Ok, here it is. It was another disappoint in my quest to help a friend with "ALS".
MUSCLE & NERVE 16:624-633 1993
Recombinant Growth Hormone Treatment of Amyotrophic Lateral Sclerosis
R.A. Smith, MD; S. Melmed, MD; B. Sherman, MD;
J. Frane, PhD; T.L. Munsat, MD; and B.W. Festoff, MD
From the Center for Neurologic Study, San Diego, California (Dr. Smith);
Division of Endocrinology, Department of Medicine, Cedar Sinai/UCLA Medical Centers, Los Angeles, California (Dr. Melmed);
Genentech, Inc., S. San Francisco, California (Drs. Sherman and Frane);
Neuromuscular Research Unit, Department of Neurology, Tufts-New England Medical Center, Boston, Massachusetts (Dr. Munsat);
and Neurobiology Research Laboratory, Department of Neurology, Veteran Affairs, and University of Kansas Medical Centers, Kansas City, Missouri and Kansas City, Kansas (Dr. Festoff)
--------------------------------------------------------------------------------
Amyotrophic lateral sclerosis (ALS) is a fatal disorder of man characterized by muscular weakness and fasciculations, progressive atrophy, and abnormal muscle stretch reflexes. As the disease inexorably progresses, there is an effort at regeneration and reinnervation manifested by collateral sprouting of still intact peripheral axons which reinnervate muscle fibers whose own neural connection have been lost. This process may be under humoral control.
The idea that an abnormality, or deficiency, of a putative motoneuron growth/trophic factor (MNGF) might underlie the cause(s) of ALS, or relate to its pathogenesis, has been inferred from the example of nerve growth factor (NGF). NGF exerts its trophic functions principally on automatic and primary sensory neurons, but is also implicated in CNS cortical and subcortical cholinergic systems. Although a MNGF has not been definitely identified, a number of candidates have been suggested. Other factors, among them insulin-like growth factors I and II (IGF-I, IGF-II) and epidermal growth factor (EGF), have all been demonstrated to exert trophic influences on CNS neurons.
IGF-I, also known as somatomedin C, is of particular interest, since it is a growth factor which shares structural and receptor similarities with insulin and is the principal mediator for growth hormone (GH). Along with insulin, IGF-I and IGF-II specifically bind throughout mammalian brain. The greatest amount of specific binding is found in cortical grey areas, in particular, the olfactory lobe, whereas binding is lowest in white matter fiber tracts. This topography suggests that IGF-I may primarily influence neuronal subsets. In cell culture it is also a potent inducer of oligodendrocyte development. In the spinal cord, cervical-thoracic ventral horns show maximum IGF-I binding. In the periphery, there is evidence to support roles for IGF-I in both muscle and nerve. Relevant to the reinnervation capacity in ALS muscle, IGF-I mRNA increases in muscle after denervation or paralysis, and direct injection of IGF-I into adult muscle induces intra-muscular nerve sprouting. Further, GH appears to regulate IGF-I mRNA in muscle.
Growth hormone has also been shown to have effects on myelination. Newborn rats rendered GH-deficient exhibit marked reduction of myelination, while mice that are genetically deficient in GH myelinate poorly. With exogenous administration of GH, these myelination deficits are reversed.
Although GH release in ALS patients is apparently normal, gthe above studies suggested the rationale that the pathophysiologic events which characterize ALSmight respond to growth hormone. We considered it likely that, at the minimum, exogenous administration of human growth hormone (hGH) in ALS patients would affect nitrogen balance as it does in fasting, producing a "protein sparing effect". In subjects administered a hypocaloric diet, GH has the effect of increasing protein synthesis and attenuating the efflux of amino acid nitrogen which follows carbohydrate loading.
More importantly, hGH might influence or prevent the loss of alpha motoneurons, secondary demyelination, or promote reinnervation of muscle. These hypothetical benefits might result from direct or indirect influences of hGH, or IGF-I, on one or more components of the motor system.
