Cancer spectrum in Mexican patients with the CHEK2 p.(Leu236Pro) variant: a retrospective study

Abstract

This study aimed to characterize, for the first time, the cancer spectrum associated with the most frequent pathogenic CHEK2 variant—NM_007194.4(CHEK2):c.707T > C p.(Leu236Pro)—in Mexican individuals. Although this variant is frequently detected through multi-gene panel testing, limited data on its associated cancer risks complicates genetic counseling and surveillance strategies. We retrospectively analyzed 5,759 patients who underwent multi-gene panel testing between August 2015 and August 2024 due to suspected hereditary cancer syndromes. Among them, 58 CHEK2 p.(Leu236Pro) carriers with confirmed cancer diagnoses were identified. Geographical clustering was observed, with 81% of patients originating from central Mexico, suggesting a possible founder effect. Ten distinct clinical indications for genetic testing were identified, with hereditary breast and ovarian cancer (HBOC) syndrome being the most common (74.1%). The mean age at first diagnosis among carriers was 43.8 ± 12 years, and 61.1% of them reported a family history of cancer in first- or second-degree relatives. A second or third primary cancer occurred in 20.7% of cases. Tumors were identified in 12 anatomical sites. Breast cancer predominated (67.6%, including one male case), followed by ovarian (8.1%), prostate (6.7%), gastric (4.1%), thyroid (2.7%), and endometrial (2.7%) cancers. Lymphoma, lung, sacrococcygeal bone, colorectal, and non-melanoma skin cancers each occurred in a single patient. Significant risk association was identified only for breast, ovarian, and gastric cancers. These results highlight the need for personalized surveillance, especially for breast cancer. Incorporating CHEK2 p.(Leu236Pro) into clinical decision-making tools may enhance risk assessment in the Mexican population, but larger studies are needed to refine risk estimates and to clarify the possible founder effect.

Supplementary Information

The online version contains supplementary material available at 10.1007/s10689-025-00490-8.

Introduction

The checkpoint kinase 2 gene (CHEK2) is implicated in the predisposition to multi-organ cancers [1], including colorectal, prostate, and breast cancer [2]. CHEK2 encodes the serine/threonine protein kinase Chk2 (hereafter referred to as Chk2), which regulates the cell cycle, DNA repair, and apoptosis. Loss-of-function CHEK2 variants confer an increased risk of breast cancer [OR = 2.54 (CI: 2.21–2.91)], comparable to the ATM and PALB2 genes [3]. This association has been well documented in European and North American populations [4, 5]. Germline pathogenic variants of CHEK2 have been associated with ovarian cancer [6], non-Hodgkin lymphoma (NHL) [7], papillary thyroid carcinoma [8], and renal cancer [9], but the specific risk for each tumor remains insufficiently characterized.

The frequency of clinically relevant CHEK2 variants varies between populations and geographical regions. The variant NM_007194.4(CHEK2):c.1100del p.(Thr367fs) is frequently observed in Northern and Eastern European populations [10], but its occurrence in Mexican populations is rare. An internal analysis by the Genomic Diagnostic Laboratory (GDL) of the National Institute of Genomic Medicine identified a single case among 4,109 individuals evaluated for hereditary cancer syndromes as of August 2024. This variant was also absent from a cohort of 327 Mexican Mestizo patients assessed for cancer susceptibility with multi-gene panel testing [11].

Among Mexican breast cancer patients meeting genetic risk criteria, the most frequently identified CHEK2 variant was NM_007194.4(CHEK2):c.707T > C p.(Leu236Pro). This variant was detected in seven of 204 women selected between 2016 and 2020, based on the National Comprehensive Cancer Network (NCCN) criteria, representing a frequency of 3.4% [12]. It has also been reported in cohorts of patients of Latino or Mexican ancestry with breast and ovarian cancer [4, 11, 13], with allele frequencies ranging from 0.9 to 2.3%. Additionally, this variant has been identified in Mexican and Latino pediatric patients with acute lymphoblastic leukemia (GDL local database, unpublished data) [14]. According to the gnomAD exome and genome database, the global allele frequency of p.(Leu236Pro) is 0.006444%, indicating its rarity. It has only been observed in the admixed American population subset of gnomAD, with a reported frequency of 0.17% (gnomAD v4.1.0). It is absent from the Mexican database on the Franklin by Genoox platform (https://franklin.genoox.com), which includes 470 non-cancer individuals as of December 2024.

Similar to other CHEK2 missense variants, NM_007194.4(CHEK2):c.707T > C p.(Leu236Pro) was initially classified as a variant of uncertain significance [15]. In 2019, a yeast-based functional study suggested that p.(Leu236Pro) impaired DNA replication and blocked mitotic entry [16]. In July 2019, we reclassified the variant as likely pathogenic based on an integrative assessment of clinical, functional, and epidemiological evidence (ClinVar variation ID: 142448). Subsequent functional studies using KAP1 phosphorylation and Chk2 autophosphorylation analyses confirmed the deleterious effect of p.(Leu236Pro) on Chk2 phosphorylation activity [17].

