SMAD4: A case-based review of the literature and current treatment options

CASE SUMMARY

History and examination

A 62-year-old man from Mumbai, with a 30-year history of chewing tobacco, presented to us in July 2021 with a 2-month history of loss of appetite and 7 kg weight loss, and a 1-month history of abdominal pain. He had not experienced any cough, breathlessness on exertion, fever, altered bowel habits, or bone pain. He had an Eastern Cooperative Oncology Group performance status (ECOG PS) of 1 and had no comorbidities or a family history of malignancy. Examination revealed non-tender hepatomegaly extending 1 cm below the costal margin, and the absence of splenomegaly or palpable lymphadenopathy. There were no significant findings on auscultation except for a mild reduction in air entry in the left lower hemithorax.

Investigations and diagnosis

The ultrasonography (USG), computerized tomography (CT) scan, positron emission tomography (PET) scan, and USG-guided liver biopsy performed by the patient as advised by the primary care physician were reviewed at the Tata Memorial Hospital, Mumbai. USG revealed hepatomegaly with multiple target lesions in both lobes of the liver suggestive of metastases; the CT and PET showed a left lower lobe lung lesion (1.7 x 2.0 cm, standardized uptake value [SUV] 6.8), multiple non-regional lymph node metastases in the left retropectoral and internal mammary, right paratracheal, aortopulmonary window, subcarinal, gastrohepatic, retroperitoneal, and mesenteric), mild left-sided pleural effusion, bilobar liver and adrenal gland metastases, polypoidal lesion in the duodenojejunal junction, and left third and fourth rib metastases. The carcinoembryonic antigen (CEA) level was 1834.9 ng/mL (normal reference levels: smokers = 0-5, non-smokers = 0-3), and CA19.9 level was 47,98,080 U/mL (normal range: 0-37). Upper GI endoscopy (UGIE), done in view of the polypoidal duodenal growth and elevated CEA/CA19.9 levels, revealed an adenomatous polyp with low-grade dysplasia. The rest of the UGIE was unremarkable. A colonoscopy was not planned as the patient did not have any large bowel symptoms or radiological evidence of large bowel disease. The liver lesion biopsy showed a moderately differentiated adenocarcinoma (acinar pattern, with mucinous differentiation), which on immunohistochemistry (IHC), was positive for cytokeratin 7 (CK7) and thyroid transcription factor-1 (TTF-1), and negative for CK20 and CDX2. Molecular testing was performed on the liver biopsy. Epidermal growth factor receptor (EGFR) status by polymerase chain reaction (PCR) was deemed uninterpretable, while the anaplastic lymphoma kinase (ALK), Ros Avian UR2 Sarcoma Virus Oncogene Homolog 1 (ROS-1) and programmed death ligand 1 (PDL-1) by IHC (Ventana SP263) were negative. The next-generation sequencing (NGS) report was still awaited at the time of treatment initiation.

Treatment administered

A working diagnosis of driver mutation negative metastatic lung adenocarcinoma with metastases to the liver, pleura, non-regional lymph nodes, bones, and adrenal glands was made, and the patient was planned for a doublet chemotherapy regimen (pemetrexed 500 mg/m² and carboplatin area under curve [AUC 6] every 3 weeks) along with a targeted anti-vascular endothelial growth factor receptor agent (bevacizumab 10 mg/kg) every 3 weeks. The reports of the NGS sent at the time of initial diagnosis (August 2021) became available after the patient had completed two cycles of chemotherapy. The NGS revealed a Tier I activating v-Raf murine sarcoma viral oncogene homolog B (BRAF) mutation in p. (Val600Glu) with variant allele frequency [VAF] of 27.9%, along with a tier I mutation in phosphatidylinositol-4,5-bisphosphate 3-kinase catalytic subunit alpha (PIK3CA), and a tier II mutation in Tumor Protein 53 (TP53) [Table 1]. After discussion in the molecular tumor board (MTB) meeting in September 2021, the patient was started on tablet dabrafenib 150 mg twice a day and tablet trametinib 2 mg once a day in view of the BRAF V600E activating mutation with high VAF. The patient had a clinical benefit in the form of improved appetite and reduction in abdominal pain, and radiologically the tumor showed a partial response, as per the Response Evaluation Criteria in Solid Tumors (RECIST), version 1.1.

