Main

Current opinion holds that treatment with GH does not place patients at risk for leukemia, leukemia recurrence, or recurrence of other tumors (1, 2). Nevertheless, prior risk factors do not explain the occurrence of leukemia in 12 Japanese patients (3). In addition, many of the malignancies described in GH-treated patients have shown distinct patterns, with most being stem cell malignancies such as leukemia, lymphomas, or thymomas (4, 5). Since the introduction of biosynthetic human GH (hGH), most endocrinologists have been using doses of GH substantially greater than used in the past, and there is the theoretical risk that higher doses of GH could affect the potential for malignancy. However, an increased risk for malignancy would need to be explained by a biologic process consistent with our knowledge of carcinogenesis.

Chromosomal aberrations are one form of genetic alteration that occur during carcinogenesis. The genetic diseases ataxia-telangiectasia, Fanconi anemia, and Bloom syndrome are characterized by increased levels of spontaneous and induced chromosomal aberrations and increased susceptibility to neoplasia. Chromosomal aberrations are not only markers of genetic damage but also may have a specific role in carcinogenesis. Hence, most, if not all, cancers exhibit chromosomal alterations. These may be specific to certain cancers, supporting the concept that chromosomal abnormalities are important in the initiation and development of tumors (6). The study of chromosomal aberrations in peripheral blood lymphocytes, in particular, has been used to gain insight into cancer development and cancer risk in humans. Hence, in a 20-y study of more than 3000 subjects, Hagmar et al. (7) found a significant association between cancer incidence and level of spontaneous chromosomal aberrations in peripheral blood lymphocytes.

Two studies have examined chromosomal aberrations in lymphocytes of children treated with hGH. Tedeschi et al. (8) found no significant increase in spontaneous or aphidicolin-induced chromosomal breaks in 10 patients sampled at 6 and 12 mo after initiation of recombinant hGH and in a second group of six patients sampled once between 6 and 24 mo after initiation of therapy. There was, however, a higher frequency of chromosome rearrangements (dicentrics and reciprocal translocations) at 6 and 12 mo into therapy. Lymphocytes from both groups also exhibited bleomycin-induced break frequencies twice that of pretherapy samples and twice that of a control group of short-stature children not receiving GH. They concluded that GH treatment increases chromosomal fragility.

Bozzola et al. (9) found no change in spontaneous chromosomal breaks in the lymphocytes of 12 patients sampled at baseline and 3, 6, 9, and 12 mo after initiation of GH therapy. The level of breaks was also no different from that of an age-matched control group sampled simultaneously at baseline. Contrary to the findings of Tedeschi et al. (8), no increase in chromosomal rearrangements was noted, whereas chromosomal fragility was not tested.

Tumorigenesis is a multistep process, and an increase in chromosomal aberrations or fragility does not imply that cancer will necessarily develop. However, the observation of higher aberration levels in GH-treated patients would be an indication for caution in the administration of this therapy and the need for intensive follow-up of these patients to exclude neoplasia. However, the limited and contradictory data derived to date do not permit one to derive conclusions regarding the genetic effects of GH treatment with any confidence. We, therefore, initiated a study of the effects of recombinant hGH therapy on spontaneous and radiation-induced chromosomal aberrations in peripheral blood lymphocytes of 24 previously untreated patients. In contrast with previous studies, the aberration data in our study are presented in descriptive terms detailing the specific aberration types observed in metaphase cells in first division, and we used ionizing radiation for assessing chromosomal fragility, thus ensuring a uniform exposure by the Go (unstimulated) cells to the clastogenic agent.

METHODS

Experimental subjects.

