|Home | About | Journals | Submit | Contact Us | Français|
Mutations in GBA1 gene that encodes lysosomal glucocerebrosidase result in Type 1 Gaucher Disease (GD), the commonest lysosomal storage disorder; the most prevalent disease mutation is N370S. We investigated the heterogeneity and natural course of N370S GD in 403 patients. Demographic, clinical, and genetic characteristics of GD at presentation were examined in a cross-sectional study. In addition, the relative risk (RR) of cancer in patients compared with age-, sex-, and ethnic-group adjusted national rates of cancer was determined. Of the 403 patients, 54% of patients were homozygous (N370S/N370S) and 46% were compound heterozygous for the N370S mutation (N370S/other). The majority of N370S/N370S patients displayed a phenotype characterized by late onset, predominantly skeletal disease, whereas the majority of N370S/other patients displayed early onset, predominantly visceral/hematologic disease, P < 0.0001. There was a striking increase in lifetime risk of multiple myeloma in the entire cohort (RR 25, 95% CI 9.17–54.40), mostly confined to N370S homozygous patients. The risk of other hematologic malignancies (RR 3.45, 95% CI 1.49–6.79), and overall cancer risk (RR 1.80, 95% CI 1.32–2.40) was increased. Homozygous N370S GD leads to adult-onset progressive skeletal disease with relative sparing of the viscera, a strikingly high risk of multiple myeloma, and an increased risk of other cancers. High incidence of gammopathy suggests an important role of the adaptive immune system in the development of GD. Adult patients with GD should be monitored for skeletal disease and cancers including multiple myeloma.
Non-neuronopathic, Type 1, Gaucher Disease (GD), the most common lysosomal storage disorder, affects ~1 in 850 Ashkenazi Jewish and 1 in 40,000 non-Jewish persons [1,2]. The metabolic defect, an autosomal recessive deficiency of lysosomal glucocerebrosidase (GBA1), results in the accumulation of glucocerebroside in the lysosomes of mononuclear phagocytes that appears to trigger a chronic inflammatory state and a complex multi-systemic phenotype. The disease manifestations are extremely heterogeneous, affecting the liver, spleen, bone marrow, skeleton, and lungs. The factors determining whether all disease compartments are equally affected or one compartment more than others are not understood. Moreover, a number of unusual associations have been described recently, including Parkinsonian syndrome [3,4], pulmonary hypertension , cholelithiasis, and multiple myeloma [6–10]. These insights have led to the current concept that GD is a continuum of phenotypes ranging from lethal disease in neonates to asymptomatic older adults . An alternative concept, however, is that this disease defined by a single gene defect and a single enzyme deficiency, in fact, engenders multiple discrete syndromes possibly under the influence of genetic and/or epigenetic modifiers. Therefore, the challenge is to delineate these distinct syndromes through careful annotation of phenotypes to improve patient management and delineate pathophysiologic mechanisms of GD.
Studies correlating GBA1 genotype with disease severity have shown that mutations in GBA account for only part of the variability of GD and prognosis cannot be predicted on the basis of genotype information alone . The most common disease mutation among Ashkenazi Jewish and non-Jewish Caucasian patients is N370S, accounting for ~70% of disease alleles in patients of Ashkenzi Jewish ancestry and ~44% of disease alleles in non-Jewish patients . This mutation arose between the 11th and 13th centuries, and it is believed to provide a selective advantage, the nature of which is unknown . Homozygosity for the N370S mutation is commonly perceived to lead to minimal symptoms in adults, although severe childhood disease may occur [15,16]. In fact, it has been suggested that the majority of N370S/N370S individuals are asymptomatic and may never come to medical attention . Therefore, screening or evaluation of such patients in high risk populations and treatment of N370S/N370S patients have been discouraged [18,19]. However, the true penetrance, range of expressivity, and natural history of N370S/N370S GD are not known. There are conflicting data from small numbers of patients in short-term longitudinal studies of the natural course of N370S homozygous GD [19,20]. It is becoming increasingly important to have a better understanding of the natural history for more accurate counseling, optimal lifelong monitoring of disease progression, and efforts to effectively direct costly enzyme replacement and other emerging therapies for GD.
