Kinase Inhibitors: Adverse Effects Related to the Endocrine System

Abstract

Context:

The use of kinase inhibitors (KIs) in the treatment of cancer has become increasingly common, and practitioners must be familiar with endocrine-related side effects associated with these agents. This review provides an update to the clinician regarding the management of potential endocrinological effects of KIs.

Evidence Acquisition:

PubMed was employed to identify relevant manuscripts. A review of the literature was conducted, and data were summarized and incorporated.

Evidence Synthesis:

KIs, including small molecule KIs and monoclonal antibodies directed against kinases, have emerged over the past decade as an important class of anticancer agents. KIs specifically interfere with signaling pathways that are dysregulated in certain types of cancers and also target common mechanisms of growth, invasion, metastasis, and angiogenesis. Currently, at least 20 KIs are approved as cancer therapeutics. However, KIs may affect a broad spectrum of targets and may have additional, unidentified mechanisms of action at the cellular level due to overlap between signaling pathways in the tumor cell and endocrine system. Recent reports in the literature have identified side effects associated with KIs, including alterations in thyroid function, bone metabolism, linear growth, gonadal function, fetal development, adrenal function, and glucose metabolism.

Conclusions:

Clinicians need to monitor the thyroid functions of patients on KIs. In addition, bone density and vitamin D status should be assessed. Special care should be taken to follow linear growth and development in children taking these agents. Clinicians should counsel patients appropriately on the potential adverse effects of KIs on fetal development.

Targeted therapies with small molecule kinase inhibitors (KIs) have become a cornerstone in the treatment of many cancers. Protein tyrosine kinases catalyze the phosphorylation of specific tyrosine residues and thus serve as important regulators of signaling pathways including cellular proliferation, differentiation, and apoptosis (1, 2). Additional small molecule KIs targeting serine/threonine kinases, including BRAF and MAPK, have recently emerged as therapies for malignancies including melanoma and thyroid cancer (3–5). Protein kinases can be inhibited pharmacologically through multiple mechanisms; the 2 most common in clinical use are small molecule KIs and monoclonal antibodies (mAbs) against receptor protein kinases or their ligands.

Most KIs are rationally designed small molecules that directly inhibit the catalytic activity of the kinase by blocking the binding of ATP (1). Small molecule KIs are extensively used as therapies for a number of different of malignancies as outlined in Table 1 (6). Greater than 500 different protein kinases are encoded by the human genome; almost all phosphorylate substrate proteins via their catalytic ATP binding region (7). Due to conservation of the structure of the ATP binding site, many KIs have inhibitory activity against a wide range of protein kinases, with the potential to affect multiple signaling pathways (7–9).

Table 1.

Major FDA-Approved Small Molecule Tyrosine KIs

Drug	Tumors treated	Main small molecule kinases targeted
Axitinib	RCC	VEGFR
Bosutinib	CML	Bcr-Abl, SRC
Cabozantinib	MTC	HGFR, RET, VEGFR
Crizotinib	NSCLC	ALK, HGFR
Dasatinib	CML, ALL	SRC, Bcr-abl, c-KIT, PDGFR
Erlotinib	NSCLC, pancreatic cancer	EGFR
Gefitinib	NSCLC	EGFR
Imatinib	CML, chronic eosinophilic leukemia, ALL, GIST	Bcr-abl, c-KIT, and PDGFR
Lapatinib	Breast cancer	EGFR, HER2
Nilotinib	CML	Bcr-abl c-KIT, and PDGFR
Pazopanib	Soft tissue sarcoma, RCC	VEGFR, PRGFR, c-KIT
Ruxolitinib	Myelofibrosis	JAK
Sorafenib	RCC, hepatocellular carcinoma	BRAF, VEGFR, PDGFR
Sunitinib	GIST, RCC, pancreatic neuroendocrine tumor	VEGFR, PDGFR, c-KIT
Vandetanib	MTC	RET, EGFR, VEGFR
Vemurafenib	Metastatic melanoma	BRAF

Abbreviations: NSCLC, non-small cell lung carcinoma; ALL, acute lymphoblastic leukemia; Bcr-abl, breakpoint cluster region-c Abl oncogene; SRC, sarcoma; ALK, anaplastic lymphoma kinase; HGFR, hepatocyte growth factor receptor; c-KIT, stem cell growth factor receptor; HER2, human epidermal receptor type 2; JAK, Janus associated kinase; BRAF, v-raf murine sarcoma viral oncogene homolog B1; RET, rearranged during transfection. All information can be accessed at http://www.cancer.gov/drugdictionary.