Table 1: Next-generation sequencing report of the patient done on initial evaluation in August 2021

The patient had a sustained clinical and radiological response for 15 months (Figures 1 and 2 depicting the contrast-enhanced CT [CECT] images on first response evaluation in December 2021 compared to the baseline scans in July 2021) following which a repeat CECT of the thorax and abdomen done in December 2022 showed oligoprogressive disease in the liver. The patient was clinically asymptomatic for the disease progression. A repeat liver biopsy from the progressive liver lesion was performed, which on histopathology revealed metastatic lung adenocarcinoma. A repeat NGS was subsequently sent on the liver biopsy. Figure 3 depicts the treatment timeline from initial diagnosis till the development of oligoprogressive liver disease.

Figure 1: Contrast-enhanced axial CT scan of the lower thorax and abdomen done after 2 cycles of chemotherapy and three months of BRAF/MEK inhibitor therapy showing; (a) near complete morphological response in the left lower lobe lung lesion (green cross), (b) initial scan done in July 2021

Figure 2: Contrast-enhanced axial CT scan of the abdomen showing, (a) partial response in the bilobar liver metastases after two cycles of chemotherapy and three months of BRAF/MEK inhibitor therapy (b) multiple bilobar liver metastases at baseline prior to initiation of treatment

Figure 3: Treatment timeline of the patient from date of diagnosis till the development of oligoprogressive disease in the liver. TACE: Transcatheter Arterial Chemoembolization, BRAF/MEKi: BRAF/MEK inhibitor; AUC: area under the curve; NGS: Next-generation sequencing

Next-generation sequencing

At the Tata Memorial Hospital, Mumbai, the testing for NGS is performed using the Solid Tumor plus Solution (SOPHiA) targeted gene panel, which identifies single nucleotide variants (SNVs), insertions, and deletions (indels) from 42 genes, 139 RNA fusions, gene amplification events in 24 genes, and microsatellite instability (MSI) status at 6 unique loci. The data generated are analyzed by the SOPHiA DDM software. The NGS done in August 2021 [Table 1] from the initial biopsy block revealed a Tier I activating BRAF mutation [mutation, c. 1799T>A; amino acid change, p.(Val600Glu); VAF, 27.9%], Tier I activating PIK3CA mutation [mutation c. 1633G>A; amino acid change, p.(Glu545Lys); VAF, 30.2%] and a loss of function mutation in TP53 [mutation, c. 1009C>T; amino acid change, p.(Arg337Cys); VAF 37.1%].

The NGS done on the repeat liver biopsy at progression in December 2022 not only showed pathogenic mutations in BRAF [Tier I mutation c. 1799T>A, p.(Val600Glu), VAF 43.5%], PIK3CA [Tier I mutation c. 1633G>A, p.(Glu545Lys), VAF 61.8%], and TP53 [Tier II mutation c. 1009C>T, p.(Arg3337Cys), VAF 74%], but also a Tier II (likely pathogenic) mutation in exon 12 of SMAD4 [mutation c. 1610A>G, p.(Asp537Gly), VAF 51.1%] [Table 2].

Table 2: Next-generation sequencing report of the patient done at progression in December 2022 (oligoprogressive liver disease)