Twenty-four patients aged 4–16 y were recruited from the Children's Hospital of Wisconsin, Milwaukee, Wisconsin and Miami Children's Hospital, Miami, Florida, there being 18 males and six females. Exclusion criteria for this study were preexisting malignancy, previous radiotherapy or chemotherapy, and syndromes associated with an increased risk of malignancy such as Down's syndrome or Fanconi pancytopenia syndrome. One patient had Turner syndrome. All patients exhibited poor growth, having a growth velocity less than the 25th percentile for age. All were prepubertal except two patients who were at Tanner stage 2 pubertal development. Eighteen patients had stimulated GH levels less than 10 ng/dL on two stimulation tests (either arginine plus clonidine or clonidine plus l-dopa) and were considered to have classic GH deficiency. Five patients had stimulated GH levels above 10 ng/dL and were diagnosed as having neurosecretory GH dysfunction. Informed consent was obtained from all patients for this protocol, which was approved by the Human Research Review Committee of the Medical College of Wisconsin.

All patients were treated subcutaneously with recombinant hGH at a dose of 0.3 mg·kg−1·wk−1. None of the patients was receiving other drug therapy except subject 7 who was being treated with Ritalin. Blood samples were drawn at baseline and 6 mo into treatment for lymphocyte culture for the determination of spontaneous chromosomal aberrations and chromosomal fragility. The 6-mo time period was chosen to make our results comparable to previous studies. Two patients, subjects 7 and 9, who contributed inordinately to an increase in chromatid aberrations at 6 mo were recalled for a third blood specimen. This was 42 mo postinitiation of treatment for subject 7 and 40 mo postinitiation of treatment for subject 9. This third specimen was used to determine spontaneous chromosomal aberrations but not chromosomal fragility. GH therapy had been discontinued on subject 7 6 mo before the third blood draw, whereas subject 9 was still on treatment.

Establishing lymphocyte cell cultures.

Peripheral blood was collected by venipuncture into heparinized tubes. Lymphocyte cultures were established in T25 flasks by adding 1 mL of whole blood to 9 mL of medium (RPMI 1640, 10% FCS, 2 mM l-glutamine). 5-Bromo-2′-deoxyuridine (BrdUrd, final concentration 20 μM) and phytohemagglutinin (PHA HA17, 1 μg/mL, Burroughs Wellcome) were added to each flask. Cultures were incubated at 37°C in a 5% CO2 atmosphere.

Assessing chromosomal aberrations.

Lymphocyte cultures were incubated for 52–56 h after mitogenic stimulation with Colcemid™ (1 × 10−7 M final concentration) being present the last 2 h. These incubation times were chosen to provide reasonable mitotic indices and metaphase cells that were primarily in first division after stimulation. After Colcemid exposure, the cells were transferred to centrifuge tubes and subjected to standard cytogenetic procedures for obtaining fixed mitotic cell suspensions (10). An independent party coded the fixed samples by using a table of random five-digit numbers. The coded samples were stored in a freezer until used for slide preparation. Microscopic slides of mitotic cells were made for groups of patients for whom both baseline and 6-mo samples were available, with a minimum of three subjects per group. This was done as a timesaving measure to minimize delay between the end of patient recruitment and completion of data collection. Samples were not decoded until the chromosomal aberration data were collected from all patients.

To obtain an accurate estimate of the level and type of chromosomal aberrations present in peripheral blood lymphocytes in vivo, we considered it imperative that only metaphase cells in first mitotic division after mitogenic stimulation be used for data collection (11–13). This would exclude metaphase cells in subsequent divisions that may have lost or rearranged their original aberration. We distinguished metaphase cells of different division status by a differential staining method that takes advantage of the BrdUrd incorporated into replicating chromosomes (14). The chromosomal aberration data collected were restricted to metaphase cells whose chromosomes exhibited a first-division staining pattern. We used descriptive terminology for identifying and recording the different chromosomal aberrations (15). Chromosomal aberrations fit into two basic categories, chromosome-type and chromatid-type. Chromosome-type aberrations are those produced before chromosome replication from cells in the G0 or G1 phase of the cell cycle. Chromatid-type aberrations are those typically produced after chromosome replication from cells in the S, G2, and M phase of the cell cycle. Chromatid-type aberrations can also be induced in G0/G1-phase cells but occur at a much lower frequency compared with chromosome-type aberrations. Chromosome-type aberrations identified in this study were dicentrics, tricentrics, rings, and acentrics. Acentrics are noncentromere-containing fragments that result from the generation of several aberration types (e.g. dicentrics). The acentric values recorded are those in excess of the number expected to be produced by the other aberrations present in the cell. These “excess acentrics” were counted as separate aberrations. Chromatid-type aberrations identified in this study were deletions (in which one of the two replicated chromatids has been broken) and exchanges, such as triradials and quadriradials. A chromatid deletion was identified as a discontinuity of the chromatid that was equal to or greater than the width of the chromatid (16). We attempted to score an average of 500 cells for spontaneous (unirradiated) aberrations and 200 cells for radiation-induced aberrations for each donor for both sample times.

