A Single Intravenous Injection of KRN5500 (Antibiotic Spicamycin) Produces Long-Term Decreases in Multiple Sensory Hypersensitivities in Neuropathic Pain

Abstract

Neuropathic pain is a significant clinical problem. Currently, there are no drugs that produce complete amelioration of this type of pain. We have previously shown that KRN5500, a derivative of the antibiotic spicamycin, produces a prolonged (7-day), and significant reduction in neuropathic pain, but not nociceptive pain. Herein, we provide further evidence for the efficacy of this drug in inhibiting pain after IV injection in a spared nerve injury model of neuropathic pain. A single IV dose of the drug produces an increase in pain thresholds to punctuate mechanical stimuli and to cold stimuli over a period of 7 days, whereas IV injection of the vehicle is without any effect. No change in pain threshold was observed in the contralateral foot. In addition, a significant antiallodynic effect to mechanical stimuli was observed at 1, 2, 4, and 6 wk. The drug may be a potential candidate for cancer-related neuropathic pain as well as a marker for discovery of effective analgesics for neuropathic pain.

In Brief

IMPLICATIONS: We examined the effect of a novel drug (KRN5500) on nerve damage pain. After the successful effects of this drug in a single human, we have shown that the drug infused as a single application at different doses in a rat model of nerve damage pain produces pain relief in this model for many weeks.

Pain generated as a result of damage to the nervous system, or neuropathic pain, is characterized by spontaneous pain and an abnormal hypersensitivity to innocuous as well as noxious stimuli. Clinically, neuropathic pain is one of the most difficult types of pain to treat, and to date there is no effective treatment that can specifically control it when it has been established ⁽¹⁾. A major focus of current pain research has involved the use of behavioral models of neuropathic pain. Such models are thought to replicate, to different degrees, the pathophysiological changes seen in patients. Thus, they are useful both in the study of mechanisms underlying neuropathic pain and for testing the effectiveness of promising candidate analgesics. Behavioral neuropathic pain models have been surgically generated by complete or partial nerve transection and chemically generated, for example, by IV injection of the vinca alkaloid, vincristine ^(2–7). Incomplete lesions (surgical and chemical), in which there is only partial nerve injury, are most common clinically and involve tactile and thermal allodynia as well as pinprick hyperalgesia ^(8,9). Thus, animal models of partial nerve injury which include loose ligation/chronic constriction of the entire sciatic nerve (Bennett model), tight ligature of half the proximal sciatic nerve (Seltzer model), tight ligature of L5 and L6 spinal nerves (Chung model), and lesion of 2 of the 3 distal terminal branches of the sciatic nerve (spared nerve injury, SNI model), are most likely to model clinical neuropathic pain and provide a venue for the analysis of novel analgesics ^(3–6).

This laboratory’s efforts to identify and characterize novel analgesics have focused on KRN5500, a component of the nucleoside antibiotic spicamycin. KRN5500 was originally identified as a potent antitumorogenic drug ^(10,11). The potential analgesic properties of this compound were first observed in a patient who received the drug for the chemotherapeutic treatment of metastatic colonic cancer involving the liver (D. Borsook, unpublished data). The patient had had significant allodynia in a glove-and-stocking distribution in all extremities for 10 yr before developing colon cancer. He was selected to receive KRN5500 in an approved experimental therapy for his cancer. At the time of receiving the drug, he had liver metastases, producing nociceptive pain in the upper right abdomen. Upon receiving the drug, the pain in his hands and feet disappeared within an hour and did not recur until his death approximately 6 mo later (he did receive subsequent doses of the drug at 3.5-wk intervals). However, KRN5500 did not affect his abdominal pain (presumably nociceptive pain, produced by an expanding liver capsule).