EXCERPTS FROM THE DISCUSSION IN THE MOLECULAR TUMOR BOARD

BRAF mutations account for 1-2% of lung adenocarcinomas and are more common in women, non-smokers, and those with aggressive histological types.^[1,2] Class I BRAF mutations are activating mutations which lead to enhanced signaling through the mitogen-activated protein kinase (MAP) kinase pathway and are sensitive to treatment with the combination of BRAF plus MEK inhibitors. Class I BRAF mutations include the classic V600E mutation which is a point mutation in the valine residue at amino acid position 600 of exon 15.^[2]PIK3CA mutations may also be considered oncogenic drivers, but more often occur as co-mutations or passenger mutations in the presence of other oncogenic drivers like EGFR, BRAF, ALK, and, most frequently, KRAS. PIK3CA mutations occur in 2-4% of non-small-cell lung cancers (NSCLC) with a higher prevalence in tumors of squamous histology, with no particular patient profile similar to that seen with BRAF mutations.^[3]PIK3CA mutations usually affect the helical binding domains (E545K or E542K, exon 9) or the catalytic subunits (H1047R or H1047L, exon 20) of PIK3CA, but targeting these mutations has not led to impressive response rates or survival benefits in metastatic NSCLC, especially in the presence of other mutations/oncogenic drivers, where rapid development of resistance to PIK3CA-directed therapy may occur.^[3]SMAD4 mutations are seen in approximately 2.24% of NSCLCs and are considered as events leading to tumor progression rather than tumor initiation. SMAD4 mutation confers a poor prognosis, especially in the presence of other mutations like TP53 and RAS and confers resistance to platinum-based chemotherapy.^[4]

Following the first NGS, the MTB recommended a change in therapy from chemotherapy to BRAF + MEK inhibitors in September 2021 due to the presence of an actionable driver mutation (BRAF V600E). At progression, the repeat NGS done on the liver biopsy in December 2022 showed a Tier II likely pathogenic SMAD4 mutation, but this mutation was not directly targetable, and majorly conferred a poor prognosis.^[4] As the patient had oligoprogressive liver disease and a sustained durable clinical benefit with dabrafenib and trametinib, the MTB recommended continuing the targeted therapy, and considering some form of directed therapy to the oligoprogressive disease. Accordingly, cisplatin chemotherapy-based Transcatheter Arterial Chemoembolization (TACE, Gelfoam beads with cisplatin chemotherapeutic drug embedded) was planned, followed by reassessment. If the disease subsequently progressed or if TACE was not feasible, the MTB suggested exploring the option of immunotherapy with chemotherapy. The use of alpelisib for PIK3CA inhibition could be tried after progression on chemo-immunotherapy regimens with questionable clinical benefit as many studies had failed to demonstrate clinically meaningful benefits with alpelisib.^[3,5]

TREATMENT PLANNED AFTER THE MOLECULAR TUMOR BOARD DISCUSSION

While awaiting liver-directed cisplatin TACE, the patient had frank progressive disease in the liver resulting in hyperbilirubinemia (total bilirubin 18.32 mg/dL, direct bilirubin 12.61 mg/dL). Majority of the chemotherapeutics (pemetrexed, paclitaxel, etoposide) and targeted therapy (alpelisib) options could not be used owing to liver dysfunction and the patient was started on systemic therapy with 3 days of intravenous cisplatin alone (25 mg/m²/day). The bilirubin level gradually decreased, and intravenous pemetrexed 500 mg/m² and bevacizumab 10 mg/kg were given 8 days after the initiation of intravenous cisplatin. The patient had complete resolution of hyperbilirubinemia and currently, he is continuing on this regimen with the addition of atezolizumab (1200 mg every 3 weeks) from the second cycle.

DISCUSSION

Introduction to SMAD4

SMAD4 (SMA- and MAD-related protein 4) is also known as SMAD family member 4, deleted in Pancreatic Cancer-4 (DPC4) and Mothers against decapentaplegic homolog 4. The name was coined in 1996. SMAD4 belongs to a family of tumor suppressor genes that regulate the transforming growth factor beta (TGF-β) SMAD4 pathway.^[6] It was originally identified in patients with pancreatic ductal adenocarcinoma as a homozygous deletion on the human chromosome 18q21.1 with an incidence close to approximately 35%.^[6] In patients being evaluated for a malignancy of unknown primary, loss of SMAD4 on IHC along with other proteins was used to suggest a pancreaticobiliary primary.^[7]SMAD4 mutations were not identified as initiating or driver events, rather, they promote the growth of tumors initiated by other molecular drivers (like KRAS) and are found late in the process of tumorigenesis, especially at the time of metastasis and tumor progression.^[8] Mutations in SMAD4 are also found in several hereditary cancer predisposition syndromes, some of which include juvenile polyposis syndrome, and hemorrhagic hereditary telangiectasia syndrome.^[9]