Irradiation of cell cultures.

Irradiations were performed at room temperature by using a J.L. Shepard & Associates model 143–45 Gamma Irradiator (Cesium-137, dose rate of 1.6 Gy/min). Cultures were divided into unirradiated (0 Gy) and irradiated (3 Gy) flasks. BrdUrd and PHA were added immediately after the irradiation. Our previous experience has shown that a 3-Gy radiation dose yields an average of one chromosomal aberration per cell. This aberration frequency satisfies the need for statistical sensitivity within a practical number of cells.

Statistics.

Differences between the baseline and 6-mo sample were determined on a per patient and per cell basis. The count data were modeled with a Poisson error structure. In addition, a variance scaling parameter was used to model the overdispersion in the Poisson mean due to variability between subjects (17). To test the significance of the differences in the count of the aberrations from baseline to 6 mo after the initiation of GH, generalized estimating equations were used with the Poison model for the error and log link. An overdispersion parameter for the Poisson and the sandwich estimator for the variance were used to provide appropriate significance levels for the generalized estimating equations (18).

RESULTS

Spontaneous chromosomal aberrations.

Table 1 shows the number of cells analyzed for each of the 23 patients and the number of spontaneous chromosome-type and chromatid-type aberrations at baseline and 6 mo after initiating therapy. Patient 23 did not have data for both sampling times and was not included in the analysis of the spontaneous aberration data. A total of 11,100 cells were scored for baseline and 10,500 cells for the 6-mo sample. The total number of chromosome-type aberrations was similar for both sample times, 62 versus 65 aberrations for the baseline and 6-mo samples, respectively. The total number of chromatid-type aberrations was higher for the 6-mo sample, 67 versus 92 aberrations for the baseline and 6-mo samples, respectively, although the difference was not significant (p = 0.370). Two patients, 7 and 9, accounted for most of the increase in chromatid-type aberrations for the 6-mo sample, with an increase of 16 and 17 aberrations, respectively. A third sample was, therefore, analyzed for these two subjects. Subject 7 showed no aberrations of the chromosome-type and four aberrations of the chromatid-type (three deletions and one exchange). Subject 9 now showed no aberrations of the chromosome-type and one aberration of the chromatid-type (one deletion).

Table 1 Spontaneous chromosomal aberrations in lymphocytes of subjects at baseline and 6 mo after initiation of GH therapy
Full size table

Table 2 shows the mean number of metaphase cells analyzed and the type of aberrations that were detected at baseline and 6 mo. The total aberrations column represents the sum of total chromosome-type and chromatid-type aberrations. The most frequent aberrations observed at both sample times were chromosome-type acentrics and chromatid-type deletions followed by dicentrics. The greatest difference in means for the different aberration categories for the two sample times was for chromatid-type deletions. However, none of the differences approached significance.

Table 2 Mean (± SE) on a per patient basis for type of spontaneous chromosomal aberrations detected at baseline and 6 mo after initiation of GH therapy
Full size table

Table 3 shows the distribution of spontaneous aberrations among the total number of cells for the two sample times. As expected, the majority of the cells showed no aberrations, and, of those that did, the majority had only one aberration. The mean frequency of chromatid-type aberrations at baseline was 0.0059 versus 0.0088 aberrations per cell for the 6-mo sample (p < 0.05). This was due entirely to an increase in the number of cells with one aberration (60 aberrations in 11,100 cells versus 84 aberrations in 10,500 cells). When patients 7 and 9 were excluded from the analysis, the difference in the number of cells with one aberration was no longer significant, with 59 aberrations in 10,100 cells versus 55 aberrations in 9505 cells. There was also no difference in mean frequencies of aberration types, with 0.0064 versus 0.0060 aberration types per cell for baseline and 6-mo samples, respectively.