We delineated the natural history of N370S/N370S GD by performing a cross-sectional study in a cohort of 403 patients who presented to a tertiary referral clinic and its collaborating clinics, encompassing the States of NY, AZ, CT, and NJ. Our study demonstrates an unexpectedly high expressivity of N370S/N370S GD characterized by adult-onset, progressive skeletal disease with relative sparing of the visceral hematologic compartments. Moreover, these patients are at increased risk of developing cancers, especially multiple myeloma and other hematologic malignancies. Underrecognition of this distinct pattern of disease may contribute to failure to diagnose the condition in a timely fashion and when indicated, to begin enzyme therapy, which is effective in reversing or preventing numerous disease manifestations [21,22]. Our findings support the concept that careful phenotypic annotation can lead to identification of specific syndromes within the GD clinical spectrum. Such approaches have the potential to improve patient management and guide future scientific investigations.
Patients were classified according to disease severity in each GD domain at age of first comprehensive evaluation before initiation of enzyme replacement therapy (ERT), presenting symptom, age of symptom onset, and phenotype (early-onset visceral vs. adult onset skeletal with mild visceral/hematologic disease), age of diagnosis, GBA1 genotype, ethnicity, whether therapeutic splenectomy had been performed previously, and number of years of follow-up (Table I). Fifty-four percent of patients were homozygous for N370S mutation and 46% were N370S/other genotype. At the time of first visit to the tertiary referral clinics of one of the authors, 29% of patients were treatment naive and 71% were on ERT. We were careful to avoid the confounding effect of ERT in our phenotype assignment. This information is usually available from referring physicians because it is mandatory to obtain for approval of ERT. Therefore, in all of the patients included in this study, we made sure that pre-ERT data were available; this information was used to assign the phenotype severity in each disease domain. Ninety percent of patients were of Ashkenazi Jewish ancestry, reflecting genotype distribution. Presenting symptoms included symptomatic splenomegaly, bone crises, thrombocytopenia, anemia, and bone pain or fracture; a minority was identified through screening. There was a significant difference in age of onset of symptoms between N370S/N370S and N370S/other patients (34 ± 18.3 years vs. 14 ± 13.7 years, respectively, P < 0.0001), also depicted as symptom-free survival in Fig. 1. There was a 5 year delay in the diagnosis of GD after the onset of symptoms in 10% of N370S/N370S and in 15% of N370S/other patients and a delay of >5 years in 13% and 9%, respectively. Therefore, in nearly one quarter of patients in each genotype category (23% and 24%, respectively) there was a significant lag from onset of symptoms to the correct diagnosis.
Next, we evaluated the spectrum of severity associated with N370S GD. We determined the severity of disease in each GD domain (hepatomegaly; splenomegaly; bone disease score; hemoglobin; white blood cell and platelet counts; and cancer type). For each disease domain, N370S/N370S to N370S/other patients were compared.
We found that presentation at a younger age was associated with worse organomegaly compared with presentation at older age for both genotype groups (see Fig. 2). Patients homozygous for the N370S mutation had less severe visceral involvement: liver volume −0.5 ×N, P < 0.001, spleen volume −4.9 ×N, P < 0.001, compared with patients with N370S/other genotype, as expected from previous reported studies. Consistent with this finding, hemoglobin, platelet, and white blood cell counts were higher in patients with N370S homozygous disease, reflecting less visceral involvement and therefore less sequestration. In marked contrast, there was worsening skeletal disease with age, as depicted by Hermann Score (HS) among N370S homozygous as well as N370S heterozygous patients. Comparison of the two genotype groups revealed more severe skeletal disease in N370S heteroallelic patients: HS −1.1, P < 0.001. SSI increased with age in both genotype groups (data not shown). Asplenic patients were excluded from these analyses because of the confounding effect of the known association of severe skeletal disease in asplenic patients. Mean HS for N370S/N370S and N370S/other patients stratified by spleen status and ERT is noted in Table II. The majority of N370S homozygous patients (187/211, 89%) expressed late onset, predominanly skeletal disease in contrast to N370S/other patients (117/187, 63%) who expressed early onset, predominantly visceral disease, P < 0.0001 (Table I).