In a previous study, we demonstrated that a single intraperitoneal injection of KRN5500 produced significant attenuation of mechanical allodynia in Chung and Bennett models of neuropathic pain ⁽¹²⁾. In contrast, KRN5500 had no effect on nociceptive pain induced by Complete Freund’s Adjuvant ⁽¹²⁾. These observations support the idea that KRN5500 can provide analgesia specifically for at least one measure of neuropathic pain. In this study, we sought to solidify this concept by testing the effectiveness of KRN5500 in a third model of neuropathic pain and by using three different measures of neuropathic pain, mechanical allodynia, cold allodynia, and pinprick hypersensitivity. After IV injection of KRN5500, behavioral tests for mechanical and cold allodynia as well as pinprick hypersensitivity demonstrated dramatic reduction of hypersensitivity. All behavioral responses were reduced nearly to presurgical baseline levels. These effects of KRN5500 were maintained, to different degrees, up to 7 days after injection. Effects of KRN5500 on mechanical allodynia also were followed up to 6 wk after injection. The results of the study, combined with our previous studies, provide strong evidence that KRN5500 has a dramatic effect on many behavioral measures of neuropathic pain, and that this effect is specific to neuropathic pain.

Methods

Male Sprague-Dawley rats (150–200 g; Charles River Laboratories, Wilmington, MA), housed in groups of 4 under a 12 h light/dark cycle, were used in these experiments. Food and water were provided ad libitum. Experimental protocols were approved by the Institutional Animal Care and Use Committee at Massachusetts General Hospital, Boston.

Neuropathic Pain Model—SNI

One week after acclimatization to laboratory conditions, baseline behavioral measurements were made and surgery was performed as previously described by Decosterd and Woolf ⁽⁶⁾. Briefly, rats were anesthetized with halothane in oxygen. The sciatic nerve and its three main branches: sural, common peroneal, and tibial nerves were exposed in the left thigh. The SNI procedure then consisted of the cutting and ligation of common peroneal and tibial nerves. After ligation, 2–4 mm of the distal nerve stumps was removed. Throughout the procedure, particular care was taken to prevent contact with or stretching of the sural nerve which remained intact. After wound closure, animals recovered from anesthesia within approximately 10 min. At 1 wk postsurgery, animals were divided into 3 groups and tested for signs of mechanical allodynia, cold allodynia, or mechanical pinprick hyper- sensitivity.

Behavioral Testing

Mechanical Allodynia.

Each animal was placed on an elevated wire grid and covered by a transparent plastic box open both at the bottom and along one side. Calibrated von Frey filaments (VFF) were applied to the lateral plantar skin of the hindpaw and each VFF was applied until the filament just began to bend ⁽¹³⁾. A trial consisted of 4 VFF applications at a frequency of 1 per 10–15 s. Threshold was taken as the least force that evoked a brisk withdrawal response ⁽¹⁴⁾. An upper cut-off of up to 15 g of force was used in these experiments because more force VFF can produce a painful response in normal animals. This cut-off was not used when testing for general loss of sensitivity after drug treatment of the uninjured foot. In this case, the maximal force tolerated, 28 g, was used.

Cold Allodynia.

While animals were on the elevated grid, a blunt needle attached to a syringe was used to gently apply a drop of acetone to the lateral plantar skin of the paw. Cooling was produced by evaporation of the acetone ⁽¹⁵⁾. The skin was not touched by the needle. The duration of paw withdrawal was recorded with an arbitrary maximal cut-off of 20 s ⁽¹⁶⁾.

Mechanical Pinprick Hyperalgesia.

While on the elevated grids, animals also were tested for mechanical hyperalgesia using a safety pin applied to the lateral surface of the left hindpaw. The pin tip was applied with enough pressure to indent but not penetrate the skin. The length of time for which the paw was subsequently withdrawn was recorded and arbitrarily cut-off at a maximum of 10 s ⁽¹⁷⁾.

Drug Administration

Throughout the procedure, rats were anesthetized with halothane in oxygen. To visualize the tail vein, the tail was held in 50°C water for 10–15 s and then wiped with an alcohol prep pad. An Angiocath 24-gauge catheter was then inserted into the vein and the needle was removed. Drug, saline, or placebo was then infused over a 1-min period using a 1-mL syringe. The catheter was then flushed with saline and removed.