This report summarizes the results of a double-blind, placebo-controlled treatment trial. An account of our research methodology and treatment strategy has been reported.
METHODS
Inclusion and Exclusion Criteria. Ambulatory patients with mild-moderate "classical ALS," as defined by clinical and electromyographic criteria as previously established, were entered into the trial in four cohorts at the three centers (Table 1). We excluded all "pure" syndromes such as progressive musclular atrophy (PMA), progressive bulbar palsy (PBP), and primary lateral sclerosis (PLS).
Drug and Placebo. Human recombinant, methionyl growth hormone (5 mg) formulated with 40 mg mannitol, 0.54 mg sodium, and 1.16 mg phosphate was reconstituted in bacteriostatic water and administered intramuscularly at a dose of 0.1 mg/kg, three times a week. Placebo consisted of the same formulation without recombinant human growth hormone (rhGH). No change in dose was made regardless of weight loss or gain. All drugs were refrigerated upon arrival at each center's pharmacy and at each patient's residence. The manner of administration of rhGH or placebo was carefully reviewed at the outset of the study. Compliance with the treatment regimen was recorded monthly, at which time used vials were returned to the trial coordinator.
This was a double-blind, placebo-controlled, parallel therapeutic trial which did not call for crossover during the study. Randomization was performed by use of computer-based tables of random numbers and was balanced for each site. This was done to ensure that an approximately equal number of treated and placebo patients were enrolled at each center. Chronological entry into the study determined the patient's identification number which was already predetermined by the randomization procedure.
Tufts Quantitative Neuromuscular Examination (TQNE). A modification of the TQNE (Tufts New England Medical Center, Boston, MA) as reported by Andres et al. was used for this trial. The isometric force measurements and manual dexterity (speed) elements were employed as described, but a hand-held microprocessor controlled spirometer (Respiradyne, Pulmonary Function Monitor Manufacturer, Chesebrough-Pond Inc., Hospital Products Division, Greenwich, CT 06830) was utilized for pulmonay function parameters. A single, trained registered physical therapist made all TQNE measureements at the three centers. This individual was trained at Tufts by Dr. T. Munsat and P. Andres, jRPT. Identical equipment was placed at each center and calibrated by the RPT prior to each measurement. A TQNE was performed on entry and every 2 months during the study.
The TQNE raw data was converted to z-scores for standardization and grouped into five megascores as outlined previously. Mega 1 represents pulmonary function; Mega 2, bulabar function; Mega 3, timed hand activities; Mega 4, isometric arm strength; and Mega 5, isometric leg stregth. The TQNE has been used recently to document the "natural histroy" of motoneuron loss in 50 ALS patients followed for as long as 67 months. In general, deterioration of motor neuron function is linear in ALS when considering the combined Mega score, although some muscle groups, for example, knee and elbow flexors, demonstrate a more linear change than others. A preliminary report validating the test/retest confidence of TQNE has been presented previously (presented at the 113th annual meeting of the American Neurological Assocaiation, Philadelphia, PA, Octovber 2-5, 1988).
Neurologist's Neuromuscular Exams. A modified Medical Research Coucil (MRC) scoring system was used. These measurements were performed every 2-3 months during the course of study. Using MRC scoring, muscle stregth was graded on an ordial scale of 1-4 for the following muscle groups: neck flexors and extensors, deltoids, bilcepts, brachiooradialis, triceps, wrist extensors, iliopsoas, quadriceps, hamstrings, and anterior tibialis.