We evaluated cancer risk in our cohort. Of 367 patients for whom there was follow-up data, 46 (12%) developed cancers. There were a total of 55 cancers reported (seven patients had multiple cancers) as shown in Table III. Interestingly, the diagnosis of cancer occurred before initiation of ERT in the majority (32 patients, i.e., 69.6%); in 8 patients the cancer diagnosis was made 6 months to 8 years (mean 3.9 years) after intiation of ERT. The temporal sequence of cancer development and ERT could not be reliably determined in 6 patients. Lifetime probability of cancer was assessed for the first presenting cancer in those patients who harbored multiple cancers. Compared with the US population, the overall lifetime risk of developing cancer was increased: RR 1.80 (95% CI 1.32–2.40, P < 0.0001). There was a striking increase in the lifetime risk of developing multiple myeloma (RR 25, 95% CI 9.17–54.40, P < 0.0001). In our study, we derived cancer risk from the first presenting cancer. Two patients developed multiple myeloma after the onset of another type of malignancy. In fact, the RR of developing myeloma in GD in the absence of selecting for the first cancer is even higher at 37.5 (CI 17.15–71.19). In the entire cohort, there was an increased risk of developing hematologic malignancies excluding multiple myeloma (RR 3.45, 95% CI 1.49–6.79, P < 0.0001).
Comparison of GD patient characteristics according to cancer type is shown in Table IV. The mean age of multiple myeloma patients was significantly higher than that for other hematologic or non-hematologic cancers (71 ± 10 years vs. 55 ± 15 years vs. 57 ± 16 years, P < 0.05). All GD1 patients with multiple myeloma were older than 50 years at the time of diagnosis, similar to the non-GD population. The majority of those with multiple myeloma (8/9, 88%) were of N370S/N370S genotype (P = 0.04). In contrast, 6/8 (75%) patients with other hematologic malignancies and 18/29 (62%) patients with non-hematologic cancers were of N370S/N370S genotype, not significantly different from the prevalence of this genotype in the rest of our patient population. A higher proportion of patients with non-myeloma hematologic malignancies and other cancers were asplenic (4/8 and 13/29, respectively) compared with the multiple myeloma patients (2/9), P < 0.05. However, overall SSI in all three cancer types was not different and indicated disease severity from extremely mild to moderately severe, SSI range 2–18 (Table IV).
Polyclonal gammopathy and MGUS have been described in GD [23–25]. We examined the natural course of multiple myeloma in our patient population by comparing Kaplan-Meier curves for event-free survival for polyclonal gammopathy, MGUS, and multiple myeloma. The results are depicted in Fig. 3. There appears to be a temporal sequence of early occurrence of polyclonal gammopathy in a large proportion of patients, followed by MGUS, and eventually multiple myeloma. Kaplan-Meier event-free survival curves support the hypothesis of sequential progression of polyclonal gammopathy to MGUS and ultimately to multiple myeloma with the mean age of development at 62, 78, and 84 years; corresponding number of events were 132, 17, and 9, respectively.
We aimed to estimate the natural history of N370S GD in a large, predominantly Caucasian, Ashkenazi Jewish patient population. We found that the phenotype resulting from homozygosity for N370S mutation in the GBA gene is distinct from commonly depicted features of GD in which visceral and hematologic disease predominates with variable bone involvement. N370S/N370S GD results predominantly in an adult-onset disease with a disproportionate burden of progressive skeletal disease in the absence of major visceral or hematologic involvement; a minority of patients present with the classic form manifesting as early onset-childhood disease with prominent visceral/hematologic disease and variable skeletal disease. These patients are at a striking 25–37.5-fold increased lifetime risk of developing multiple myeloma and are also at increased risk of developing other cancers. The predilection of N370S homozygous GD patients to develop progressive skeletal disease with only mild visceral and hematologic disease may contribute to diagnostic delays . Furthermore, our findings suggest that in adults with N370S/N370S GD, the risk from occult skeletal disease, multiple myeloma, and other cancers should be addressed when evaluating patients who may appear to have insignificant visceral and hematologic disease.
