and Raber (30); the high/low levorphanol analgesia mouse lines of Belknap et al. (13); and the high/low SSIA mouse lines of Panocka and colleagues (83).
We have been studying the lines of Panocka et al., which were bred specifically
for high (HA) and low (LA) SSIA at a water temperature (20oC)
and swim duration (3 min) producing a mixed opioid/nonopioid analgesia.
The selection process of Panocka and colleagues began with 150 animals
of the outbred Swiss-Webster line being tested for their analgesic response
on the hot-plate test (56oC), using an
arbitrary cut-off latency of 60 sec. Animals displaying postswim hot-plate
latencies shorter than 10 sec were chosen to be the progenitors of the
LA (``low analgesia'') line, and animals displaying latencies greater than
50 sec were chosen to be the progenitors of the HA (``high analgesia'')
line. LA males and females and HA males and females, so defined, were bred
in all successive generations using these same criteria. Figure
3
shows the progressing divergence of SSIA over the first 20 generations
of selection. After only five generations, post-swim hot-plate latencies
(measured 2 min after swim) in HA mice had increased by more than twofold
(83) over randomly bred controls (C).
We have found that HA and LA mice differ from one another whether the SIA is opioid or nonopioid (62). SPA threshold is five times higher in LA than HA mice (66); the analgesic dose of morphine 100 times greater (84). HA mice display four times more analgesia following ethanol administration than LA mice (73). These are large and reliable effects. These findings suggest that genetic factors are very important in determining central nervous system mechanisms of pain control, their overall strength and their neurochemical bases.
We have conducted another study in this series that appears to be particularly exciting (72). Following mendelian logic, we have crossed HAs and LAs to make F hybrids (HA x LA), and then F hybrids (F x F), and backcrossed the F hybrids to the parental stock to make high (HA x F) and low (LA x F) backcrosses. All of these groups were tested for opioid (32oC), nonopioid (15oC), and mixed opioid/nonopioid (20oC) SSIA, and also for morphine analgesia (10 mg/kg, i.p.). Following 32oC swim or morphine administration, the two opioid phenomena assessed, the proportions of animals in each genetically segregating group (i.e., backcrosses and F hybrids) exhibiting the HA, LA, and F hybrid phenotypes are close to the values predicted by mendelian analysis if the magnitude of analgesia is determined by one or very few gene(s).
The true power of the genetic approach is that it can allow the localization (i.e., gene mapping), isolation, and ultimately cloning of the gene(s) underlying inherited traits. The genetic dissection of a trait by gene mapping is facilitated by use of an appropriate genetic model. According to Lander and Botstein (51): ``The ideal situation occurs when (a) the phenotypic difference between the strains is large compared to the ... within-strain standard deviation; (b) breeding experiments indicate that the number of effective factors ... is small; and (c) the strains are a result of selective breeding for the trait.'' Clearly, the HA/LA mice fit these criteria perfectly. It seems likely, therefore, that the powerful techniques of molecular biology can be profitably applied to the task of gene mapping and cloning, representing the first steps toward the eventual use of gene therapy for chronic pain syndromes.
Working with rats, we have been studying the effects of stress on natural killer (NK) cell cytotoxicity and, in parallel, its effects on tumor growth. NK cells are a class of lymphocyte that, without prior sensitization, recognize and destroy virally infected cells and certain tumor cells, thus constituting a first line of defense against infectious diseases and playing a role of special importance in surveillance against cancer (39). We have used standard techniques and also developed new ones for assessing NK cell cytotoxicity in vitro and in vivo.
In our earliest work we found that intermittent footshock stress, which causes opioid SIA, suppressed NK cell cytotoxicity, whereas footshock of seemingly comparable severity (same intensity; same total shock duration, but applied continuously), which causes nonopioid SIA, did not (96). The suppressive effects of the opioid stress on NK cell activity were blocked by naloxone and mimicked by high (<20 mg/kg) doses of morphine given systemically (96). The same opioid stress, but not the nonopioid one, caused a higher rate of mortality in animals inoculated with a mammary ascites tumor (MAT13762B) (55). Among animals dying from the tumor, those given the opioid stress died sooner than those given the nonopioid stress or no stress. Again, an opiate antagonist blocked this effect (55), and a high dose of morphine mimicked it (54). Morphine's effect on NK cell activity is seen whether the cells are harvested from spleen, bone marrow, or blood, suggesting that this is a widespread effect visible in at least three major immune compartments (97). This suppressive effect of opiates on NK cells shows stereospecificity (i.e., levorphanol, but not dextrorphan, is active) (67). Furthermore, fentanyl and sufentanil (potent opiate drugs sometimes used clinically at high doses to induce opiate anesthesia) have effects similar to morphine in these assays at doses predictable from their analgesic potencies relative to morphine (12).
Morphine given intracerebroventricularly (i.c.v.) suppressed NK cell cytotoxicity at doses of 20 mg or higher (95), and although this dose is high relative to the i.c.v. analgesic dose, this observation nonetheless suggests a central site of action for morphine's NK-suppressive effect. This conclusion is further supported by the finding that N-methyl-morphine, which does not cross the blood-brain barrier, is ineffective in altering NK cell killing ability when given systemically (95).