SMAD4 mutations have been observed in other tumors like colon (12%), kidney (0.2%), biliary tract (8.8%), ovarian, appendiceal, esophageal (2.4%), and lung cancer (2.2%).^[10] The incidence of SMAD4 loss by mutation and homozygous deletion is in the range of 2-6%, but this does not correlate with the number of samples with SMAD4 loss seen on IHC, which may reach up to 58%, as described in a study by Haeger et al.^[11] in 168 NSCLC samples. In patients with NSCLC, SMAD4 mutations are associated with lymph node metastases, increased angiogenesis, and a more aggressive cellular biology.^[10] Novel mouse models were induced to develop SMAD4 loss (SMAD4-/-) to better understand the role of SMAD4 in lung cancer tumorigenesis. Around 48% of mouse lung cancer models developed a single large adenoma or adenocarcinoma 12-18 months after SMAD4 loss signifying its importance in lung cancer tumorigenesis.^[11] Reduced SMAD4 expression is associated with increased genomic instability which is the basis for tumor initiation, especially in lung cancer models.^[12] Increased genomic instability is associated with reduced ability of tumor cells to repair DNA damage, increased apoptosis, and increased susceptibility to topoisomerase inhibitors and inhibitors of non-homologous end joining (NHEJ) method of DNA repair.^[11] Double-stranded DNA breaks were assessed in SMAD4 deleted (SMAD4 -/-) mouse lung cancer models using the pH 2AX immunostaining. In both, KRAS.SMAD4-/- and SMAD4-/- mice, SMAD4 loss was associated with increased DNA damage, reflected by the increased pH 2AX staining.^[11]

TGF-β/SMAD4 signaling pathway

The SMAD family consists of eight regulatory proteins that are classified according to their function, into three groups: receptor-activated SMADS (R-SMADS), inhibited-SMADS (I-SMADS), and co-mediator SMAD (co-SMAD).^[13] The SMAD family of regulatory proteins (SMAD2/3/4) mediate downstream signaling of the TGF-β pathway and play a critical role in tumor progression. Upon activation, TGF-β phosphorylates SMAD2 and SMAD3 proteins (R-SMADS) which then form a trimer complex with SMAD4 (co-SMAD). The trimer complex translocates into the nucleus where it functions as a regulator of gene expression [Figure 4].^[13] SMAD6 and 7 are I-SMADS which compete with the phosphorylation of SMAD4 and thus negatively regulate the trimer formation and signaling via the TGF-β/SMAD4 pathway. TGF-β functions as a tumor suppressor by promoting apoptosis and inhibiting cell proliferation during the early stages of the disease, whereas it promotes angiogenesis, invasion, and metastasis during the later stages of tumorigenesis.^[8] The initial regulator of TGF-β is the TGF-β receptor type 2 (TGFBR2). Mutations in the TGFBR2 or loss of its function have been associated with lung cancer formation especially squamous NSCLC.^[4] The effect of the TGF-β signaling pathway on ERK1/2 and SOX2 expression was studied in both lung cancer adenocarcinoma and squamous cell carcinoma cell lines. Inhibition of TGF-βR1 signaling or SMAD4 loss was associated with increased ERK1/2 and SOX2 expression which in turn mediated lung cancer tumorigenesis.^[14]

Figure 4: SMAD4 pathway (SMAD4: SMA- and MAD-related protein 4, TGF-β: Transforming Growth Factor Beta, TGFBR1 and 2: Transforming Growth Factor Beta Receptor 1 and 2)