Two studies are frequently cited as depicting a benign course of GD. One study  reported that there was no change in the severity score index (SSI) of 29 patients after 2–13 years of follow-up; 13 patients in this series were N370S homozygous and the remainder had severe genotypes, including homozygous L444P, known to result invariably in progressive and frequently, neuronopathic disease. A later study  concluded that there was no progression of disease in 68 N370S homozygous GD patients followed for a mean of 2.6 years. However, this conclusion is surprisingly at variance with the finding in that study of a statistically significant increase in the SSI and significant decreases in hemoglobin and platelet counts. Despite these observations indicating progressive disease in N370S homozygous GD, the prevailing view is that it is a benign disorder of extremely low penetrance, and regular comprehensive monitoring or consideration of ERT should be given a lower priority for this group of patients [18,19]. This false perception may be perpetuated by the atypical presentation we have delineated in N370S/N370S GD. Moreover failure to recognize this previously undefined presentation may contribute to diagnostic delays of many years, which could result in severe complications, a number of which are preventable or ameliorated by currently available enzyme therapy, the standard of care for GD1 . In our study, we found that nearly 25% of patients experience diagnostic delays from 1 to 30 years after the onset of symptoms. In a recent study, we found patients with N370S/N370S GD are most likely to suffer diagnostic delays associated with disabling complications, that are preventable or reversible by timely therapy [21,22].
Our data underscore the importance of evaluating individual domains of disease activity to reliably assess the effects of GD. This approach permitted delineation of divergent manifestations of N370S GD on the viscera and hematologic profile versus the skeleton. We have found the HS to be a useful instrument to assess the diversity and severity of skeletal involvement in GD. Bone disease in GD is a complex, multidimensional process, involving osteopenia, marrow expansion, lytic lesions, avascular necrosis, fracture, and/or collapse. HS can be derived simply from plain films. Although the MR bone marrow burden score proposed by Maas et al.  appears to be a sensitive tool to assess marrow disease, the advantage of the HS is that it ascertains severity of all of the diverse aspects of GD bone involvement and it may reflect the natural history of bone involvement.
Among the entire cohort, the risk of developing multiple myeloma, non-multiple myeloma hematologic malignancies, and other cancers is increased. It is of interest to note that among GD1 patients who developed cancers, the diagnosis of cancer preceded ERT in 70% of patients in keeping with many previous reports suggesting this association is an intrinsic feature of the disease in some patients. The association of GD with cancers has been reported for more than 3 decades in case-reports and small case-series [7–9], but the data are not consistent, and there are no genotype correlations or natural history studies. There are numerous case reports of multiple myeloma and GD . Analysis of data of 2742 patients within ICGG Gaucher Registry demonstrated a 6-fold increased life time risk of multiple myeloma but not of other cancers . The study suffered from incomplete ascertainment since the objective of the registry is to track treatment responses, and it comprises a predominantly younger population, i.e., 65% of patients were younger than 40 years old, an age at which cancer complications are less frequent and an age at which multiple myeloma does not occur. Findings similar to the ICGG Registry were reported in a single center study . A more recent study by Landgren et al., found no increase in overall cancer risk in a cohort of 1,525 patients identified in the VA system with an ICD code for lipidosis . However, the diagnosis of GD in this study was flawed, because patients were identified based on diagnostic codes (272.2 or 272.7) that represents all lipidoses (such as primary hypercholesterolemia and all lipid storage diseases, such as Fabry disease and Niemann-Pick disease) as well as GD .
The high incidence of sequential polyclonal gammopathy, MGUS, and multiple myeloma suggests important interactions between pathologic glycolipid-laden macrophages and the adaptive immune system in the pathogenesis of GD. This finding together with increased risk of cancers reported by us and others [7–9] as well as coexistence of multiple cancers among seven patients in our study, suggest that alterations of the cellular milieu in GD create a microenvironment conducive to malignancy.
The pathways linking GD and cancer are not known, but several features of GD may offer clues to the pathogenesis of malignancy that will undoubtedly be the focus of future scientific investigations. It has been suggested that accumulation of glucocerebroside and related lipids in tissue macrophages results in chronic stimulation of the immune system and consequent lymphoproliferation including CD1d-mediated imbalances in regulatory and NKT cells [24,28,29]. Another factor promoting malignancy in GD may be endoplasmic reticulum (ER) stress . Recent studies suggest that N370S glucocerebrosidase is likely a protein folding mutation and thus may contribute to ER stress . There are clearly several putative and likely additive or synergistic mechanisms to promote malignancy in GD.