The data were presented as mean ± sem. Statistical differences between the groups were determined by the Mann-Whitney U-test. Dunnett’s test after analysis of variance was used to assess the difference between baseline and each time point within the groups. Statistical significance was accepted when P ≤ 0.05.

Results

Effect of IV KRN5500 on Mechanical Allodynia

All rats displayed normal presurgical responses to innocuous mechanical stimulation as measured using VFF. At 7 days postsurgery, the earliest time point taken, animals displayed only minor allodynic responses to VFF stimulation. Within 10–14 days of SNI surgery, essentially all animals exhibited marked mechanical allodynia on the lateral surface (sural nerve skin area) of the affected hindpaw. The force required to elicit foot flexion with VFF decreased from 15 g (most force used in this study) down to <1 g 14 days postsurgery (Fig. 1A). IV tail injection of 1 mg/kg KRN5500 resulted in recovery to near baseline levels of mechanical sensitivity within 4 h and this effect was sustained through postinjection day 1 (Fig. 1A). At 3, 5, and 7 days postinjection, animals showed some increase in hypersensitivity to VFF stimulation but were still significantly improved compared with the postsurgical baseline. Rats that received tail injections of saline retained postsurgical baseline levels of mechanical allodynia (hypersensitivity) throughout the 7-day period (Fig. 1A).

Figure 1: A, Effect of KRN5500 on mechanical allodynia for 7 days after a single IV injection. Time course of changes in mechanical allodynia (pain threshold in grams). Various forces of von Frey filaments (up-down method) were applied to the plantar surface of the left paw before and after spared nerve injury. Fourteen days postsurgery, saline (n = 12) or KRN5500 (n = 10) (1.0 mg/kg) was administered IV, and the rats were studied for 7 days. B, Effect of KRN5500 on mechanical allodynia (von Frey hair) over a 6-wk period. The histograms show the time course of changes in mechanical allodynia (pain threshold in grams) after a single IV injection of KRN5500. Various forces of von Frey filaments (up-down method) were applied to the plantar surface of the left paw before and after spared nerve injury. Fourteen days postsurgery, saline (n = 5) or KRN5500 (n = 5) (2.5 mg/kg) was administered IV, and the rats were studied for 6 wk. All of the values represent the mean ± sem within the group. *Value is significantly different from baseline (S) 7–10 days postsurgery (P < 0.05, Dunnett’s test after analysis of variance). #P < 0.05 between groups (Mann-Whitney U-test). BL = preoperative baseline, Surg. = baseline value 14 days after spared nerve injury surgery.

Time Course of the Effect of KRN5500 on Mechanical Allodynia in SNI Model

To gain a better understanding of the long-term effects of a single injection of KRN5500 on neuropathic sensory hypersensitivity, animals received a large (2.5 mg/kg) dose of KRN5500 and were tested with VFF up to 6 wk postinjection. With this large dose, animals recovered normal presurgical responses to VFF within 4 h of IV injection (Fig. 1B). From 1 to 7 days postinjection, there was a steady decline in effect; however, at 7 days, the mechanical threshold to VFF stimulation was still 50% of the presurgical baseline—a substantial improvement compared with the mechanical threshold of saline control animals (Fig. 1B). Two, 4, and 6 wk after a 2.5-mg/kg injection of KRN5500, mechanical threshold as measured by VFF had substantially decreased but was still more than that measured in animals that had received IV injections of saline (Fig. 1B).