Laboratory Tests. Routine chemistries, coagulation assays, and fasting blood sugar (FBS) were obtained at entry and every month during the study and peformed in a central laboratory (SK Beckman, Los angeles, CA). Whole blood was collected in EDTA-treated tubes, the plasma separated from cells, aliquoted, and frozen at -20 degrees Celsius with 4 hours of collection. Total somatomedin C (IGF-I), insulin, and anti-hGH antibody levels were obtained at entry and every 2 months. These were determined using commercial kits as described (Nicholls Institute, San Juan Capistrano, CA). Similarly, 24-hour urine creatinine and 3-methylhistidine (3-MH) levels, obtained on the same schedule, were measured from 30-ml frozen aliquots using previously reporeted methods in a central laboratory (Calcium Research Lab, KCVAMC, Kansas City, MO).
Statistics. A double-boind placebo controlled trial was underteaken to minimze bias. Seventy-five patients were recruited so as to insure a poswer of at least 95% for a single one-sided test at the 0.05 level when the difference of the means of the treated and placebo groups was equal to 1 standard deviation for any variable.
RESULTS
Patient Entry. Four cohorts (total of 75 patients) were entered into the trial relatively evenly amongst the three centers (Table 1), beginning in March 1987. The oldest patient was 74, the youngest 20. The mean age was 57.18 +/- 10.68 (SD). A positive family history for AlS was elicited in 6 patients (8.0%). Two patients knew nothing about their families. Considering those who commenced treatment, there were 41 men and 34 women, approaching the expected sex ratio as reported for AlSS in the literature. Forty-one patients completed the 1-year study, 11 died, and the remainder (n - 23) withdrew for a variety of reasons. It was agreed at the outset that preliminary analysis of the data would be undertaken midway in the study to determine if a positive trend existed. If such a trend was found, then the study would have ben declared open and alll patients would have been provided growth hormone for an additional 12 months.
Unfortunately, such a trend was not found at the 6-month point. Accordingly, the study was continued as planned. The last cohort completed their 12 months of therapy. Consequently, some patients recieved therapy for as long as 18 months.
TQNE Reliability Analysis. Reliability data, using a single rater at all three centers, was strikingly good at 0.97-0.98 for all megascoresusing the TQNE. These results (Fig. 1), using identical equipment all calibrated by the same RPT, were equal to or better than those previously reported at a single center and have been communicated in preliminary form. Factor analysis of individual items demonstrated clustering of spediific megascores to related neuraxis regions (i.e., bulbar and arm strength).
Nutritional Data. Since growth hormone might conceivably affect the nutritional status of AlS patients, we reviewed the amount of weight loss that occured over time in the control and treatment groups. In the control group, the mean weight was approximately 155 lbs at entry versus 152 lbs. for the treatment group. At 24 weeks, chosen to attain a larger sample size, mean weight dropped 5.8 lbs. in the control group compared with the treatment group which lost 4.7 lbs. This difference did not reach significance (Table 2). HOwever, analysis was confounded by the fact that some patients developed edema. This prompted the use of antidiurietic therapy in several instances. A total of 33 of the combined group of patients lost weight, several losing 15% of their initial weight, and 1 almost 22% over a 6-month period.
Rate of Disease Progrssion. As an estimate of disease progression weutilized a modified TQNE consisting of 28 components. This allowed for interval data on functional motor units to be quntitaatively compared between the two treatment aarms. Test/retest data for all patients at entry indicated a high degree of reliability for all megascores (r=0.96-0.98). For all megascores (hand, arm, leg, and speech) there was a downward trned between the placebo and hGH-treated groups (Fig. 2). The decrease from visit to visit averaged 0.122 for the placebo patients and 0.149 for the protropin group. This difference was not significant. Because of the fact that losses occurred as the trial proceeded, some caution is warranted in interpreing the data. The implication is that patients who died or dropped out would have demonstrated similar decreases in their TQNE scores had they remained in the study.
Survival Analysis. At the completion of the 48-week study, 6 placebo and 5 rhGH-treated patients had died. Follow-up data was obtained on as many patients as possible approximately 12 months afeter termination of treatment in order to maximize the statistical power for the survival analysis. At this juncture, 19 of 38 growth hormone-treated patients had died as had 16 of 36 placebo-treated patients. The difference in the proportion dying in the two groups was not statistically significant (P=0.65 using Fisher's Exact Test).