Our findings have the potential to further improve patient monitoring and management as well as highlight areas of future scientific investigation. First, our findings underscore the importance of monitoring patients with GD for gammopathies and especially for development of multiple myeloma in patients older than 50 years. The finding of increased risk of non-myeloma hematologic malignancies and other cancers not related to GD severity emphasizes the need to have a high index of suspicion in routine monitoring adult GD patients even though their GD may appear indolent. Furthermore, our results and those of others should encourage further investigations to determine whether the carrier state for GD is a risk factor for the development of multiple myeloma. Additionally, it will be of great interest to determine whether the new generation of patients who tend to begin ERT early in the course of their GD will have an altered cancer risk.
Cancer risk data in our study of predominantly Ashkenazi Jewish patients were derived from a comparison with the US White population as well as the Israeli population. Finding valid cancer incidence rates to serve as a reference for our GD patient population or indeed for any other GD patient population is challenging. As previously stated, most cases of GD in our cohort are of Ashkenazi Jewish ancestry. This ethnic group corresponds demographically to Caucasians who are represented in the US SEER database as non-Hispanic Whites. However, this group has a variety of unique characteristics, which can influence their cancer risk estimates. Cancer rates are determined by behavioral exposures, by genetic constituents, and most likely by their combination. For example, the high prevalence of founder mutations in the BRCA gene in the Ashkenazi Jewish population leads to higher probability of developing breast cancer and to younger age at detection of this cancer . Thus, it is unclear to what degree cancer rates from the SEER program would represent the expected cancer rates in the Ashkenazi Jewish population.
To further refine our analysis, we compared cancer rates in our study to those for Jews from the National Israeli Cancer Control Center and our initial findings were confirmed. However, it is important to note that disease rates are also dictated by other parameters, such as the degree of screening and differences in health behaviors between countries. For example, the use of hormone replacement therapy in post-menopausal women and prostate specific antigen to screen for prostate cancer in Israel are much lower compared with the US (G. Rennert, unpublished observations). A comparison of cancer incidence rates between European–American born Israeli Jews and SEER non-Hispanic Whites shows the rates in the latter to be 2–3-fold higher for leading cancers, such as lung cancer and prostate cancer . Thus, such an adjustment is also imperfect.
Our study is not a prospective study. Because of heterogeneity in clinical outcomes of N370S/N370S GD and its relatively slow natural course, it is not feasible to determine an accurate natural history by prospective follow-up of a cohort of untreated patients over many years. However, despite this difficulty, it remains important to estimate the natural history for counseling, enabling optimal lifelong monitoring of disease progression, and directing costly therapies effectively. We used quantitative determination of disease severity at age of evaluation prior to commencing ERT in a cross-sectional study of consecutively evaluated patients to define the natural history of our patient population. Every patient was included and there was no selection of patients for disease severity.
In conclusion, our observational study reveals the progressive nature of GD due to N370S homozygous mutation with the brunt of the disease on the skeleton in adult patients. Uniquely, this lysosomal storage disease is associated with an increased risk of multiple myeloma and other malignancies. Adults with N370S/N370S GD should be carefully monitored for disease activity in all domains, especially the skeleton and there should be a high index of suspicion for cancers. There should be further exploration of cancer risk in GD patients representing the spectrum of genotypes, age, and disease severity, to develop an evidence-based cancer screening strategy in this population and to evaluate the eventual impact of various therapies . As with Parkinsonian syndrome  and pulmonary hypertension  associated with GD, the risk of cancer in individual GD patients in the clinic is low and therefore patients should be reassured. However, it is hoped that these emerging insights into the nature of GD will enhance delivery of sophisticated personalized medicine to this group of patients.