The SMAD4 gene is 1659 bp long and is composed of 11 exons spanning the entire length of the gene. The translated SMAD4 protein has three domains, Mad Homology1 (MH1) domain at the N-terminus, Mad Homology2 (MH2) domain at the C-terminus, and a linker region between the MH1 and MH2 domains.^[15] Mutations in the MH2 domain are more common than mutations in the MH1 domain or the linker region.^[15] The MH1 domain, encoded by exons 1 and 2, is a negative regulatory domain that contains multiple sub-domains. The nuclear localization signal (NLS, responsible for SMAD4 transcription) sub-domain is located on exon 1, the DNA binding motif sub-domain (binding to SMAD responsive elements) is located at the junction of exon 1 and exon 2, and the leucine-rich nuclear export signal (NES, responsible for signaling in the TGF-β SMAD pathway) sub-domain is located at the junction of exons 2 and 3. The MH2 domain is encoded by exons 8/9/10/11 and is the domain that is phosphorylated in SMAD2/3 upon activation by TGF-β.^[15] These two domains are closely linked with each other. MH1 disturbs the MH2 interaction with phosphorylated R-SMADS, and the connection between MH1 and complementary DNA (cDNAs) is prevented by MH2. The linker region is encoded by exons 2/3/7/8 and is responsible for subcellular localization of SMAD4 due to its nuclear export signal and it also contains a SMAD-activation domain (SAD) responsible for its role in the SMAD4 gene transcription. The frequency of mutations in the linker region is high, and includes missense mutations, silencing, frameshift mutations, insertions, and deletions.^[15] Exon deletions in the linker region create alternatively spliced variants which have significant effects on the epithelial-to-mesenchymal transition (EMT), and regulation of E-cadherin expression. Wan et al.^[16] studied this role of TGF-β stimulated alternatively spliced SMAD4 variants in cell proliferation, migration, and invasiveness. They used the MTT assay, colony formation assay, and wound healing assay to show that alternatively spliced variants were upregulated in the TGF-β-induced EMT process with alteration in E-cadherin and VIM protein expression. E-cadherin expression is reduced by the transcriptional factors, SNAIL and TWIST, which are upregulated by the TGF-β-induced SMAD4 complex.

PAK3 is a downstream effector molecule of SMAD4, which mediates the metastatic signal transduction via PAK3-JNK-Jun pathway. SMAD4 negatively regulates PAK3 via transactivation of miR-495 and miR-543 expression. The SMAD4-PAK3-JNK-Jun axis is an important signal transduction pathway as was highlighted in the study by Tan et al.,^[17] where reduced cellular levels of SMAD4 increased the levels of PAK3, p-JNK, and p-Jun expression. This in turn led to higher levels of p53 and RAS (G12D), reflecting the accumulation of mutant p53 and enhanced RAS activity, especially in the late stages of cancer. This is likely the mechanism of the development of metastasis in human lung cancers. The SMAD4 pathway has shown crosstalk with other important signaling pathways, especially the RAS and p53 pathway. Thus, combined KRAS and TP53 mutations are responsible for progression and acceleration of lung cancer metastasis.^[11]

SMAD4 expression in cancer

IHC expression of SMAD4 proteins correlates strongly with the presence or absence of underlying SMAD4 alterations. SMAD4 samples were analyzed by Wang et al.,^[18] in 6564 patients diagnosed with NSCLC who underwent surgical resection between June 2017 and July 2019. Wild-type SMAD4 samples showed high IHC expression, whereas mutated SMAD4 cases showed loss of IHC expression, with the areas of IHC loss ranging from 10 to 100%. It was interesting to note that the area of tumor showing loss of SMAD4 protein expression on IHC showed a higher correlation with the underlying SMAD4 mutation than the intensity of SMAD4 protein staining.^[18] The frequency of loss of SMAD4 IHC staining (58%) is far higher than the underlying incidence of SMAD4 mutation (2-5%), thus indicating that there are other molecular events leading to reduced SMAD4 expression and not just SMAD4 mutation (other possible explanations include homozygous deletion and heterozygous loss).^[18] SMAD4 loss, irrespective of the underlying mechanism, has similar clinical implications in terms of its prognostic and predictive value. There is a correlation between the proteomic loss as seen in the IHC panels with the underlying transcriptomic loss as seen with the reduced mRNA expression levels.^[18] Reduced SMAD4 IHC expression and mRNA levels are seen at all stages of lung cancer possibly indicating a link or even correlation between SMAD4 and the development of NSCLC. Levels are reduced more in lung adenocarcinoma than in squamous cell carcinoma.^[18] To calculate the IHC score for SMAD4, a semi-quantitative approach utilizing the staining intensity and distribution can be utilized. Based on the intensity of staining and expression in both the nucleus and cytoplasm, the IHC expression is characterized as not present (0), weak (1+), distinct (2+), and very strong (3+). The H-score (0-300) is then calculated and a positive IHC assay result is given to a score with threshold value ≥75.^[18]