In low- and middle-income countries, epidemiological studies evaluating the specific cancer risks of pathogenic or likely pathogenic variants are scarce [2]. The NM_007194.4(CHEK2):c.707T > C p.(Leu236Pro) variant is frequently identified in Mexican patients undergoing multi-gene panel testing for cancer risk, but data on its cancer-specific risk and associated tumor spectrum remain limited. Risk management for p.(Leu236Pro)-positive patients is primarily based on cancer risk estimates derived from populations in which NM_007194.4(CHEK2):c.1100del (p.Thr367fs) is the predominant variant. This lack of data complicates genetic counseling and hinders the implementation of organ-specific surveillance strategies. Analyzing population-specific CHEK2 genetic variants is crucial for comprehensive risk assessment and management [18, 19]. To address this limitation, we assessed the spectrum of cancers identified in a cohort of Mexican patients carrying the p.(Leu236Pro) variant and their affected relatives, and evaluated whether a statistically significant risk association exists between this variant and the identified tumor types.

Methods

Study population

NM_007194.4(CHEK2):c.707T > C p.(Leu236Pro)-positive patients were selected from a group of 5,759 individuals who underwent multi-gene panel testing for suspected hereditary cancer predisposition. These patients belonged to three distinct cohorts: (1) 204 Mexican women diagnosed with breast or ovarian cancer who met at least one genetic risk criterion defined by the National Comprehensive Cancer Network (NCCN) guidelines (versions 2016.1 through 2022.1); (2) 1,650 oncology patients referred to the Hereditary Cancer Clinic at the National Cancer Institute in Mexico; and (3) 3,905 individuals referred to the GDL for genetic testing. The distribution of diagnoses among the 5,759 individuals is detailed in Supplementary Material 1 (Table S1). A genetic counselor or oncologist performed the clinical evaluation process and referred the patients for genetic testing. Pre- and post-genetic testing counseling was offered by a certified medical geneticist to all patients. A family history of cancer was classified as positive if at least one first- or second-degree relative had been diagnosed with any malignancy, irrespective of cancer type or age at diagnosis (excluding non-melanoma skin cancer).

Variant identification

A multi-gene panel test using next-generation sequencing (NGS) was performed for all individuals. Genomic DNA was extracted from EDTA-anticoagulated peripheral blood. The multi-gene panels used were either the TruSight Cancer version 1 or 2 (Illumina, San Diego, CA, USA) or the Hereditary Cancer Solution panel (Sophia Genetics SA, Saint-Sulpice, Switzerland). Library preparation was performed according to the manufacturer’s instructions for each kit. Details of the kits, including gene content and limitations, are provided in Supplementary Material 2. Sequencing was performed on either the MiSeq or NextSeq 500 instrument (Illumina, San Diego, CA, USA), depending on sample throughput. Bioinformatic analysis was conducted using the commercial Sophia DDM^® platform for all kits. This analysis enabled the detection of single-nucleotide variants (SNVs), insertions/deletions (indels), and copy-number variants (CNVs) in all panels. Details on the bioinformatic analysis and quality control metrics are summarized in Supplementary Material 2. All CNV positive cases were confirmed using orthogonal methods. Variant classification followed the guidelines of the American College of Medical Genetics and Genomics (ACMG) [20], incorporating evidence from databases and the scientific literature.

Statistical analysis

Continuous variables were summarized as means and standard deviations, while categorical variables were expressed as frequencies and percentages. To assess the association between the variant and tumor type, a case-control study was conducted comparing the distribution of CHEK2 p.(Leu236Pro) among cancer cases stratified by tumor type and Mexican individuals from the Mexico City Prospective Study (MCPS) (https://rgc-mcps.regeneron.com/home) [21]. Two separate comparisons were performed using MCPS controls: (1) the entire cohort of Mexican individuals in the MCPS database (N = 138,200), and (2) the subset classified as “Indigenous Mexican” (N = 93,978). Fisher’s exact test was applied for comparisons involving small sample sizes (< 5), while chi-square tests were used otherwise. A two-tailed p-value < 0.05 was considered statistically significant. Odds ratios (ORs) with 95% confidence intervals (CIs) were also calculated. All analyses were performed using Epi Info™ version 7.2 software.

Results

Fifty-eight NM_007194.4(CHEK2):c.707T > C p.(Leu236Pro)-positive patients with documented cancer diagnoses were identified; 91.3% (53/58) were female. All of them self-reported as Mexican Mestizo. Clinical characteristics are summarized in Table 1. Additional details regarding cancer in relatives and their relationship to the proband are provided in the Supplementary Material (Table S2). As shown in Fig. 1, most patients (47/57, 82.4%) were born in the central region of Mexico (Mexico City, State of Mexico, Hidalgo, Puebla, Morelos, Guanajuato, and Tlaxcala). The remaining individuals were born in the northern (Chihuahua and Coahuila) or southeastern (Veracruz) regions of the country. For one case, the state of birth was unknown.

Table 1.