To confirm that the recovery of more normal mechanical sensitivity is specific to the KRN5500 component of the drug solution, a series of SNI animals were injected either with saline or the vehicle in which KRN5500 was dissolved. From 4 h to 7 days postinjection, both groups displayed profound mechanical allodynia. There appeared to be no difference in the hypersensitivity exhibited by saline- or vehicle-injected rats, confirming that there was no effect of the vehicle (Fig. 2A). To further demonstrate the specificity of KRN5500 and determine that the reduction in mechanical hypersensitivity was not caused by a general decrease in sensitivity, VFF were used to test the uninjured hindpaw of SNI rats that received either 1 mg/kg KRN5500 or saline. As seen in Figure 2B, the force required to elicit foot flexion did not increase over the 7-day postinjection period, which indicates that there was no generalized loss of sensitivity after drug injection at this dose. The only observable change was an extremely slight increase in sensitivity on day 3.

Figure 2: A, Effects of KRN5500 vehicle on neuropathic pain behavior. Time course of changes in mechanical allodynia (pain threshold in grams). Various forces of von Frey filaments (up-down method) were applied to the plantar surface of the left paw before and after spared nerve injury. Fourteen days postsurgery, saline (n = 4) or KRN5500 (n = 4) vehicle was administered IV, and the rats were studied for 7 days. B, Effects of KRN5500 on uninjured hindpaw to mechanical withdrawal in rats with contralateral spared nerve injury exhibiting allodynia. Time course of changes in mechanical allodynia (pain threshold in grams). Various forces of von Frey filaments (up-down method) were applied to the plantar surface of the left paw before and after spared nerve injury. Fourteen days postsurgery, saline (n = 5) or KRN5500 (n = 4) (1.0 mg/kg) was administered IV, and the rats were studied for 7 days. All of the values represent the mean ± sem within the group. BL = preoperative baseline, Surg. = baseline value 14 days after spared nerve injury surgery.

Effect of IV KRN5500 on Cold Allodynia

Decosterd and Woolf ⁽⁶⁾ demonstrated that in addition to a profound and persistent mechanical allodynia, SNI rats also develop a strong and persistent hypersensitivity to cold, as demonstrated by prolonged hindpaw flexion withdrawal in response to a drop of cold acetone. Similarly to the Decosterd and Woolf study ⁽⁶⁾, this study shows that a cold stimulus applied to the lateral plantar hindpaw elicits no withdrawal during presurgical testing (Fig. 3). As early as 7 days postsurgery, animals displayed a clear allodynic response to cold acetone (data not shown). By 14 days postsurgery, a cold stimulus applied to the lateral plantar hindpaw evoked prolonged duration of flexion withdrawal (6–9 s) as compared with responses obtained in the presurgical period (Fig. 3). Four hours after IV injection of 1 mg/kg KRN5500, there was a dramatic shortening of flexion withdrawal duration (1–2 s) in response to cold acetone, indicating a loss of hypersensitivity with drug treatment (Fig. 3). Over the duration of the post-IV testing period (7 days), the duration of withdrawal did increase but remained substantially reduced as compared with the duration of flexion withdrawal measured in animals that received saline injection as a control (Fig. 3).

Figure 3: Effect of KRN5500 on cold allodynia. Time course of changes in cold allodynia (duration of flexion withdrawal in seconds). A cold stimulus (acetone drop, see Methods) was applied to the plantar surface of the left paw before and after spared nerve injury. Fourteen days postsurgery, saline (n = 10) or KRN5500 (n = 9) (1.0 mg/kg) was administered IV, and the rats were studied for 7 days. All of the values represent the mean ± sem within the group. *Value is significantly different from baseline (S) 7–10 days postsurgery (P < 0.05, Dunnett’s test after analysis of variance). #P < 0.05 between groups (Mann-Whitney U-test). BL = preoperative baseline, Surg. = baseline value 14 days after spared nerve injury surgery.