A Cox-model survival regression analysis was performed to evaluate the effects of growth hormone treatment and prognostic baseline variables including age, and TQNE megascores for arm strength, leg strength, hand strength, and speech on survival time. Theere were 62 patients with data available for this analysis. In this analysis, only leg strength was found to be statisticallly significnat (P=0.044), with poorer prognosis for patients with less strength on entry into the study.
A seperate Cox-model survival regression analysis was performed, including all of the above variables plus the rates of decline in the TQNE megascores. Data from 56 patients were available for this analysis. In this analysis, each of the rates of decline in the TTQNE megascorres were individually statistically significant:
Rate of Decline P-value
Hand 0.0047
Speech 0.0046
Leg Strength 0.0002
Arm Strength <0.0001
Using all of the abbove variables jointly (baseline values and rates of decline) in Cox multiple regression, only age (P=0.0062) and the rate of decline in arm strength were statistically significant (P<0.0001). In this last analysis, older patients and patients experiencing the most rapid decline in arm strength had the worst prognosis. The relationship of decline in arm strength to survival is illustrated in Figure 3.
Comparison of MRC with TQNE. Comparing the average MRC scores with the average arm and leg megascores from the TQNE a strong correlation was noted. However, at the lowest MRC values, the variability in the TQNE limb megascores was greater than at higher MRC values (Fig. 4).
Chemistries and Coaagulation Profile. No significant abnormalities were detected in routine blood chemistries or hemograms. There was a slight prolongation of prothrombin time with a mean of approximately 2 seconds. This was statistically significant and merits further evaluation. Bleeding and/or clotting disorders have not been described as being overrepresented in the ALS patient population, however, a recnet report suggested an inverse assocaition between ALS and atherosclerosis.
Hormonal Data. It has been reported (after the initiation of this study) that total plasma IGF-I levels in ALS patients are normal. We analyzed total IGF-I in most patients entered iinto the trial along with fasting blood sugars and insulin leveles. The minmal total somatomedin C (IGF-I) detectable level was 2.3 mU/mL. After correction for dilution (1:20) the results were reported in aunits/mL. Normal values for controls, obtained by sampling a population of 220 adults (Nichols Institute normal data) were 0.89 U/mL (range 0.34-2.2). Approximately 25J% (18 of 69) of ALS patients' baseline values were abnormal-10 being lower and 8 being higher than expected. The loweest values were not restricted to older males. At baselikne, 8 of the placebo-treated patients were abnormal. On serial sampling over monthhs two of these values remained lower than our nomrla and the value recently rreported by Braunstein and Reviczky for ALS patients. More striking are the results of the treated ALS patients' responses to rhGH. Over the course of the study, total average IGF-I serum levels almost doubled. In the control group IGF-I went from 0.9 U/mL, whereas in the protropin gorup the mean went from 1.2 U/mL to 2.3 Y/mL. This is the first study demonstrationg the intactness of the IGF-I response to exogenous rhGH response in ALS patientrs (Fig. 5), and, in fact, in an adult population.
Shown in Table 4 are the blood glucose and insulin values. We had anticipated that most patieents wuold develop elevated blood glucose, and this was found, but only two developed frank hyperglycemia (over 149 mg/dL). In contrast to the IGF-I response, and unexpectedly, no chagne was found in insulin leveles after prolonged administration of rhGH. This was paradoxical given the glucose and IGF-I responses, and clearly different from non-ALS adult subjects administered protropin.
3-MH Excretion. Analysis of the excretoin of 3-MH and the creatinine: 3-MH ratios failed to show a difference between rhGH-and placebo-treated ALS patients.