Clinical features of patients carrying the NM_007194.4(CHEK2):c.707T > C p.(Leu236Pro) variant

No.	Sex at Birth		Primary Tumor Site (Age)	Primary Tumor Histology	Second Primary Tumor Site (Age)	Second Primary Tumor Histology	Third Primary Tumor Site /Histology (Age)	Indication for Germline Testing/Panel&	Family History of Cancer#	Other Genomic Findings*	State of Birth
1	F		Bone (sacrum-coccyx) (32)	Sarcoma	Breast (41)	IDC	Breast/IDC (47)	Li-Fraumeni syndrome/3	ND	Yes, PALB2 (P)	State of Mexico
2	F		Breast (59)	ND	Skin (62)	BCC	Uterus: Endometrial (Moderately Differentiated AC) (82)	HBOC/3	Yes	No	Morelos
3	F		Breast (38)	IDC	Ovary (53)	Mature teratoma	Breast/IDC (68)	HBOC/3	Yes	No	State of Mexico
4	F		Breast (21)	IDC	–			HBOC/3	Yes	No	Hidalgo
5	F		Breast (29)	IDC	Ovary (40)	H-GS		HBOC/1	Yes	No	Chihuahua
6	F		Breast (29)	IDC	–			HBOC/3	No	No	State of Mexico
7	F		Breast (32)	IDC	Ovary (53)	H-GS		HBOC/1	No	No	State of Mexico
8	F		Breast (32)	IDC				HBOC/2	No	No	Mexico City
9	F		Breast (33)	IDC				HBOC/3	Yes	No	State of Mexico
10	F		Breast (34)	IDC				HBOC/3	No	No	Mexico City
11	F		Breast (34)	IDC				HBOC/3	No	No	State of Mexico
12	F		Breast (35)	IDC				HBOC/1	No	No	Coahuila
13	F		Breast (35)	IDC				HBOC/3	Yes	No	State of Mexico
14	F		Breast (37)	IDC				Cascade testing/2	Yes	No	Puebla
15	F		Breast (37)	IDC				HBOC/3	No	No	Mexico City
16	F		Breast (37)	IDC				HBOC/3	No	No	Tlaxcala
17	F		Breast (37)	IDC				Cowden syndrome/3	No	No	Mexico City
18	F		Breast (37)	IDC				HBOC/3	Yes	No	Mexico City
19	F		Breast (38)	IDC	Breast (45)	IDC		HBOC/3	No	No	State of Mexico
20	F		Breast (38)	IDC				HBOC/1	No	No	Mexico City
21	F		Breast (39)	IDC	Breast (51)	IDC		HBOC/1	Yes	Yes, BRCA2 (P)	Mexico City
22	F		Breast (39)	IDC				HBOC/3	Yes	No	Mexico City
23	F		Breast (40)	IDC				HBOC/3	Yes	No	Mexico City
24	F		Breast (41)	IDC	Ovary (61)	H-GS		HBOC/1	Yes	No	Hidalgo
25	F		Breast (41)	IDC				HBOC/3	Yes	No	State of Mexico
26	F		Breast (41)	IDC				HBOC/3	Yes	No	State of Mexico
27	F		Breast (42)	IDC	Breast (42)	IDC		HBOC/3	Yes	No	Mexico City
28	F		Breast (42)	IDC				HBOC/2	Yes	Yes, Homozygous CHEK2 p.(Leu236Pro)	Guanajuato
29	F		Breast (42)	IDC				HBOC/3	Yes	No	Mexico City
30	F		Breast (43)	IDC	Breast (56)	IDC		HBOC/3	Yes	No	State of Mexico
31	F		Breast (43)	IDC				HBOC/2	Yes	No	Mexico City
32	F		Breast (46)	IDC				HBOC/3	No	No	Mexico City
33	F		Breast (47)	IDC	Thyroid (47)	Papillary		HBOC/3	Yes	No	State of Mexico
34	F		Breast (47)	IDC				HBOC/3	Yes	Yes, ATM (P)	Mexico City
35	F		Breast (47)	IDC				HBOC/3	No	No	Tlaxcala
36	F		Breast (47)	LIC				HBOC/3	No	No	State of Mexico
37	F		Breast (51)	ES				HBOC/2	Yes	No	Mexico City
38	F		Breast (51)	IDC				HBOC/2	Yes	No	Mexico City
39	F		Breast (57)	IDC				HBOC/3	No	No	Mexico City
40	F		Breast (63)	IDC				HBOC/3	Yes	Yes, CDKN2A (LP)	Mexico City
41	F		Breast (75)	LIC				HBOC/3	Yes	No	Guanajuato
42	M		Breast (50)	IDC				HBOC/3	Yes	No	Mexico City
43	F		Colorectal (33)	AC				Lynch syndrome/3	No	Yes, ATM (P)	Hidalgo
44	M		Prostate (59)	AC				HBOC/3	Yes	No	Tlaxcala
45	M		Prostate (65)	AC				Treatment selection/3	No	No	State of Mexico
46	M		Prostate (68)	AC				Treatment selection/3	No	No	Guanajuato
47	M		Prostate (71)	AC				Treatment selection/3	ND	No	State of Mexico
48	M		Prostate (77)	AC				Treatment selection/3	ND	No	ND
49	F		Gastric (40)	DSRCC				Lynch syndrome/3	Yes	No	State of Mexico
50	F		Gastric (44)	PD AC with SRC				Hereditary diffuse gastric cancer/3	Yes	No	Mexico City
51	F		Gastric (37)	DGC				Hereditary diffuse gastric cancer/3	No	No	Hidalgo
52	F		Ovary (36)	EC				HBOC/3	Yes	No	Mexico City
53	F		Ovary (47)	AC				HBOC/3	ND	No	Mexico City
54	F		Thyroid (45)	Medullary				MTC/3	No	No	Veracruz
55	F		Brain (46)	Meningioma				CNS Tumor/3	Yes	No	Mexico City
56	F		Lung (68)	AC				Lung cancer/3	ND	No	State of Mexico
57	F		Uterus (Endometrial) (30)	AC				Lynch syndrome/3	Yes	Yes, BRCA1 (P)	Mexico City
58	F		Oral cavity (40)	NHL	Breast (68)	IDC		HBOC/3	No	No	State of Mexico