Effect of IV KRN5500 on Mechanical Pinprick Hyperalgesia

To further characterize the effects of KRN5500 on sensory hypersensitivity that result from nerve injury, the duration of paw withdrawal after pinprick was measured in SNI animals that had received IV injection of either KRN5500 or saline. Baseline (presurgery) withdrawal duration was always very short, <1 s (Fig. 4). However, after transection of the tibial and saphenous nerves, the duration of flexion withdrawal to pinprick of the lateral plantar hindpaw surface (sural nerve skin area) was increased to 5 ± 1 s and 7.5 ± 1.5 s in the 2 treatment groups by 14 wk postsurgery (Fig. 4). Mechanical pinprick hyperalgesia was clearly established along a time course similar to mechanical and cold allodynia. Like mechanical and cold allodynia, pinprick hyperalgesia was dramatically reduced at 4 h after injection of KRN5500. Unlike mechanical and cold allodynia, the pinprick response steadily increased over the 7-day period of testing. Nonetheless, there still was approximately a 50% reduction in pinprick hyperalgesia 7 days after a single injection of 1 mg/kg KRN5500 (Fig. 4).

Figure 4: Effect of KRN5500 on pin-prick hyperalgesia. Time course of changes in mechanical pin-prick hyperalgesia (duration of flexion withdrawal in seconds). A pin prick was applied to the plantar surface of the left paw before and after spared nerve injury. Fourteen days postsurgery, saline (n = 12) or KRN5500 (n = 10) (1.0 mg/kg) was administered IV, and the rats were studied for 7 days. *Value is significantly different from baseline (S) 7–10 days postsurgery (P < 0.05, Dunnett’s test after analysis of variance). #P < 0.05 between groups (Mann-Whitney U-test). All of the values represent the mean ± sem within the group. BL = preoperative baseline, Surg. = baseline value 14 days after spared nerve injury surgery.

Effect of IV KRN5500 on Motor Function and Weight Gain

Finally, the effect of different doses of KRN5500 was tested on motor function using an accelerating Rota-rod treadmill and on weight gain. Because SNI affects function of the affected hindpaw, Rota-rod testing was done using surgically naïve animals. After Rota-rod training and baseline measurements, animals received IV injection of saline, 2.5, or 1.0 mg/kg KRN5500. When compared with baseline measurements, there was no significant difference in performance on the Rota-rod in any of these groups over a 7-day period (Fig. 5). To further assess the specificity of KRN5500, rats were weighed over a 6-wk period after injection of saline, 0.5, 1.0, or 2.5 mg/kg KRN5500. All groups showed a steady and consistent weight gain over the 6-wk period. At 3 and 5 days postinjection, the 2.5-mg/kg animals trailed behind the saline group (2% and 6% increases versus 7% and 7.5% increases) in weight gain but by 7 days both groups exhibited an 11% weight gain. Animals that received 1- and 2.5-mg/kg injections exhibited larger weight gains than saline over the time period tested.

Figure 5: Effect of KRN5500 on motor function. Time course of changes in motor function (time on wheel in seconds). Before and after IV injection, the amount of time the rat could stay on the accelerating Rota-rod was recorded. Three days after being trained on the Rota-rod, saline (n = 4), KRN5500 (n = 15) (1.0 mg/kg), or KRN5500 (n = 6) (2.5 mg/kg) was administered IV, and the rats were studied for 7 days. All of the values represent the mean ± sem within the group. BL = preinjection baseline.

Discussion

Spicamycin, a nucleoside antibiotic obtained from Streptomyces alanosinicus, is a mixture of several fatty acid components that have antitumor activity in vitro and in vivo^(10,11). In our studies of neuropathic pain, we have used a semisynthetic derivative of spicamycin called KRN5500. This derivative has been shown to have high chemotherapeutic efficacy (at 12 mg/kg) in animal models of hepatic metastasis ⁽¹¹⁾. KRN5500 also is the derivative that provided prolonged relief of neuropathic pain to our patient who took the drug for metastatic colon cancer (D. Borsook, unpublished data). At doses larger than those used in our studies, KRN5500 shows a broad spectrum of antitumor activity and has potent inhibitory activity against protein synthesis in in vitro models ^(11,18).