Receptor kinase signaling can also be inhibited by mAbs against the receptor or its ligand. These mABs are distinct from the small molecule KIs in that they target a single kinase rather than affecting the function of multiple kinases. Because of this specificity, mABs with activity against protein kinases are less likely to exert off-target effects. Table 2 lists the major US Food and Drug Administration (FDA)-approved mAbs with activity against protein kinases. In 2004, bevacizumab, an anti-vascular endothelial growth factor (anti-VEGF) mAb, became the first FDA-approved drug in this class for its use in metastatic colorectal cancer (10). This was later followed by cetuximab and panitumumab, both anti-epidermal growth factor receptor (anti-EGFR) mAbs (11, 12). Trastuzumab is a mAb against HER-2, a tyrosine kinase that may be overexpressed in breast cancer (13).

Table 2.

Major FDA-Approved MAbs With Activity Against Tyrosine Kinases

Drug	Tumors treated	Main tyrosine kinases targeted
Bevacizumab	Colorectal cancer, non-small cell lung cancer, RCC, glioblastoma	VEGF
Cetuximab	Colorectal cancer	EGFR
	Head and neck cancer
Panitumumab	Colorectal cancer	EGFR
Trastuzumab	Breast cancer	HER2
	Gastric cancer

In contrast to the adverse effects from conventional cytotoxic chemotherapeutic agents, KIs seem to have a distinct adverse effect profile, providing a relatively high therapeutic window with less toxicity than conventional therapy. However, KIs have important side effects, including fatigue, hypertension, rash, impaired wound healing, myelosuppression, and diarrhea (14). The overall toxicity of KIs, although less life-threatening than conventional cytotoxic chemotherapy, nevertheless is common and may require dose reduction. Endocrine-related adverse effects of these agents include abnormalities in thyroid function, bone metabolism, linear growth, gonadal function, fetal development, adrenal function, and glucose metabolism (15). This review will outline the endocrine-related side effects associated with KIs in order to bring the practicing clinician up to date with the current status of the field.

Methods

Data used to obtain information for the review were identified by search of PubMed citations from January 2001 to November 2012 in English. Search criteria included the keywords “tyrosine kinase inhibitor,” and as well as the individual generic names for each KI listed in Tables 1 and 2, in conjunction with the terms “thyroid dysfunction,” “hypothyroidism,” “bone density,” “parathyroid,” “linear growth,” “ovarian failure,” “osteonecrosis,” “spermatogenesis,” “fertility,” “pregnancy,” “hypogonadism,” “adrenal insufficiency,” and “hypo/hyperglycemia.”

Thyroid Function

As the use of KIs has become widespread, adverse effects of these agents on thyroid hormone function and metabolism have been discovered. A number of excellent reviews have been published on this topic (16–19). As a result of these studies, current recommendations call for prospective thyroid function testing in all patients starting therapy with KIs (16, 20). There does appear to be a distinction in terms of the frequency with which certain KIs are associated with hypothyroidism as outlined in Table 3. However, it should be pointed out that the number of reported cases of hypothyroidism in the literature is influenced by how commonly a particular agent is used as well as how long it has been available.

Table 3.