(AC) Adenocarcinoma; (BCC) Basal Cell Carcinoma; (DGC) Diffuse Gastric Carcinoma; (DSRCC) Diffuse Signet Ring Cell Carcinoma with Carcinomatosis; (EC) Endometrioid Carcinoma; (ES) Eccrine Syringoid; (F) Female; (HBOC) Hereditary Breast and Ovarian Cancer Syndrome; (H-GS) High-Grade Serous; (IDC) Invasive Ductal Carcinoma; (LIC) Lobular Invasive Carcinoma; (LP) Likely Pathogenic; (M) Male; (MTC) Medullary Thyroid Cancer; (ND) Not Determined; (NHL) Non-Hodgkin Lymphoma; (P) Pathogenic; (PD) Poorly Differentiated; (SRC) Signet Ring Cells

^&Panel used for genetic testing in this patient: 1-TruSight Hereditary Cancer Panel (Version 1), Illumina; 2-TruSight Hereditary Cancer Panel (Version 2), Illumina; 3-Hereditary Cancer Solution v2, Sophia Genetics

^#Details about the family history of cancer in each case can be found in Supplementary Material 1, Table S2, column M

^*Other genomic findings: The identified variant is listed in Supplementary Material 1, Table S2, column O

Fig. 1.

Geographic distribution by state of birth of individuals carrying the NM_007194.4(CHEK2):c.707T > C p.(Leu236Pro) variant. Lines on the map indicate administrative boundaries between Mexican states. Shades of green represent the number of p.(Leu236Pro)-positive patients per state (shown in parentheses). The state of birth was unknown for one individual

Multi-gene panel testing was performed for ten distinct clinical indications. Hereditary breast and ovarian cancer (HBOC) accounted for 43 of 58 cases (72.8%), followed by treatment selection in metastatic castration-resistant prostate cancer (4/58, 6.7%). Less frequent indications included early-onset central nervous system tumors, suspected Lynch syndrome, diffuse gastric cancer, suspected Cowden syndrome, Li-Fraumeni syndrome, medullary thyroid carcinoma, and cascade testing (i.e., testing of at-risk relatives once the pathogenic variant was identified in the index case).

Spectrum of identified tumors

All patients in the cohort had a prior cancer diagnosis. Most patients (79.3%, 46/58) presented with a single primary tumor, while a smaller subset (15.5%, 9/58) developed a second primary tumor, and three patients (5.2%, 3/58) had a third primary malignancy. The mean age at initial cancer diagnosis was 43.8 ± 12 years (range: 21–77). Furthermore, 61.1% of patients reported a family history of cancer in first- or second-degree relatives. Family history information was unavailable for four patients.

Among the 58 patients, a total of 73 primary tumors were identified, including bilateral, synchronous, and metachronous cancers. The tumors were distributed across 12 anatomical sites (Table 1; Fig. 2). The most prevalent tumor site was the breast (67.1%, 49/73), which included one case of male breast cancer (MBC) and five cases of bilateral breast cancer. Other primary sites included the ovary (8.2%, 6/73), prostate (6.8%, 5/73), stomach (4.1%, 3/73), thyroid (2.7%, 2/73), and endometrium (2.7%, 2/73). Lymphoma, lung, sacrococcygeal bone, colorectal, and non-melanoma skin cancers were each observed in one patient. (Table 1; Fig. 2).

Fig. 2.

Anatomical distribution of malignant tumors in the NM_007194.4(CHEK2):c.707T > C p.(Leu236Pro)-positive patients. The silhouette depicts male and female organs to illustrate tumor locations

Among the 49 breast tumors, 83.7% were the first primary malignancy, with a mean age at diagnosis of 41.2 ± 9.4 years (range: 21–75 years). Invasive ductal carcinoma (IDC) was the predominant histologic subtype, accounting for 93.5% (43/46) of tumors. In contrast, invasive lobular carcinoma (ILC) and syringoid eccrine carcinoma (SEC) were rare, comprising 4.3% (2/46) and 2.1% (1/46), respectively. Histologic data were unavailable for three tumors.

Among patients who developed a second primary cancer, IDC of the breast was the most common (n = 6), followed by high-grade serous ovarian carcinoma (n = 3). Three patients developed a third primary tumor: two with breast cancer (IDC) and one with moderately differentiated endometrial carcinoma.

Other genomic findings

Pathogenic or likely pathogenic variants in other cancer-predisposition genes were identified in six patients: ATM (n = 2), BRCA1 (n = 1), BRCA2 (n = 1), CDKN2A (n = 1), and PALB2 (n = 1) (Table 1). One patient carried CHEK2 p.(Leu236Pro) in a homozygous state. These additional variants are listed in Supplementary Material 1, Table S2, Column O.