DISCUSSION
This trial sought to determine whether exogenous administration of recombinant human growht hormone could modify or arrest the course of amyotrophic lateral sclerosis or provide ancillary benefits in the form of improved nutrition. At the comletion of the trial an almost eqaula number of placebo and treated patients had succumbed to their disease. This sugests that rhGH did not appear to influence the course of ALS as judged by careful measurement of disease progression using the TQNE.
More treated versus control patients dropped from the study. This did not apprear to be related to side effects which were not noteworthy except in two instances when blood sugar was affected. Withdrawal from ALS trilas is commonplace. Olarte et al. reported that 24 of 55 patients left a 1-year trial. Accordingly, our experience is not unique, suggesting between both studies that withdrawal rates of 30-40% are to be expected in ALS trials. This may be an indictment of the trial process, but most likely represents the fact that ALS patients become discouraged when their disease progresses inexorably in spite of treatment. THere are other reasons as well. A few patients switched to other treatments based on sensational reports in the "ALS media." At least one patient was told she did not have ALS and would benefit from immunosuppressive therapy. The basis for these conclusions was not apparent, but does illustrate the influences that nonparticipating physicians can exert upon subject participation.
Since abnormalities of carbohydrate metgabolism have previously been described in ALS patients, it was expected that administration ofrhGH might trigger chemical diabetes in the majority of patients. This was not the case, although two patients developed hyperglycemia. Prior studies have demonstrated insulin insensitiviy and reduction in moncyte insulin receptors in AlS without hyperinsulinism. Although mean insulin levels were generally higher in the treated group, these values were not statistically significant.
In a small number of control patients, IGF-I values remained abonormal throughout the study. With administration of growth hormone there was a marked rise in IGF-I levels in most patients. In at least one of the several growth homrone-treated patients, in whom IGF-I levels did not change with treatment, noncompliance was suspected sicne th patient soon dropped form the study. As judged by the response of IGF-I to administraiton of growth homrone it appreas that the peripheral, princippally hepatic, effects of growth hormone are normal in ALS patients. Earlier workers have reported flattening of the circaian pattern for human growth hormone secretion oand impairment of the stimulated release of hGh in ALS. In males (ages 35-63) with postpolio syndrome, a neuromuscular condition sharing some clinical similarity with JALS, serum IGF-I values have been reported to be below normal. This finding has led to the speculation that denervation of muscle in postpolio patients may be influenced by the loss of pulsatile GH releae that occurs in some older males. Our data would not support a similar hypothesis for ALS.
With the recognition that treatement trials in ALS have often been faulty in design, considerable empahsis was placed on trial methodology. Intorduction of the TQNE has provided a new standard for evaluation of ALS patients in clinical trials. This measurement proved to be highly reproducible and appeared to quantitatively describe the course of the disease. This result, previously reported by others, was expected. Comparing the TQNE to manual testing, it was found that the TQNE more acurately chronicles deterioration of strength that characterizes ALS, especially when weakenss is advanced. Several clinical assessment led to the interesting observation that deterioration of arm stregth was a predictor of early death or withdrawal from the study, independent of the mode of therapy. Since diaphragmatic failure is the most common cause of death in ALS it is not surprising that loss of arm strength is an omnious sign. The diaphragm and upper extremity muscles both receive innervation from the midcervical region, which is preferentially effected in ALS.
Unfortunately, measurement of 3-methylhistidine in the urine, a chemical marker for muscle degeneration, did not correlate with the course of ALS. Excretion of this amino acid did not change appreciably over time and differences between placebo and rhGh=treated ALS patients werenot eveident. This is consistent with results in normal subjects in whome exogenous administration of GH did not effect 3-methylhistidine excretion, which is thought to be a marker of myo-fibrillar protein breakdown. This dat further suggests that GH cannot moderate muscle breakdown in normal subjects or patients with ALS.