Frequency of Hypothyroidism Reported With Different KIs

Common	Moderate	No reports
Sorafenib	Axitinib	Bosutinib
Sunitinib	Cabozantinib	Crizotinib
Imatinib	Dasatinib	Gefitinib
	Erlotinib	Lapatinib
	Nilotinib	Ruxolitinib
	Pazopanib	Vemurafenib
	Vandetanib

There appear to be 2 distinct types of thyroid disturbances associated with the use of KIs. In the first case, patients with preexisting hypothyroidism develop grossly increased TSH and reduced T₄ while on therapy, whereas most patients with normal pretreatment thyroid functions and intact glands show no change in thyroid function tests (TFTs). This drug effect has been linked to the use of imatinib, sorafenib, and vandetanib (21–25). The second type of thyroid disturbance associated with KIs is de novo hypothyroidism in patients with previously normal thyroid function, associated with sunitinib, sorafenib, imatinib, dasatinib, nilotinib, and axitinib (26–32).

In 2005, imatinib became the first KI reported to cause hypothyroidism. In 8 thyroidectomized medullary thyroid cancer (MTC) patients, marked elevations in TSH up to a mean of 384 ± 228% of the upper limit were recorded during treatment (21). Free T₄ was decreased but generally remained within the reference range. Although the average dose of levothyroxine was doubled, thyroid functions normalized in only 3 of 8 patients. In this initial description of thyroid dysfunction associated with KIs, the patients were athyroid; therefore “thyroid dysfunction” could be better described as “alterations in TSH and free T₄” or “worsening postsurgical hypothyroidism,” because the pathophysiology could not be explained by the thyroid gland. One study evaluating the thyroid function of 68 patients with chronic myeloid leukemia (CML) and intact thyroid glands who received imatinib did not find any adverse effect on thyroid function, suggesting that possibly the effect of imatinib was unique to postthyroidectomy patients (24). In the randomized phase III trial of vandetanib for patients with MTC, 90% of the patients had prior thyroidectomy, and increases in the dose of thyroid replacement were required in 49% of the patients randomized to vandetanib compared to 17% of the patients on placebo (25).

A number of other studies have followed thyroid functions in patients receiving KIs who have intact thyroid glands. In patients with metastatic renal cell cancer (RCC), hypothyroidism developed in 7 of 39 patients (18%) and was first observed 2–4 months after initiation of sorafenib (30). Sunitinib is the KI most frequently associated with hypothyroidism, with the incidence ranging from 20 to 85% (26, 28, 29, 33–36). Newer second generation KIs have also been associated with hypothyroidism, which was reported in 69 of 359 patients (19%) receiving axitinib and 29 of 355 patients (8%) receiving sorafenib for RCC (32). Thyroid abnormalities were detected in 25, 55, and 70% of individuals with CML taking imatinib, nilotinib, and dasatinib, respectively (31). Thyroid dysfunction has not, however, been reported in the literature associated with mAbs with activity against protein kinases.

There are still uncertainties regarding the mechanism by which KIs cause thyroid dysfunction, and theories are presented in Table 4. One hypothesized mechanism is destructive thyroiditis. A number of studies have found an association between sunitinib use and the presence of antithyroglobulin antibodies, low TSH before the development of hypothyroidism, and/or atrophic thyroid tissue on ultrasound, all thought to be consistent with destructive thyroiditis (26, 29, 37, 38). Another proposed mechanism of KI-induced thyroid dysfunction is impaired iodine uptake. Thyroid ultrasound scans and ¹²³I thyroidal uptake were done at the end of several periods of sunitinib treatment in patients with gastrointestinal stromal tumor (GIST), in which ¹²³I uptake was significantly reduced at the end of treatment periods (36). An alternative hypothesis for KI-induced thyroid function is capillary regression induced by VEGF receptor (VEGFR) inhibition. This theory is supported by the observation of marked shrinkage of the thyroid gland during treatment with sunitinib in 3 patients (39, 40). Some of the proposed mechanisms cannot account for the observed alterations of TFTs in studies of postthyroidectomy patients. TSH elevation in postthyroidectomy patients might be explained by an indirect effect of sunitinib on the metabolism of thyroid hormone, or with thyroid hormone action at the pituitary level. One recently induced hypothesis based on in vitro studies of MCT8-mediated iodothyronine transport showed that partial inhibition by KIs of pituitary or hypothalamic thyroid hormone feedback may increase TSH or increase the levothyroxine requirement of thyroidectomized patients (41). Rather than a single unifying drug effect, different types of KIs may in fact have more than 1 mechanism affecting thyroid function.