NGS- and PCR-based techniques are the mainstays for diagnosing SMAD4 mutations. The most common alterations in SMAD4 (in order of frequency) are mutations (3.21%), loss (0.7%), R361H (0.4%), R361C (0.26%), and R445 (0.08%).^[19] The mutation hotspot region in the SMAD4 gene is the MH2 domain.^[15] For example, 40-60% patients with juvenile polyposis were diagnosed with germline mutations of the SMAD4 gene, and approximately 85% of these mutations were seen in the MH2 domain.^[8] As compared to germline mutations, somatic missense mutations in the MH2 domain were seen in as many as 78.8% patients.^[8] In patients with colorectal cancer, which carries a variety of SMAD4 genetic alterations (missense mutations, silencing, frameshift mutations, insertions, and deletions), the missense mutations are the most common and most deleterious.^[8] Some missense mutations alter protein stability (e.g., the K45N mutation in the MH1 domain disrupts the SMAD4 localized to the nucleus), while others do not affect the functioning of the SMAD4 protein (like the L172M and T197I mutations in the linker region).^[8] Another important site of SMAD4 mutations is the NLS and NES regions in the MH1 and MH2 domains, respectively. SMAD4 shuttles between the nucleus and cytoplasm using the NES and NLS proteins. Mutations in the NLS protein disrupt this shuttling. It leads to reduced nuclear accumulation and inactivation of the transcriptional function of SMAD4.^[20]

In patients with NSCLC, copy number deletions and mutations are the most common genetic variations. Among the 819 patients with NSCLC from five studies analyzed for SMAD4 alterations, mutations in SMAD4 were the most common genetic alteration with an incidence of 4%, and four patterns of SMAD4 mutations, i.e., missense mutations (most common at 66.7% and mostly in MH2 domain), deletions (20.8%), nonsense mutations (8.3%), and frameshift mutations (4.2%) were identified.^[21–25] D351 and R361 were the two most common hotspots identified among patients with SMAD4 mutations. Despite different hotspots, mutations were the most common genetic alterations in the SMAD gene, and MH2 domain was the most common site of mutation.^[21–25]

Clinical relevance of SMAD4

Wang et al.^[4] in his comprehensive analysis of SMAD4 mutations in NSCLC highlighted the clinicopathological features, prognostic, and predictive implications of SMAD4 mutations in NSCLC. Of the 963 patients with NSCLC analyzed, more than 90% had adenocarcinoma as the primary histology, and 24 (2.5%) patients had SMAD4 mutations.^[4] Patients with SMAD4 mutations had a higher T (tumor) category and poorer tumor differentiation, but there was no correlation with N (nodal) category, M (metastasis) stage, age, gender, histopathology, or the type of driver mutation.^[4] There was no association between overall survival (OS) and low or high SMAD4 IHC expression (n = 910), but significantly longer disease-free survival ([DFS], HR, 0.84, P = 0.027) was observed in the patients with high SMAD4 IHC expression.^[4]

Wang et al.^[4] also studied the level of SMAD4 IHC expression and survival outcomes in relation to the chemotherapy regimens received. Patients who had been treated with platinum agents and had a low SMAD4 IHC expression had a poorer OS (HR, 0.83, P = 0.038) as well as DFS (HR, 0.85, P = 0.048) compared to those who had received non-platinum compounds, especially topoisomerase inhibitors. Haeger et al.^[11] studied various drugs (gemcitabine, irinotecan, topotecan, etoposide, doxorubicin) causing DNA double-stranded breaks (DSBs) and reported that reduced SMAD4 expression sensitized the cells (Beas2B [SV40 T antigen immortalized] airway epithelial cells and purchased A549 lung adenocarcinoma cells) to drugs that caused DNA DSBs, while drugs like platinum, 5-fluorouracil, and mitomycin-C had only modest activity, as the DNA adducts that were formed with these agents were predominantly intra-strand. The retrospective study by Ziemke et al.^[26] analyzed 21 patients with SMAD4 mutations between 2004 and 2014 and found a non-significant trend (P = 0.18) towards increased responsiveness to gemcitabine and etoposide-based chemotherapy, although the study was underpowered to definitively prove this.