Case-control analysis

The results of the comparative analysis using the entire MCPS cohort are shown in Table 2. A significantly increased risk was identified for breast (OR = 3.2, 95% CI [2.4–4.3]), ovarian (OR = 4.2, 95% CI [1.8–9.5]), gastric (OR = 9.9, 95% CI [3.1–32.2]), and thyroid cancers (OR = 4.6, 95% CI [1.1–18.8]). The associations remained statistically significant only for breast, ovarian, and gastric cancers when the Indigenous Mexican subpopulation was used as the control group (Supplementary Table S3). No significant associations were detected for lung, colorectal, endometrial, non-melanoma skin, prostate, or brain malignancies.

Table 2.

Case-control comparative analysis of CHEK2 p.(Leu236Pro) carriers using the MCPS cohort as the control group

Tumor Site	Carriers (cases)	Total Cases	Frequency (%)	Carriers (controls)	Total Controls	Frequency in Controls	OR	95% CI	p-Value
Breast	49	2,223	2.2	961	138,200	0.7	3.2	2.4–4.3	< 0.001
Ovary	6	209	2.9	961	138,200	0.7	4.2	1.8–9.5	< 0.001
Lung	1	249	0.4	961	138,200	0.7			ns
Colorectal	1	214	0.5	961	138,200	0.7			ns
Endometrial	2	119	1.7	961	138,200	0.7			ns
Skin&	1	39	2.6	961	138,200	0.7			ns
Brain	1	43	2.3	961	138,200	0.7			ns
Gastric	3	46	6.5.	961	138,200	0.7	9.9	3.1–32.2	< 0.001
Thyroid	2	64	4.3	961	138,200	0.7	4.6	1.1–18.8	0.02
Prostate	5	317	1.7	961	138,200	0.7			ns
Others*	2	753	0.2

The 1,483 non-cancer patients are not included in Table 2

^&Non-melanoma, OR: Odds Ratio, CI: Confidence Interval, ns: not significant

^*No association was evaluated in the “Others” category because it comprises a heterogeneous group of tumors

Discussion

This study describes the tumor spectrum observed in a cohort of cancer patients harboring the NM_007194.4(CHEK2):c.707T > C p.(Leu236Pro) variant. Among the 58 individuals analyzed, 73 primary malignancies were identified across 12 anatomical sites (Fig. 2). Genetic testing was performed for 10 distinct clinical indications. HBOC was the most common indication, followed by treatment selection in metastatic castration-resistant prostate cancer and personal or family history suggestive of Lynch syndrome. The mean age at initial cancer diagnosis was 43.8 ± 12 years (range: 21–77 years), suggesting a predisposition to early-onset disease, especially breast cancer.

Characterizing the tumor spectrum associated with this variant is particularly relevant for the Mexican population, as CHEK2 p.(Leu236Pro) is the most frequently identified pathogenic CHEK2 variant in Mexican individuals undergoing multi-gene panel testing for hereditary cancer. Based on the internal database of the Genomic Diagnostic Laboratory (GDL) at the National Institute of Genomic Medicine (INMEGEN), this variant accounts for approximately 56% of the pathogenic CHEK2 alleles detected in patients tested for suspected HBOC, suggesting a founder effect in Mexico (unpublished data, GDL, INMEGEN).

Comparative analysis of the variant’s frequency across different ethnic groups supports the hypothesis that CHEK2 p.(Leu236Pro) may have originated in the central region of Mexico within an indigenous population. Data from the Mexico City Prospective Study (MCPS) [21] show that the variant is present exclusively in individuals genetically classified as “Indigenous Mexican” (allele frequency 0.005113) and is absent in Mexican individuals of European or African ancestry. It is also absent from other Latin American subpopulations, including Peruvians from Lima, Colombians from Medellín, Puerto Ricans, and Mexican mestizos from Los Angeles (https://rgc-mcps.regeneron.com/rsid/rs587782471). This observation is consistent with gnomAD data, which reports the variant only in “admixed American populations” (0.17%) and not in other groups (gnomAD v4.1) (https://gnomad.broadinstitute.org/variant/22-28711994-A-G).

To further elucidate the ethnic and geographical origin of the CHEK2 p.(Leu236Pro) variant and its potential founder effect, future studies should incorporate haplotype analysis, assessments of the variant’s geographical distribution, and evaluations of ancestry-informative markers from specific Mexican indigenous groups. Such investigations would provide valuable insights into the demographic history underlying this variant and its implications for cancer risk assessment in the Mexican population.

CHEK2 is classified as a moderate-risk gene for breast cancer [1]. An absolute breast cancer risk of 23–27% is outlined in the NCCN Guidelines for Genetic/Familial High-Risk Assessment: Breast, Ovarian, Pancreatic, and Prostate [22]. Its inclusion in breast cancer multi-gene risk panels highlights its clinical importance [23].

The personal and family cancer histories of individuals carrying the CHEK2 p.(Leu236Pro) variant resembled those typically seen in BRCA1/2-related HBOC. This similarity has been previously reported for CHEK2-positive individuals [4, 6, 23, 24]. In our cohort, most individuals (72.9%) were diagnosed with breast cancer at an early age (mean 41.2 years), and 63.4% reported a positive family history. The majority were referred for genetic testing due to suspected HBOC, including one case of male breast cancer (MBC). Our case-control analysis supports this similarity, with significant associations observed between CHEK2 p.(Leu236Pro) and both breast cancer (OR = 3.2, 95% CI 2.4–4.3) and ovarian cancer (OR = 4.2, 95% CI 1.8–9.5). These associations remained statistically significant even when using a control group composed exclusively of Indigenous Mexican individuals, who have a higher frequency of the variant (Supplementary Material 1, Table S3).