In a previous study, we have shown that a single IV injection of KRN5500 can attenuate mechanical allodynia in both the Chung and Bennett models of neuropathic pain ⁽¹²⁾. These models are well-established experimental models of neuropathic pain resulting from partial nerve injury ^(3,5). However, clinical neuropathic pain is produced by multiple etiological factors that inevitably result in pain attributed to different underlying mechanisms. Therefore, to best characterize the clinical relevance of a novel analgesic, it should be tested in multiple animal models resulting from different types of nerve injury. The SNI model is perhaps ideal to further our understanding of the clinical potential of KRN5500 because it involves a very different type of nerve injury compared with the Chung and Bennett models. In the SNI model, two terminal branches of the sciatic nerve (tibial and common peroneal) are cut and the sural nerve is left intact ⁽⁶⁾. The Chung and Bennett models involve more proximal (before the branch point) partial injuries and there is co-mingling of distal intact and degenerating axons, whereas in the SNI model, this is greatly restricted. As a result, the SNI model may represent a different pathophysiological mechanism and thereby broaden our understanding of the clinical potential of KRN5500.

In this study, all SNI animals began to show behavioral responses as early as seven days after injury. In contrast, Decosterd and Woolf ⁽⁶⁾ have reported effects on pain behavior in the first 24 hours postsurgery. However, we did obtain robust behavioral responses, consistent with those reported by Decosterd and Woolf within 10–14 days of surgery. These behavioral effects were highly reproducible and were measured in almost all animals that received surgery. Subsequent to establishment of mechanical allodynia, cold allodynia, and pinprick hyperalgesia, animals received a single injection (1 mg/kg) of KRN5500. Injection of KRN5500 IV resulted in nearly complete recovery of presurgical responses in each behavioral test of pain within one day of injection. Up to 7 days after a single injection, there was still significant attenuation of all pain behaviors. Moreover, a single injection of 2.5 mg/kg KRN5500 was even able to maintain a small level of analgesia as long as 6 weeks after injection. Remarkably, KRN5500 seems to have both rapid (within 24 hours) and prolonged (days to weeks) effects on behavioral measures of neuropathic pain.

Our studies do not identify mechanisms underlying the analgesic properties of KRN5500. However, some biologically relevant effects of this drug in other systems have begun to emerge and may shed light on our understanding of its role in analgesia. KRN5500 and its active metabolite, SAN-Gly, inhibit protein synthesis in vitro^(11,18). In addition, Kamishohara et al. ⁽¹⁹⁾ found that treatment with KRN5500 leads to swelling of the Golgi apparatus and alters the distribution of a newly synthesized marker protein in a colon adenocarcinoma cell line. Compounds that block axonal transport (e.g., colchicine, vincristine) can reduce hyperalgesia in animal models of pain ^(20,21). Quite interestingly, an increasing body of data shows that nerve injury results in the new synthesis and redistribution of proteins that are thought to have a role in the development and maintenance of neuropathic pain [reviewed in Woolf and Mannion ⁽¹⁾ and Woolf and Salter ⁽²²⁾]. Thus, any effects of KRN5500 on new protein synthesis and distribution/transport are candidate mechanisms in analgesia. In addition, Sakai et al. ⁽²³⁾ have demonstrated that the antitumor activity of spicamycin analogs depends on the presence of a glycine moiety. Given that neuronal hyperexcitability, both at the level of primary sensory neurons and at the spinal cord (central sensitization), is important to pain hypersensitivity, it is interesting to speculate that this glycine moiety may have a functional, inhibitory role in sensory pathways [reviewed in Woolf and Mannion ⁽¹⁾ and Woolf and Salter ⁽²²⁾]. Preliminary electrophysiological data from our laboratory suggests that KRN5500 can rapidly modulate neuronal excitability in spinal cord slices ⁽²⁴⁾, suggesting a rapid inhibitory neurotransmitter-type modulation. Our studies show that KRN5500 has a profound and long-lasting analgesic effect on three behavioral pain tests of neuropathic pain. Studies to determine the underlying mechanisms of action are still required.