Table 4.

Proposed Mechanisms of Tyrosine KI-Induced Hypothyroidism

Decrease in enteric thyroid hormone absorption

Increased hepatic metabolism of T4 and T3

Inhibition of T4 deiodination

Impaired iodide update

Stimulation of T4 and T3 clearance

Destructive thyroiditis

Direct inhibition of thyroid peroxidase activity

Capillary regression induced by VEGF inhibition

Interference with thyroid hormone action at the pituitary gland level

The optimal management of KI-induced hypothyroidism remains to be elucidated. At a minimum, baseline thyroid function tests should be performed before the initiation of KI therapy, and TFTs should be frequently monitored during treatment. Levothyroxine treatment should be started when clinical hypothyroidism develops. One algorithm for testing recommends measuring serum TSH and free T₄ levels and thyroid antibodies at baseline, and measuring serum TSH levels on day 1 of each chemotherapy cycle (16). However, the recommendations for management of asymptomatic subclinical hypothyroidism (TSH, 5–10 μIU/mL) are unclear in cancer patients in whom symptoms of hypothyroidism, such as fatigue, might coincide with symptoms of the malignancy and its treatment. Individuals with elevated TSH can be effectively managed with thyroid hormone replacement; therefore, hypothyroidism alone is not an indication for dose reduction or discontinuation of KI therapy (16). To fully understand the underlying molecular mechanism of KI-related hypothyroidism, additional studies are necessary.

Altered Bone Density and Secondary Hyperparathyroidism

Recent reports have described altered bone and mineral metabolism and secondary hyperparathyroidism in patients taking KIs (42–48). Because patients often continue KI treatment indefinitely, awareness of potential effects of these agents on the skeleton is imperative. Long-term treatment with imatinib has been associated with altered bone metabolism, decreased bone remodeling, as well as hypophosphatemia (44, 49–51). The proposed mechanism for imatinib-induced secondary hyperparathyroidism is an increase in bone mineral density due to both decreased bone resorption and increased bone formation (45, 52, 53). In 9 patients prospectively studied for 2 years while on imatinib, secondary hyperparathyroidism and decreased bone turnover were found (53). The effect of imatinib over 4 years in 17 CML patients showed a high incidence of secondary hyperparathyroidism, yet with stable mean areal and volumetric bone mineral density on repeat dual-energy x-ray absorptiometry scans (54). In 26 patients receiving sunitinib for RCC, hyperparathyroidism developed in 69% of the patients (46). In 39 patients with thyroid cancer on vandetanib, serum 25-hydroxyvitamin D level decreased and serum PTH levels increased, suggesting possible decreased intestinal absorption of vitamin D (47).

Nonspecific inhibition of kinases expressed by osteoclasts and osteoblasts, such as c-KIT and platelet-derived growth factor receptor (PDGFR)-A, may explain the effects of KIs on bone metabolism (44). Preclinical studies have shown that the dasatinib can cause dysregulation of bone remodeling via inhibition of osteoclasts (55, 56). Recent cancer treatment approaches have therefore attempted to use KIs to target the bone microenvironment in the treatment of metastatic bone disease in breast and prostate cancer (57, 58). In terms of the clinical management of skeletal adverse effects of patients on KIs, no clear recommendations exist, but based on clinical data obtained thus far, it is prudent to monitor bone density, PTH levels, and 25-hydroxyvitamin D levels in these patients at baseline and at yearly intervals, or more frequently as clinically indicated.