There were a number of reasons we had been optimistic about the treatment of ALS with growh hormone. Exogenous administration of protropin exers a profound effect on muscle mass and strength in adults with growth hormone deficiency. This effect is predominantly observed in limb girdle muscles, probably due to the pressence of a proximal myopathy associated with growth hormone deficiency. Somewhat similar results have been observed in malnourished elderly males in whom administration of growth hormone has been seen to have a pronounced effect on muscle mass. Along with a direct trophic effect on muscle, ther is reason to believe that growth hormone mihgt effect neuronal sprouting which accompanies ALS. Although compeltgeely denervated muscle would be unlikely to respond to treatment it ws hoped that remaining motor units might have benefited from hormonal stimulation. While muscle mass and nerve sprouting were not quantified, it is apparent that growth hormone had minimal therapeutic value as judged by the steady decline of strength in the treated group, suggesting that any salutary effects of treatment were overwhelmed by the extent of the degenerative process which characterizes ALS.
It is now known that most, if not all, GH effects are mediated by IGF-I. It has also been shown that GH regulates gene expression of IGF-I in the liver. Although hepatic synthesis appears to be the major source of serum IGF-I, evidence exists for the local production by various tissues that also respon dto GH, including muscle. Although we have demonstrated that administration of exogenous growth hormone results in an almost twofold increase in total serum IGF-I, we did not determine whether IGF-I was increased in muscle of treated patients. Furthermore, we measured only total, rather than free and bound IGF-I in our study. Recently, circulating binding proteins for the IGF's have been demonstrated. Of the six IGF-bindingn proteins (IGFBPs) which have been sequenced and cloned, IGFBP-1 and IGFBP-3 are the most abundant. It is now known that IGFBP-3 is induced by GH administration, wheras direct administration of ministration, wheras direct administration of rhIGF-I does not induce this GH-regulated IGFBP. Accordingly, one explanation for the lack of therapeutic effect of exogenous rhGH could be that much of the IGF-I that was indcued was in the bound form. Furtehr, the lack of effect could be accounted for on the basis of IGF-I inhibittors or antagonists, faulty interaction with muscle IGF-I receptors or a postreceptor defect in neuron and/or muscle which may characterizet this system in ALS patients. These defects might not be corrected thorugh adminstration of parenteral rhGH. It remains to be determined if delivery of pharmacologic doses of rhIGF-I would dresult in a different therapeutic outcome.
Acknowledgments: Support for these studies was provided by the Muscular Dystrophy Association; Genentech, Inc.; the Joseph Drown Foundation; and the Medical Research Service of the Department of Veterans Affairs. The authors gratefully acknowledge the participation of our patients and their families, and Drs. Neal Lewis and David Cooper, both formerly associated with the Muscular Dystrophy Association. Clinical co-ordination and oversight was adeptly provided by Sheri Jacoby, RN (Harbor View Hospital, San Diego); Gabriela Molna, PhD (Cedar Sinai Medcical Center, Los Angeles); Mary Brown, RN (University of Kansas Medical Center, Kansas City, KS); Lori Peterson, BA (Kansas City Veterans Affairs Medical Center, Kansas City, MO); and Joyce Kuntze, RN (Geneentech, Inc., San Francisco). This manuscript is dedicated to the memory of Arthur Levien and Warren Irwin and to Arlene and William Doro, Milton Fillius, Sheila Irwin, Curtis Robert Jr. and Sr., Wendy Wachtel-Schine, and Joan and George Thagard. Further, we are in debt to Shyla Hernandez, who nurtured this manuscript through its numerous incarnations.
Address reprint requests to Barry W. Festoff, MD, Neurobiology Research Laboratory (151), Veterans Affairs Medical Center, 4801 Linwood Boulevard, Kansas City, MO 64128.
Accepted for publication November 25, 1992.
CCC 0148-639X/93/060624-10
Copyright 1993 John Wiley & Sons, Inc.
Nautica
Please Scroll Down to See Forums Below 