This is an observational study in which all patients received the standard of care. The data collection has been approved by the Institutional Review Boards of Yale University and NYU Schools of Medicine. Between 1996 and 2006, consecutive patients with GD (confirmed by demonstration of low, <10% of normal, leukocyte acid β-glucosidase activity, and genotyped as N370S/N370S or N370S/other) underwent comprehensive phenotyping by staging of severity in all disease domains. The database was maintained and developed by PKM and includes patients in the GD clinics of PKM, MB, KA, and GP. The database recorded information on age of symptom onset, diagnosis, and first comprehensive evaluation. Quantitative phenotypic data in each disease compartment at the age of evaluation were used to estimate the natural course of the disease. Statistical analysis was performed by KK with input from TT, RY, and PKM.
The range of expressivity of N370S/N370S and N370S/other GD was determined using a standard protocol for phenotyping in each GD disease domain. This included volumetric MRI measurement of the liver and spleen , a skeletal series to determine bone disease score as described by Hermann et al. [36,37], and blood tests (CBC, hepatic function, biomarkers of GD, serum immunoglobulins, and serum protein electrophoresis) . A HS of 0 denotes no skeletal disease; 1 denotes diffuse osteopenia; 2 is medullary expansion; 3 is localized destruction (osteolysis); 4 is ischemic necrosis of long bones; sclerosis; osteitis; 5 is diffuse destruction; epipheseal collapse; osteroarthrosis [36,37]. Two predominant phenotypes were evident, one marked by early onset visceral disease (age of symptomatic presentation < 18 years, spleen volume > 10 × normal, and bone disease score < 3) and the other marked by late onset skeletal disease (age of symptomatic presentation > 18 years, spleen volume < 10 × normal, and bone disease score > 3). Presence and type of cancer was also included, as there is a growing body of literature suggesting an increased incidence of multiple myeloma in patients with GD [6,7,10].
Overall disease SSI was determined as described previously . The presence of polyclonal gammopathy was determined based upon results of serum immunoglobulins and serum protein electrophoresis. Standard diagnostic criteria for monoclonal gammopathy of undetermined significance (MGUS) and multiple myeloma were applied .
GBA gene analysis was performed as previously described by PCR and DNA sequencing . Ascertainment of the presence of a 55 base pair deletion spanning the N370S mutation was performed by PCR to avoid misclassifying N370S/55 base pair deletion as N370S/N370S .
Descriptive statistics were computed for gender, age at presentation, genotype, ethnicity, mode of presentation, spleen status, treatment status, and years of follow-up. Additional descriptive statistics were calculated stratifying patients with multiple myeloma and hematologic (non-multiple myeloma) cancers.
A linear model was constructed for each of the four GD domains, specifically, liver (in multiples of normal, MN), spleen MN, HS, and SSI by age and according to N370S/N370S or N370S/other. Model fit was assessed using standard methods such as residual plots and checks for non-linearity and heteroscedasticity.
The relative risk (RR) of cancer by type and overall risk was calculated for our population and compared with the United States (US) as well as the Israeli population (because > 90% of our patients were of Ashkenazi Jewish ancestry) by attained age and gender. The probability of developing cancer was calculated using DevCAN (Probability of Developing or Dying from Cancer). The probability of developing cancers by specific attained ages was calculated using publicly available SEER (Surveillance, Epidemiology, and End Results) software of National Cancer Institute, Bethesda, MD. Same analysis was repeated using the Israeli Cancer Registry data to ensure ethnically matched control population to derive the most accurate risk estimate.
A Kaplan-Meier curve was constructed to show polyclonal gammopathy-, MGUS-, and multiple myeloma-free survival by age. The prevalence of each of these conditions was also calculated.
No adjustments for multiple hypothesis testing were made, as this study was exploratory in nature. Instead, a systematic approach was taken where increased risk of all cancers was assessed, and if significant, all other groups of cancers were studied individually.
Contract grant sponsor: NIH NIDDK; Contract grant numbers: T32DK007356, P30DK34989; Contract grant sponsor: NIDDK; Contract grant number: K24DK066306; Contract grant sponsors: the AASLD/ALF Sheila Sherlock Award, NIH CTSA/Yale Center for Clinical Investigation Scholar Award, Howard Hughes Patient-Oriented Research Pilot Project Grant, National Gaucher Foundation.
The authors thank the patients who participated in their studies.
Conflict of interest: PKM receives a research grant from Genzyme Corporation for participation in the International Gaucher Registry.