Osteonecrosis of the Jaw

Due to the antiangiogenic properties of KIs, they have the potential to exacerbate bisphosphonate (BP)-induced osteonecrosis of the jaw (ONJ). Five patients with a history of BP and sunitinib administration presented with ONJ in close relation to initiation of sunitinib, whereas 2 case reports describe ONJ associated with the use of bevacizumab alone (59–62). The incidence of ONJ among 116 patients receiving BPs with or without KIs (including bevacizumab, sorafenib, or sunitinib) was 16 and 1.1%, respectively (P = .008), suggesting that the combination of BPs and KIs induces ONJ more frequently than BPs alone (63). However, a case review of 27 patients with breast cancer treated with either BPs alone or BPs plus bevacizumab showed no effect of the addition of bevacizumab on the development of ONJ (64). A larger retrospective study of incident ONJ cases among BP and/or bevacizumab-treated patients demonstrated an incidence of ONJ in 72 (1.1%) of the 6561 patients treated with iv pamidronate and/or zoledronic acid alone, none (0%) of the 1711 patients treated with bevacizumab alone, and 8 (2.0%) of the 409 patients treated with the combination of bevacizumab and iv BPs. The authors concluded that there is a modest risk of ONJ with BP administration that may be compounded by bevacizumab administration; yet administration of bevacizumab alone does not appear to be associated with significant ONJ risk (65). Additional retrospective analyses of 3 large trials of breast cancer patients also showed no increase in the incidence of ONJ in patients who received bevacizumab with BPs compared to BPs alone (66). Although studies do not uniformly show that KI administration affects the development of ONJ, endocrinologists involved in the care of patients treated with these agents should be aware of the potential complication of ONJ, especially because patients with breast cancer are often treated concurrently with bevacizumab and BPs.

Linear Growth

Normal childhood growth relies on many of the same pathways implicated in tumor pathogenesis, such as angiogenesis. Preclinical studies have raised concern for potential adverse effects of KIs on longitudinal growth (67). In preclinical studies, imatinib has been associated with thickening of the epiphyseal growth plate and impaired bone length, especially in prepubertal animals (68, 69).

Several case reports have documented growth deceleration in children with CML treated with imatinib (70–73). One such case focuses on the disturbance of GH secretion as the mechanism of growth impairment. A 7-year-old identical twin with CML fell from the 95th to the 25th percentile for height over 5 years of treatment with imatinib and dasatinib, whereas her twin sister continued to grow at the 95th percentile. The patient was diagnosed with acquired GH deficiency (GHD) based on inappropriately low GH in response to arginine and clonidine stimulation (73). Three larger retrospective clinical series have evaluated growth in children with CML on imatinib. The first studied 34 children, median age of 10 (range, 2–13) years, showing a reduction in height SD score (SDS) (P = .002) at the fifth year of treatment, with prepubertal children affected more severely (74). The second studied 48 children with a median decrease in height from 0.01 to −0.85 SDS (P < .001); severity of growth impairment was related to younger age at start of treatment (75). Finally, a third group showed a median height decrease of −0.37 SDS during the first year of treatment in 22 children (76). Although most reported data on growth effects of KIs are specific to imatinib, other KIs may have similar effects (77).

Imatinib therapy has also been linked to GHD in adults. Seventeen adults with CML on imatinib therapy underwent glucagon stimulation testing, and 70% had severe GHD (serum GH level < 3 μg/L after glucagon stimulation testing) (78). Due to the growing evidence that KIs are associated with growth impairment and possibly GHD, careful monitoring of growth velocity as well as IGF-I is recommended for children treated with these agents. The mechanism of growth retardation may be a direct effect of the KI on growing bone and/or interference with protein kinase-mediated GH secretion. Treatment with GH is contraindicated in children with an active malignancy; however, some have suggested that GH replacement may be a possible treatment in children with GHD who are chronically treated with KI (73). Because GH has been shown to play a role in the malignant transformation and progression of a variety of cancers, including breast, lung, and colon, any recommendation to administer GH to individuals with a malignancy would likely be controversial (79).

Gynecomastia

Several cases of gynecomastia have been reported with the use of imatinib, dasatinib, and sunitinib (71, 80–83). One larger study analyzed testosterone levels in 38 men receiving imatinib for CML at baseline and during treatment. Seven cases of gynecomastia were noted (18% of patients), associated with a significant decrease in testosterone concentrations (84). The mechanism by which sunitinib and other KIs may induce gynecomastia remains unknown; however, one potential pathophysiological link is the expression of c-KIT and PDGFRA in the testis where they are involved in testosterone production (85). Male patients who receive KIs should be monitored clinically for the potential development of hypogonadism.