Three breast cancer patients carried an additional pathogenic variant in a cancer gene: CDKN2A (patient ID40), BRCA2 (patient ID21), and ATM (patient ID34). While BRCA2 and ATM are associated with high and moderate breast cancer risk, respectively, the role of CDKN2A in breast cancer risk remains unclear [22]. To further clarify the association between CHEK2 p.(Leu236Pro) and breast cancer risk, the three cases were excluded from the case-control analysis. The results continued to support a significant association (OR = 3.02, 95% CI [2.2-4], p < 0.00001).

The association between pathogenic and likely pathogenic variants of CHEK2 and MBC remains controversial [25, 26]. In a cohort of 715 Caucasian MBC patients, CHEK2 pathogenic variants were present in 4.1% of patients (OR = 3.7, p < 6.24 × 10^− 24), representing the second most common predisposition gene after BRCA2 [27]. Hallamies et al. [28] analyzed 68 MBC cases from the Finnish population and showed that CHEK2 c.1100delC was associated with an increased risk of MBC (OR = 4.47, 95% CI 1.51–13.18, p = 0.019). A meta-analysis also supported the association between CHEK2 c.1100delC and an increased risk of MBC (OR = 3.13, 95% CI 1.94–5.07) [29]. Other CHEK2 variants have also been investigated, with some studies identifying a significant association between MBC and pathogenic or likely pathogenic variants [25, 30]. However, other studies do not support an association between CHEK2 variants and MBC risk [31–36]. The risk of MBC may vary depending on the specific pathogenic CHEK2 variant.

In our cohort, we identified a female breast cancer patient who was homozygous for the CHEK2 p.(Leu236Pro) variant (patient ID28; Table 1 and Supplementary Table S2). She was diagnosed with IDC at age 42 and had a positive family history of cancer. Two of her sisters were diagnosed with breast cancer before age 50, but they were not genetically tested to confirm the presence of the variant (Table 1 and Supplementary Material Table S2). The clinical consequences of homozygosity for CHEK2 pathogenic variants were initially investigated for NM_007194.4(CHEK2):c.1100del p.(Thr367fs), due to its high carrier frequency in European populations. Findings suggest that homozygosity is associated with a higher breast cancer risk than heterozygosity in female carriers [37]. Additional studies on patients homozygous or compound heterozygous for other CHEK2 pathogenic or likely pathogenic variants indicate increased susceptibility to multiple malignancies and earlier onset in both sexes [19, 38]. Multiple cytogenetic anomalies in peripheral lymphocytes have been reported in such patients; however, karyotyping was not performed in our case [39]. Estimating lifetime risk in CHEK2 homozygous or compound heterozygous individuals remains challenging due to the small sample sizes in the existing studies. Therefore, the NCCN Guidelines recommend incorporating both personal and family history into cancer risk management decisions [40].

Ovarian cancer was a recurrent tumor type among CHEK2 p.(Leu236Pro) carriers, occurring as a primary tumor in two cases and as a second primary tumor in four, with most being serous adenocarcinomas (4/6 cases). A significant association between CHEK2 p.(Leu236Pro) and ovarian cancer was identified in our cohort (OR = 4.2, 95% CI [1.8–9.5]). No additional pathogenic variants related to ovarian cancer risk were identified in these six patients. Analyses of large European and North American cohorts have identified carriers of pathogenic or likely pathogenic CHEK2 variants among ovarian cancer patients [41, 42]. Additionally, a small study in patients of Hispanic ancestry reported ovarian cancer in individuals carrying other pathogenic CHEK2 variants [11]. However, the ACMG and NCCN guidelines (v2.2025) do not currently consider CHEK2 as a gene conferring increased ovarian cancer risk [2, 40]. Prospective studies in diverse populations are required to clarify the contribution of specific pathogenic CHEK2 variants to ovarian cancer susceptibility.

Prostate cancer was the third most frequent tumor type in our study. Among the affected patients, 4 of 5 had metastatic castration-resistant prostate cancer diagnosed between the ages of 65 and 77 and underwent genetic testing to determine eligibility for poly(ADP-ribose) polymerase (PARP) inhibitors. In one case, HBOC was suspected because the patient’s sister had breast cancer at age 32. (Table 2 and Supplementary Material 1, Table S2). Pathogenic germline CHEK2 variants have been associated with increased prostate cancer risk. In one study, CHEK2 pathogenic variants were identified in 2.7% of 1,022 prostate cancer patients, compared to a population prevalence of 1.4% (RR: 1.9, 95% CI: 1.3–2.8, p < 0.001) [43]. A multicenter cross-sectional study of 384 patients with metastatic castration-resistant prostate cancer from 11 cancer centers in seven Latin American countries, including Mexico, identified CHEK2 as the most frequently mutated gene (1%), comparable to BRCA2 (0.8%) and ATM (0.8%) [44].