Such findings suggest the frightening conclusion that high doses of opiates, such as those often required in the long-term management of cancer pain, for example, are themselves responsible for increased morbidity and mortality. However, we have also reported that morphine's immunosuppressive and tumor-enhancing effects show complete and rapid tolerance (54); after just a few days of morphine administration, its effects on NK cells and tumors were no longer seen. In marked contrast, the immunosuppressive effects of opioid-mediated footshock stress did not diminish even after three weeks of stress administration (54). These findings suggest that not only is it safe to use opiate drugs to manage pain over extended periods even at high doses, but that it may be unsafe not to, since failing to manage the pain might itself be stressful and hence immunosuppressive and tumor-enhancing in a manner that would not be expected to show tolerance (see ref. 56).
Our most recent work in this area employs the MADB106 tumor line both as the radiolabeled target cell in the in vitro NK assay and also as the tumor whose metastatic spread in the lungs we measure in vivo three weeks after intravenous inoculation. The MADB106 tumor is syngeneic to the Fischer 344 rats we study and its metastatic development is known to be NK sensitive (9,17). Several groups, including ours, have shown that reducing NK activity by antiasialo GM or by selectively depleting large granular lymphocyte (LGL)/NK cells with the monoclonal antibody 3.2.3 (17) causes a vast increase in the number of MADB106 pulmonary metastases, an effect that can be prevented by the adoptive transfer of LGL/NK cells, but not other lymphocytes (9).
We find (18) that forced swimming, of a sort that produces nonopioid SIA, suppresses NK cell cytotoxic activity measured 1 hr after stress and causes a pronounced increase in the number of MADB106 pulmonary metastases if the tumor is injected at the same time. These immunological effects of swim stress are also insensitive to naloxone (18). Clearly, not all stressors affecting the immune system do so via opioid mechanisms. If animals are stressed 24 hr after tumor inoculation, no effect on metastatic growth is seen (17). We know that the MADB106 tumor is sensitive to NK cell activity only for 24 hr postinoculation. Taken together, these results suggest that stress can enhance tumor metastasis by suppressing NK cell cytotoxicity.
To provide more direct evidence for this suggestion, we have studied lung clearance of MADB106 tumor cells, a process fundamental in preventing lung metastases of this tumor. There is recent evidence that this process is largely controlled by NK activity. For example, selective depletion of LGL/NK cells, using the monoclonal antibody 3.2.3, causes a 100-fold decrease in lung clearance of the MADB106 tumor (16) and a similar increase in number of lung metastases (17). Studying the effects of swim stress on lung clearance and NK mediation of such effects, we found a fourfold stress-induced decrease in lung clearance in normal rats but no such stress effect in rats previously depleted of LGL/NK cells (16). These findings strongly implicate LGL/NK cells in mediating the effects of forced swimming on lung clearance and metastatic colonization. Validating the ability of this depletion method to distinguish between LGL/NK-mediated and non-LGL/NK-mediated effects, we showed in the same experiment (16) that a different in vivo manipulation caused similar effects in both normal and LGL/NK-depleted rats. Although it is always difficult to prove causation, we feel our results offer the most convincing evidence to date of a causal relationship between stress, NK suppression, and enhanced metastatic development.
We commend to others this MADB106 model as one permitting parallel in vitro/in vivo studies of immunologic and oncologic outcome variables that are, at the very least, closely interrelated. Until now our technique has required sacrificing the animal in order to measure NK activity. However, we are currently using a whole blood NK cytotoxicity assay in which 0.5 ml of blood is sufficient to assess NK activity. In this assay, NK activity is measured per ml blood without excluding any blood cells. Original serum can serve as the assay medium, preserving some of the hormonal changes induced by an in vivo manipulation; or it can be replaced by neutral artificial medium, as is the standard practice. This method enables us to track NK activity over time in the same animal from successively drawn blood samples. We have found that the whole blood assay better reflects in vivo levels of NK activity than the standard NK assay.
Finally, in ongoing work from our laboratory, Dr. G. Page is investigating in rats the effects of an experimental surgery (laparotomy) on both NK cytotoxicity against the MADB106 tumor and on the number of lung metastases measured three weeks after injection of this same tumor (81). The surgery is performed under halothane anesthesia; animals are allowed to recover from the anesthesia and to survive for at least 10 hr. At that time, they are either sacrificed in order to perform the NK assay or are inoculated with the tumor and sacrificed three weeks later for determination of number of lung metastases. The surgical/postsurgical experience appears to be an adequate stressor for causing NK suppression and enhanced metastatic growth compared to untreated or anesthesia-only controls.
Chapter 3 Table 1: Effects of morphine on the
increased number of lung metastases of the MADB106 tumor induced by experimental
laparotomy in rats
|
|
|
||
| Vehicle | Morphine | Vehicle | Morphine |
| 42 (± 5)a | 40 (± 6) | 70 (± 9)b | 47 (± 8) |
Of special interest, Page is finding in this work that an analgesic dose of morphine delivered in a slow-release preparation during and immediately after the surgery is able largely to prevent these effects (see Table 1). It seems that the experience of pain, no doubt by triggering a robust stress reaction, can suppress NK cell activity and cause an experimental tumor to metastasize more rapidly (81). We believe these findings speak to the potentially critical importance of pain relief generally, and postoperative pain management specifically, for human health. Perhaps inadequately managed postsurgical pain will prove to be especially dangerous for patients with tumors that are in the metastatic phase of development. In sum, pain can apparently affect morbidity and mortality; contrary to popular belief, it seems that pain can kill (56).