Table 3 Distribution of spontaneous chromosomal aberrations listed in Table 1 at baseline and 6 mo after initiation of GH therapy Mean represents the mean frequency of the aberrations listed in Table 1. †p < 0.05.
Full size table

Radiation-induced chromosomal aberrations.

Table 4 shows the mean and SEM for the number of cells and the different chromosome- and chromatid-type aberrations observed. A total of 4884 metaphase cells (212 ± 7 cells per patient) were analyzed for the baseline samples and 5000 metaphase cells (217 ± 7 cells per patient) for the 6-mo samples. The majority of aberrations (98%) were of the chromosome-type, as expected after irradiation of lymphocytes in G0 phase. Dicentric aberrations were most prevalent, followed by acentrics and rings. There were no significant differences between the baseline and 6-mo samples for any of the aberration types on a per patient basis. Table 5 shows the distribution of dicentrics, acentrics, and rings among the total number of cells at each sample time. The frequency of rings in the baseline samples was significantly higher than in the 6-mo samples, 0.138 versus 0.113 per cell, respectively (p = <0.001). However, no significant differences were found for the other aberration types.

Table 4 Mean (± SE) of radiation-induced chromosomal aberrations on a per patient basis for baseline and 6-mo samples
Full size table
Table 5 Distribution of radiation-induced chromosome-type aberrations among the total no. of cells at baseline and 6 mo after initiation of GH therapy †p < 0.001
Full size table

DISCUSSION

This study failed to demonstrate a significant change in spontaneous chromosomal and chromatid aberration levels in peripheral blood lymphocytes 6 mo after initiation of hGH therapy. However, when the distribution of spontaneous aberrations among all cells was examined, a significant increase in the number of chromatid-type deletions was seen in the 6-mo samples. This increase was attributable to two subjects. Because the study group as a whole did not exhibit a trend toward higher aberrations, this could have represented an idiosyncratic response for these two patients. These patients were, therefore, recalled for a third blood sample. Cells from this sample demonstrated no change in chromosome or chromatid-type deletions from baseline for either patient. It seems likely, therefore, that this was a spurious result, and that GH therapy has no influence on spontaneous chromosomal aberrations.

This conclusion accords with results from other studies. Bozzola et al. (9) found no change in spontaneous chromosomal breaks in 12 patients sampled at 0, 3, 6, 9, and 12 mo after initiation of hGH therapy. They also found the level of breaks to be no different from that of an age-matched control group sampled at baseline. Our results are also in accord with most of the data of Tedeschi et al. (8) who found no increase in chromosomal aberrations in 10 patients sampled at 6 and 12 mo after initiation of therapy and in a second group of patients sampled once between 6 and 24 mo after commencing therapy. However, contrary to Tedeshi et al. (8), we were unable to demonstrate a higher frequency of chromosome rearrangements in our study.

However, it should be noted that there are several factors preventing direct comparison of our three studies. One factor is the form in which the chromosomal aberration data have been presented. Both Tedeschi et al. (8) and Bozzola et al. (9) expressed their aberration data as chromosomal breaks but used different definitions for the term “breaks.” Different criteria have been used to classify chromosomal discontinuities as breaks (16, 19, 20). Tedeschi et al. (8) used the term in a derivative sense, in that they converted each aberration into the minimal number of breaks needed for its formation. This assumes that the number of breaks resulting in an aberration is the same for each cell and that these numbers are not affected by GH therapy. Bozzola et al. (9), on the other hand, used the term “breaks” in a descriptive sense, referring to a discontinuity in the chromosome. These correspond to the chromatid-type deletions presented in our study. In our study, we applied descriptive terminology and presented the number of each aberration type observed, thereby avoiding the assumptions inherent with derivative interpretations (15, 21). Neither Tedeschi et al. (8) nor Bozzola et al. (9) presented a descriptive classification of the aberration types contributing to their chromosomal breaks other than describing a few structural rearrangements.