Prostate cancer risk associations have also been reported for other pathogenic CHEK2 variants. A meta-analysis identified an association with c.1100delC (OR = 3.2; 95% CI: 1.85–5.9) and p.(Ile157Thr) (OR = 1.80; 95% CI: 1.5–2.1) [45]. Additionally, the c.349 A > G (p.Arg117Gly) variant was associated with prostate cancer in Portuguese patients with early-onset disease or a family history (OR = 1.9; 95% CI: 1.1–3.2) [46]. In contrast, no statistically significant association was observed between the CHEK2 p.(Leu236Pro) variant and prostate cancer risk in our case-control analysis. However, the frequency of CHEK2 p.(Leu236Pro) carriers was higher in the patient group than in the control group, suggesting a trend toward increased risk (1.7% of cases vs. 0.7% of controls) (OR = 2.2 95% CI: [0.94–5.4]; p = 0.05, χ² = 3.5, data not shown in Table 2). Analyzing a larger cohort, including men with early-stage prostate cancer or younger age at diagnosis, may help clarify whether an associated risk exists in specific subgroups.

Although endometrioid carcinoma (EC) is the most common malignancy of the female reproductive tract, it was not a recurrent tumor type among CHEK2 p.(Leu236Pro) carriers in our cohort, with one case occurring as a first tumor and another as a second primary malignancy. No definitive association has been established between CHEK2 pathogenic variants and EC risk; similarly, our results showed no statistical association. A multi-gene panel analysis in EC patients reported CHEK2 as the most frequently mutated gene after excluding cases with pathogenic variants in the mismatch repair genes MLH1, MSH2, MSH6, and PMS2 [47]. Conversely, an analysis of EC risk associated with the CHEK2 p.(Ile157Thr) variant in a cohort of 268 European patients and 449 female controls found no association. However, the variant was more frequently observed in EC patients over 75 years of age and those with deep myometrial invasion [48]. Depending on the specific CHEK2 pathogenic variant, heterogeneity in EC risk may exist. In our cohort, the patient with EC as a primary tumor was diagnosed with adenocarcinoma at age 30 and referred for genetic testing due to suspected Lynch syndrome. In addition to CHEK2 p.(Leu236Pro), a loss-of-function pathogenic variant in BRCA1 was identified, which likely explains the family history of cancers in this patient and may have contributed to EC development [49]. Given the clinical overlap between Lynch syndrome and other hereditary cancer syndromes, multi-gene panel testing provides a more comprehensive approach than syndrome-specific testing. This strategy enables the detection of additional clinically actionable mutations, as illustrated by this patient [47].

Previous NCCN guidelines reported a 5–10% increased risk of colorectal cancer in individuals with pathogenic or likely pathogenic CHEK2 variants. Additionally, a colon cancer phenotype has been described in CHEK2 c.1100delC carriers, leading some authors to refer to it as ‘hereditary breast and colorectal cancer [50]. However, the 2024 NCCN Genetic/Familial High-Risk Assessment: Colorectal, Endometrial, and Gastric guidelines now state that individuals with CHEK2 pathogenic or likely pathogenic variants are not at increased risk for colorectal cancer [40, 51]. No significant association with colon cancer was found in our study. The only case of colon adenocarcinoma occurred in a 33-year-old patient who also carried a pathogenic ATM variant. This variant may have contributed to the early-onset disease, as ATM is associated with a low-to-moderate increased risk of colorectal and other cancers [52, 53]. No pathogenic variants or variants of uncertain significance were identified in Lynch syndrome genes or other colorectal cancer predisposition genes in this patient.

Additionally, gastric and thyroid cancers were observed in Mexican patients carrying the CHEK2 p.(Leu236Pro) variant. Several studies have suggested a role for pathogenic CHEK2 variants in gastric cancer predisposition [54–56]. In a Polish population, an increased risk (OR = 1.6; p = 0.004) was reported [57], and a separate study found a stronger association for the loss-of-function CHEK2 variant NM_007194.4:c.444 + 1G > A (OR = 3.5) [1]. A literature review using natural language processing also identified an association between CHEK2 variants and gastric cancer [55]. Two studies of Chinese patients with stomach cancer identified CHEK2 as one of the most frequently mutated genes [56, 58]. Our findings also support an association between CHEK2 p.(Leu236Pro) and gastric cancer (OR = 9.9; 95% CI: 3.1–32.2). These results contrast with a Japanese study that found no pathogenic CHEK2 variants in gastric cancer patients. However, the limited sample size and use of single-strand conformation polymorphism analysis may have affected its results [54].

The association between thyroid cancer and germline pathogenic variants in CHEK2 remains uncertain. A case-control study reported an increased risk for papillary carcinoma (OR = 4.54; p = 0.0116) and a higher risk among carriers of CHEK2 loss-of-function alleles (OR = 5.7; p = 0.006) [8, 59]. However, thyroid cancer was infrequent in our cohort, with only one case of medullary carcinoma as a primary tumor and one case of papillary carcinoma as a second primary tumor. In the case-control analysis, the association between CHEK2 p.(Leu236Pro) and thyroid cancer was not significant when compared with the “Indigenous Mexican” control group (Supplementary Table S3). In the case of medullary thyroid carcinoma, pathogenic variants or variants of uncertain significance in the coding region or exon–intron boundaries of RET were excluded. These findings align with a recent review concluding that the risk of thyroid cancer associated with CHEK2 is low and that routine thyroid cancer screening is not currently recommended [60].