Fertility

KIs inhibit proteins with known roles in gonadal development, suggesting that they may have an adverse impact on fertility; however, data regarding the effects of KIs on gonadal function and subsequent fertility are lacking (86). In a prospective study of 179 premenopausal women randomized to receive chemotherapy with or without bevacizumab, the incidence of ovarian failure was higher in the bevacizumab arm (34%) compared to the control arm (2%). After discontinuation of bevacizumab and chemotherapy, recovery of ovarian function occurred in 22% of bevacizumab-treated patients; therefore, practitioners are advised to inform women of the risk of ovarian failure with bevacizumab (87). One case of severe oligospermia in a male patient with CML who started imatinib before puberty has been reported (88). One case of primary ovarian insufficiency in a 30-year-old woman within 2 years of initiating imatinib therapy has been reported; however, this has not been substantiated (89). One recent case describes a 17-year-old female with CML attempting oocyte retrieval with a compromised response to gonadotropins while on imatinib that normalized when retrieval was repeated after stopping imatinib (90). However, recent animal studies investigating spermatogenesis and folliculogenesis in mice treated with imatinib showed no difference in spermatogenic activity, testicular morphology, or follicular development (91). Future long-term evaluation of the effects of KIs on ovarian function and fertility are required; however, it remains prudent to recommend strategies to preserve fertility from the time of diagnosis.

Pregnancy

The long-term use of KIs has led to the need to address their effect on fertility and pregnancy outcome. Adequate and well-controlled studies of KIs in pregnant women are lacking. Increasing evidence demonstrates that children born to men who are actively being treated with imatinib at conception have no increased risk for congenital malformations (92–95). Current recommendations are that male patients who wish to father children may continue imatinib treatment (96). Unfortunately, a worrisome cluster of rare congenital malformations has prevented imatinib from being recommended safely during pregnancy (96). The most comprehensive data set on the effect of imatinib during pregnancy comes from a study of 125 women known to have conceived while on therapy (97). Twelve of the pregnancies resulted in infants with fetal abnormalities, including exomphalos, renal agenesis, cardiac and bony defects; similar abnormalities were seen in preclinical studies of exposed pregnant rats and in mice homozygous for mutations in PDGFR-α (37, 97, 98).

As its use becomes more widespread, data are emerging on the effects of trastuzumab in pregnancy. The first large randomized trial assessing its effect on pregnancy course and outcome was recently published. A total of 5102 patients with HER2-positive breast cancer were randomized to either 1 or 2 years of trastuzumab or observation after completion of chemotherapy; 70 pregnancies were reported (99). Trastuzumab did not appear to affect fetal outcome, and none of the newborns had congenital abnormalities (99). However, 8 case reports have described oligo- or anhydramnios in pregnancies associated with trastuzumab exposure (100). The package insert cautions that use of trastuzumab during pregnancy resulted in cases of oligohydramnios, pulmonary hypoplasia, skeletal abnormalities, and neonatal death (www.herceptin.com/pdf/herceptin-prescribing.pdf).

Current recommendations advise for female patients who wish to conceive to discontinue imatinib during conception and pregnancy due to its recognized teratogenic effects (97). However, in cases where alternative therapy or stopping therapy are not acceptable alternatives, KIs have been used during pregnancy. A number of KIs have been associated with oligohydramnios, necessitating the close ultrasound follow-up of growth and amniotic-fluid index in these patients (102). Because of the observed teratogenic effects of KIs in humans, they should be used during pregnancy only if the potential benefit to the pregnant woman justifies the potential risk to the fetus.

Adrenal Insufficiency

Due to the role of protein kinases in the hypothalamic-pituitary-adrenal (HPA) axis and in steroid secretion, it follows that KIs may be linked to the development of adrenal insufficiency. In practice, however, adrenal insufficiency is an extremely rare adverse event and has only been reported thus far in 2 patients receiving saracatinib, a KI in trials for use in breast cancer (103). Adrenal toxicity was noted in animal studies of sunitinib, yet in clinical studies, ACTH stimulation testing performed in 400 patients found only 1 patient with consistently abnormal test results during treatment, with no clinical evidence of adrenal insufficiency in the patients followed (38).