Another factor preventing direct comparison of our three studies is the replication status of the metaphase cells contributing to the aberration data. We cultured lymphocytes for 52 to 56 h after mitogenic stimulation, thus assuring that most metaphase cells were in first division, whereas Tedeschi et al. (8) and Bozzola et al. (9) cultured lymphocytes for 72 h. Because the number of mitotic divisions that mitogen-stimulated lymphocytes complete in culture is variable (21, 22), restriction of data collection to metaphase cells in first mitotic division provides a more accurate representation of the aberration level of peripheral blood lymphocytes in vivo. To accomplish this, we used a staining method that permits identification of the number of mitotic divisions completed by a metaphase cell after mitogenic stimulation, and this staining method was used for estimating spontaneous and radiation-induced aberrations at both sample times. Although Tedeschi et al. (8) used a similar protocol for measuring spontaneous aberrations, they did not use it for measuring drug-induced aberrations. Bozzola et al. (9) make no mention of the use of any technique to identify the replication status of metaphase cells. If no specific technique was used, some of their aberration data could be from metaphase cells that had completed more than one mitotic division, thereby leading to possible underestimation or misclassification of aberrations.

In the second part of this study, we examined the effect of hGH therapy on chromosomal fragility by analyzing radiation-induced chromosomal aberrations. The only notable finding was a significant decrease in the frequency of ring chromosomes at 6 mo for the pooled data. This result could suggest that hGH decreases the fragility of chromosomes after radiation. However, an effect limited to the formation of ring chromosomes is not easily explained by proposed mechanisms of radiation-induced chromosomal aberrations. Both rings and dicentrics are chromosome-type exchange aberrations and are thought to be produced by the same DNA repair mechanism. Thus, one would expect both these aberrations to exhibit similar changes. The most likely explanation for the difference in rings may lie in the technical aspects of identifying aberrations: Chromosome aberrations exhibit a range of size, and small rings can be difficult to distinguish from small acentrics. Because rings were present at a lower frequency than acentrics, their frequencies would be more sensitive to misclassification. In support of this, the sum of rings and acentrics shown in Table 4 are equal for both sample times.

The discrepant results between our study and that of Tedeschi et al. (8) for chromosomal fragility may be due in part to the agent that they used to induce aberrations. Bleomycin is associated with variability in cellular uptake and metabolism after exposure, and this can lead to a nonrandom distribution of chromosomal damage among cells (23). We used ionizing radiation to externally irradiate cells, thus avoiding the problems associated with chemical agents such as bleomycin.

Differences in lymphocyte cell cycle phase at the time of agent exposure could also account for the discrepant results. The protocol used by Tedeschi et al. (8) yields metaphase cells that are almost exclusively in the G2 phase of the cell cycle when exposed to bleomycin, thus producing aberrations of the chromatid-type. Bleomycin-induced chromatid break frequencies obtained from G2-treated cells are dependent upon the extent of drug exposure (24) as well as the system used to define the term “breaks” (25). Aberration frequencies obtained from a single sample time of G2 exposed cells are also subject to variations in mitotic delay (21). In the present study, lymphocytes were irradiated before mitogenic stimulation and, thus, were exclusively in the G0 phase of the cell cycle. The level of radiation-induced chromosomal aberrations in G0-phase lymphocytes is independent of sample time when data collection is restricted to first-division metaphase cells (21). The use of irradiated G0 cells, thus, avoids the problems of differential sensitivity and mitotic delay that are inherent in G2 cells exposed to drugs.

In conclusion, our data demonstrate that spontaneous chromosomal aberrations and chromosomal fragility are not influenced by recombinant hGH therapy. Our results are consistent with the findings of Bozzola et al. (9) and some of the observations of Tedeschi et al. (8). They also accord with the results of Lin et al. (26) who found normal mutation frequencies at three gene loci for patients receiving GH therapy.