Non-melanoma skin cancer, non-Hodgkin lymphoma (NHL), sacrococcygeal bone cancer, and lung cancer were each observed in a single patient. A moderately increased risk of NHL and poorer progression-free survival have been suggested in CHEK2 carriers [7]. However, the role of CHEK2 in the development of NHL and Hodgkin lymphoma remains unclear. Although some studies have explored the risk of melanoma and non-melanoma skin cancers in CHEK2 carriers [61], current evidence is insufficient to draw definitive conclusions. Further research is warranted to clarify the risk associated with these tumor types.

The patient with a sacrococcygeal bone tumor developed three primary malignancies, with breast cancer diagnosed as the second and third tumors. In addition to CHEK2 p.(Leu236Pro), a pathogenic PALB2 variant was detected, fulfilling the criteria for multi-locus inherited neoplasia allele syndrome (MINAS). Approximately 28% of reported MINAS cases develop multiple primary tumors, and some exhibit atypical tumor phenotypes. The tumor spectrum in MINAS generally reflects the independent effects of the germline-mutated cancer genes. However, the occurrence of unusual phenotypes or multiple primary tumors, as in our patient, may indicate complex interactions between coexisting mutations [62].

Lung cancer was observed in only one patient, and no significant association between CHEK2 p.(Leu236Pro) and lung cancer risk was identified. However, recent evidence suggests that pathogenic CHEK2 variants may contribute to lung cancer predisposition. A large retrospective study of 7,788 lung cancer patients reported a significantly higher prevalence of pathogenic variants in BRCA2, ATM, CHEK2, and BRCA1 compared with controls [63]. Additionally, cross-sectional cohort study of 201 Mexican patients with lung adenocarcinoma, selected based on a family history of lung cancer, early-onset disease, a history of never or light smoking, or actionable genomic alterations, identified pathogenic CHEK2 variants in 6.9% of positive cases (3/43), including one CHEK2 p.(Leu236Pro) carrier [64]. Germline testing is not currently recommended for all lung cancer patients [19], except for untreated carriers of the EGFR p.(T790M) variant [65]. However, prospective studies involving germline testing of lung cancer patients from different ethnic backgrounds may help clarify the contribution of the CHEK2 gene and specific pathogenic variants to lung cancer risk.

Limitations

This retrospective cross-sectional study has inherent limitations, including a small sample size of CHEK2 p.(Leu236Pro)-positive cases, and it exclusively includes individuals of Mexican ancestry. As most genetic testing was performed for HBOC indications, selection bias likely led to overrepresentation of breast cancer cases and underestimation of tumors at other anatomical sites. The occurrence of prostate cancer among CHEK2 p.(Leu236Pro) carriers may also be underestimated because the cohort primarily comprised metastatic castration-resistant cases. The frequency of CHEK2 p.(Leu236Pro) in early-stage prostate cancer or younger men has not yet been evaluated. Moreover, CHEK2 p.(Leu236Pro) is considered a moderate-to low-penetrance variant, and several patients also carried pathogenic variants in high-penetrance cancer genes, complicating efforts to establish clear associations. Environmental and lifestyle factors, which may significantly influence cancer risk, were not evaluated. The absence of a non-carrier control group limited the contextualization of the observed tumor spectrum relative to background cancer incidence in the general population. Prospective follow-up of CHEK2 p.(Leu236Pro) carriers, particularly among unaffected family members, is warranted to refine the tumor spectrum and better characterize the associated cancer risks.

Conclusions

This study provides the first comprehensive characterization of the tumor spectrum associated with the NM_007194.4(CHEK2):c.707T > C p.(Leu236Pro) variant. Our findings demonstrate a statistically significant increased risk of breast, ovarian, and gastric cancers in carriers of this variant. No significant association was observed for other malignancies, although further studies are warranted to clarify a potential link with prostate cancer.

Incorporating these findings into clinical practice could improve early cancer detection strategies for carriers, particularly in the context of breast cancer. Prospective population-based studies are required to better define the overall cancer risk associated with CHEK2 p.(Leu236Pro). Additionally, haplotype analysis combined with ancestry-informative markers from Indigenous Mexican populations may help elucidate the variant’s ethnic origin and potential founder effect. Such research would provide valuable insights into the demographic history of this variant and its implications for population-specific cancer risk assessment in individuals of Mexican ancestry.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

LLFL and CAV carried out the study conception and design. Material preparation was performed by LLFL, RMAG, MADFH, VZFO, RERC, CMG and MJO. Data collection was conducted by LLFL, CAV, and ECMC. Data analysis was performed by LLFL, CAV, JGS, and PAAM. The first draft of the manuscript was written by LLFL, SPR and CAV, and all authors commented on previous versions of the manuscript. All authors read and approved the final manuscript.

Funding

The authors did not receive support from any organization for the submitted work.

Data availability

No datasets were generated or analysed during the current study.

Declarations

Competing interests

The authors declare no competing interests.

Ethics approval

All procedures involving human participants in this study were conducted per the ethical standards of the institutional and/or national research committee, the 1964 Helsinki Declaration, and its subsequent amendments or comparable ethical standards. This study was approved by the Research and Bioethics Committee of the National Institute of Genomic Medicine in Mexico (approval numbers: 13/2017/E, 22/2018/I, and CEI2017/17).

Informed consent

All patients signed a written informed consent for germline multi-gene panel testing and agreed to the use of their data for research and statistical purposes. All participants received pre- and post-test genetic counseling by a certified clinical geneticist.