In 25 patients with CML treated with imatinib, glucagon stimulation testing and low-dose (1 μg) ACTH testing were performed to evaluate the HPA axis. Twelve (48%) patients were diagnosed with adrenal insufficiency in this study (peak serum cortisol level < 18 μg/dL), indicating an increased prevalence of subclinical HPA dysfunction in patients receiving imatinib (104). The FDA drug approval summary cautions that despite the absence of clinically important adrenal suppression in patients taking sunitinib, subclinical toxicity may be unmasked by physiological stress; therefore, monitoring for adrenal insufficiency is recommended in patients undergoing stressors such as surgery, trauma, or severe infection (105).

Glucose Metabolism

KIs have been implicated in alterations of blood glucose (BG) metabolism; however, both elevated and decreased BG levels have been described, and the mechanism by which they alter BG levels is unknown. A number of case reports document improved glycemic control in diabetic patients after the administration of KIs (106–110). In addition, there are 3 reports of clinically significant hypoglycemia developing in nondiabetic patients while on KIs (111–113). A retrospective study of BG concentrations in 17 diabetic and 61 nondiabetic patients treated with KIs including dasatinib, imatinib, sorafenib, and sunitinib showed statistically significant but modest decreases in mean BG, which were reversible upon discontinuation of the drug in almost all cases (114); 47% of the patients with diabetes were able to discontinue their diabetes medications while on a KI (114). The authors suggest that clinicians should be aware of the potential hypoglycemic effect of KIs because modification of diabetes management may be required.

In contrast to the other KIs, nilotinib has been associated with transient hyperglycemia. The phase II trial of nilotinib for patients with CML reported hyperglycemia as a grade 3 or 4 toxicity in 12% of patients (115). A prospective analysis was performed to determine the effects of nilotinib therapy on glucose metabolism in a subset of patients with CML with and without preexisting type 2 diabetes (116). Of 836 patients, hyperglycemia occurred as grade 3 or 4 toxicity in 10% of patients on nilotinib. No patients discontinued the study due to hyperglycemia, and there were no diabetic serious adverse events. In the subset of patients with preexisting type 2 diabetes (n = 57), 74% did not have a change in diabetes therapy on study (116).

The clinical significance of altered BG levels in patients receiving KIs is difficult to interpret, given that some agents have opposite effects on glucose levels and most changes appear to be modest. In diabetic patients, careful assessment of glycemic control while on KIs is recommended. Monitoring hemoglobin A1C and BG levels at baseline and at yearly intervals for nondiabetic patients while on treatment is a reasonable recommendation.

Overlapping Signaling Cascades?

Although PTH and TSH elevations have been associated with the use of KIs, the etiology is still unknown. Review of endocrine physiology reminds us that both the TSH and TRH receptors are members of the G protein-coupled receptor (GPCR) superfamily, whereas the thyroid hormone receptor is a nuclear receptor; thus, none of the thyroid signaling pathway receptors belong to the tyrosine kinase class. The same can be said for the PTH receptor 1 and the calcium-sensing receptor, both GPCRs. Although these signaling cascades were once thought to be discrete, there may be integrated cross talk between certain receptor tyrosine kinase and GPCRs, whereby these signaling pathways join to form complex signaling networks (101). This “cross talk” may have consequences when considering drug side effects because these agents may be less specific than initially thought.

Conclusions

As the use of KIs becomes increasingly common, practitioners must be familiar with the recognition and management of endocrine-related side effects associated with these agents. Patients on these agents should have their thyroid function monitored, and bone density, PTH, and 25-hydroxyvitamin D levels should be assessed periodically. Children on KIs should be followed closely for abnormalities in growth and pubertal development. Women of childbearing age need to be counseled about the known adverse effects of KIs on fetal development. Clinicians should be aware of the potential association between ONJ and these agents. At this time there does not appear to be enough evidence to require routine monitoring for adrenal insufficiency. Finally, diabetic patients on KIs should be monitored closely for changes in glucose metabolism. Further investigation is required to better understand the exact molecular mechanisms underlying endocrine dysfunction in patients receiving KIs.