Immunotherapy in SMAD4-mutated tumors has not shown promise. Principe et al.^[27] studied the immunogenicity of SMAD4 mutant pancreatic adenocarcinomas and found that SMAD4 mutant tumors have poor T cell infiltration into the tumor microenvironment due to reduced expression of several chemokines involved in T cell recruitment. SMAD4 enhances T cell infiltration and IFNγ biosynthesis indirectly promoting PD-L1 expression, and therefore tumors with SMAD4 loss or mutations in the cell surface receptor, TGF-βR2, have lower expression of PD-L1 and represent a poorly immunogenic molecular subtype.^[27] Bromodomain and extra-terminal motif inhibitors (BETi) in colorectal cancer cell lines have demonstrated synthetic lethality with SMAD4 loss by restoring MYC repression. SMAD4 represses MYC, which in turn represses p21 expression and induces G1 cell cycle arrest. In SMAD4 deficient cells, BETi significantly reduces MYC levels and increases p21 levels to restore cell cycle arrest.^[28]

Patients with a mutated SMAD4 had a higher chance of progression, especially in the presence of a concurrent driver mutation like EGFR, ALK, KRAS, and MET (P = 0.042) as compared to those with wild-type SMAD4.^[4] Patients with KRAS G12D mutated lung tumors or those with PTEN deletion and concurrent SMAD4 mutations (homozygous or haploinsufficiency or increased copy numbers) had increased tumor cell size as evidenced by an approximately 20% reduced tumor cellularity and increased malignant transformation rate evidenced by a greater number of adenocarcinomas as compared to adenomas in mouse models.^[4] Genetically engineered mouse models having mutations in KRAS G12D and loss of function in the TP53, and SMAD4 genes were evaluated to understand lung cancer metastasis. When SMAD4 was suppressed in the presence of a KRAS mutation, the incidence of metastasis was 5.6% which increased to 51.2% when all three genes (KRAS, SMAD4, TP53) were mutated.^[4] Tumors in the mice that had all three mutations showed increased invasiveness, migration, and metastasis. The survival of these mice also differed, with 12.8 weeks median survival for those mice with only KRAS mutations versus 34.6 weeks for those mice with all three genes mutated. Patients with coexisting mutations in the EGFR and SMAD4 genes had a significantly worse PFS and OS than those with single EGFR mutations. SMAD4 was an independent prognostic factor for OS in patients with NSCLC (P = 0.038) while the smoking status (P = 0.02), TNM stage (P < 0.001) and SMAD4 status (P = 0.002) were independent prognostic factors for PFS according to the univariate and multivariate regression models in patients with NSCLC.^[4]

Small cell transformation in patients with EGFR mutated NSCLC is usually attributed to the EGFR mutated status and the use of EGFR directed therapies, but the study by Ding et al.,^[29] showed that small cell lung cancer transformation is irrespective of the EGFR mutation status and accelerated by SMAD4-mediated ASCL1 transcription independent of RB1 mutation status. Out of 1474 patients with NSCLC that were evaluated for small cell transformation, there were 24 cases with small cell transformation (13 EGFR mutant, 2 ALK positive, 9 wild type) with an incidence rate of 5.65% in the EGFR mutant cohort and 9.73% in EGFR wild-type cohort (P = 0.16), 7.5% in lung adenocarcinomas and 4.7% in lung squamous cell carcinomas (P = 0.41).^[29] Thus, the group led by Ding et al., concluded that small cell transformation occurred irrespective of EGFR mutation status and histopathology. When paired pre- and post-transformation samples were sequenced, the transformed small cell lung cancer cells inherited RB1, EGFR, and/or TP53 mutations from the primary lung adenocarcinoma tumors and harbored acquired gene alterations such as RICTOR, SMAD4, and RET. In patients with TP53 inactivated NSCLC, SMAD4 mutations were associated with a neuroendocrine phenotype. With regard to treatment, the presence of SMAD4 loss (usually with RB1 co-inactivation) conferred decreased sensitivity and increased IC50 (concentration at which 50% of the cells respond to the therapeutic agent) to many chemotherapeutic agents, especially EGFR directed tyrosine kinase inhibitors (TKIs) and pemetrexed. Additionally, PD-L1 expression was downregulated in the presence of SMAD4 loss indicating poor response to immunotherapy. They also identified that SMAD4 can regulate ASCL1 transcription competitively with Myc and Myc inhibitors can act as potential therapeutic drugs for SMAD4-mediated resistant lung cancers. The IC50 for SMAD4-mutated lung cancers treated with a pan-Myc inhibitor was significantly lower than that in the control group with normal cellular expression of SMAD4. BCL2 inhibitors and DLL-3 antibodies are also being explored as potential targeted therapies for small cell lung cancers. The IC50s for etoposide and irinotecan were significantly lower in the SMAD4-mutated subgroup.^[30–32]

PIK3CA gain of function mutation was another potential target in our patient, but targeting PIK3CA in metastatic NSCLC has not produced clinically meaningful results, especially in the presence of other concomitant oncogenic driver mutations. Scheffler et al.^[5] found other oncogenic drivers or concomitant mutations in 77% patients on NGS, raising the question of whether PIK3CA mutations are true oncogenic drivers in NSCLC, when other concomitant mutations are present or whether they are merely passenger mutations. For patients that had inoperable tumors in this study, the OS did not differ significantly for the PIK3CA-mutated subgroup as compared to the mutation negative subgroup.^[5] A meta-analysis of 13 randomized trials by Wang et al., on the clinical significance of the PIK3CA gene in NSCLC showed that only two (Song et al. and Zhang et al.) of the 13 trials found the PIK3CA mutation to be an independent factor for OS and PFS. This meant that most studies did not find a meaningful clinical impact of PIK3CA in patients with NSCLC.^[33] A study by Tan et al.^[34] focused on the targeting of the PI3K/Akt/mTOR pathway in NSCLC with mutations in the PIK3CA gene reported that the initial results of targeted therapies in this population were disappointing. Real-world clinical data in PIK3CA-mutated NSCLCs have been similarly disappointing.

CONCLUSION

SMAD4 mutations occur in 2-5% of patients with NSCLC, whereas SMAD4 IHC expression loss is seen in about 58% of patients with NSCLC. The mechanism leading to SMAD4 loss does not impact its prognostic and predictive value. SMAD4 mutations confer a poorly immunogenic molecular subtype due to reduced levels of various chemokines leading to T cell infiltration, thus explaining the predictive role of SMAD4 mutations in determining immunotherapy response. Chemotherapeutic agents (topoisomerase inhibitors, gemcitabine) causing double-stranded breaks have shown a survival benefit as compared to alkylating agents and agents causing intra-strand adducts and single strand breaks. BETi, MYC inhibitors, DLL3 antibodies, BCL-2 inhibitors need further exploration to define their role in SMAD4 mutant tumors. Thus, SMAD4 mutation is a key step in progression and acceleration of lung cancer metastasis, but effective treatment and real-world clinical data are lacking. Our patient had additional oncogenic driver mutations, and it would be intriguing to investigate the intricate interplay between these mutations. It is essential to determine which pathway is ultimately responsible for driving tumor progression. Newer algorithms are being studied in this regard.

Declaration of patient consent

The authors certify that they have obtained all appropriate patient consent forms. In the form, the patient has given his consent for images and other clinical information to be reported in the journal. The patient understands that his name and initials will not be published, and due efforts will be made to conceal his identity, but anonymity cannot be guaranteed.

Financial support and sponsorship

Nil.

Conflicts of interest

Vanita Noronha, Anuradha Chougule, and Kumar Prabhash are members of the editorial board of Cancer Research, Statistics and Treatment. As such, they may have had access to information and/or participated in decisions that could be perceived as influencing the publication of this manuscript. However, they had recused themselves from the peer review, editorial, and decision-making process for this manuscript, to ensure that the content is objective and unbiased.