Correspondence *Division of Hematology–Oncology, Case School of Medicine, Case Western Reserve University, University Hospitals of Cleveland and Ireland Cancer Center, Case Comprehensive Cancer Center, Wolstein Research Building WRB 2-123, 10900 Euclid Avenue, Cleveland, OH 44106, USA

Email patrick.ma@case.edu

The case

A 70-year-old Japanese-American woman, who had never smoked, was diagnosed with T2N2M1 stage IV non-small-cell lung cancer (NSCLC) with rib metastases in April 2002. She had initially presented to her primary-care physician in December 2001 with a cough of several months' duration. She lost 3.6 kg over 6 months and weighed 41 kg at presentation. She was otherwise healthy with no family history of malignancy. A chest X-ray performed in February 2002 showed a right hilar mass. In March 2002, a chest CT confirmed a right middle lobe mass, measuring 5 6 cm, and enlarged hilar and mediastinal lymph nodes. A bone scan 8 days later showed multiple rib metastatic foci. The patient underwent bronchoscopy and mediastinoscopy in April 2002, and biopsies of two involved R2 lymph nodes confirmed adenocarcinoma (Figure 1A). Histopathologic and immunohistochemical examination of this biopsy specimen revealed strong positive staining for cytokeratin 7 and thyroid transcription factor-1 (data not shown). Epidermal growth factor receptor (EGFR) expression was strongly positive and predominantly membranous (Figure 1A, inset).

Figure 1 Lung adenocarcinoma tumor cells and EGFR somatic mutations

(A) Hematoxylin and eosin staining of lung adenocarcinoma in the involved mediastinal R2 lymph node at time of diagnosis, demonstrating the typical acinar pattern of glandular differentiation (upper panel). Intense EGFR staining of the adenocarcinoma tumor cells (inset). Laser microdissection of the tumor cells in the lymph node is shown (lower panels) with digital micrographs taken before (I, left), during (II, middle) and after (III, right) the UV laser firing. Magnification 200. (B) Cerebrospinal fluid (CSF) with non-small-cell lung cancer tumor cell (black arrow) and normal leukocyte (arrow outline) identified. Magnification 400. (C) Chromatogram of the EGFR DNA sequencing, showing missense mutations L858R (CTGCGG, heterozygous) in exon 21 and E884K (GAAAAA, heterozygous) in exon 22 seen in the involved R2 lymph node (top panel) and also in the metastatic CSF tumor cells (middle panel). Wild-type EGFR sequence seen in normal leukocytes isolated from CSF (bottom panel) is shown as a control, and confirms the somatic nature of L858R and E884K mutations. Presence of the two mutations was confirmed in both tumor samples by reamplification and sequencing in both sense (Forward) and antisense (Reverse) directions.

Full figure and legend (63K)Figures & Tables indexDownload Power Point slide (271K)

The patient was immediately referred to the chest oncology clinic of a large cancer center in the US midwest. Her physical examination and laboratory findings were unremarkable. She was enrolled in a phase III clinical trial and received first-line treatment with carboplatin (area under the curve 6) and paclitaxel (200 mg/m² every 3 weeks). She was also randomized to receive concurrent erlotinib (150 mg daily).¹ She completed six cycles of this therapy and attained a maximal tumor reduction of 42%. After 11 months of single-agent erlotinib and disease stabilization, however, she developed ataxia and leg weakness. An investigative head CT scan showed new multiple brain metastases in her right cerebellum and left parietal lobe. Erlotinib was stopped, and in November 2003 the patient underwent standard whole-brain radiation therapy. Thereafter, she was enrolled in a phase II trial of combination oral temozolomide (75 mg/m² daily on days 1–15) and intravenous irinotecan (100 mg/m² on days 8 and 15 in a 21-day cycle); this therapy was not tolerated, and was discontinued after one cycle because of severe diarrhea and dehydration. The patient then remained clinically stable, off chemotherapy, until August 2004 when she developed new-onset hemiparesis associated with radicular pain in her left leg, left-gaze diplopia, and incontinence of the bowel and bladder. She became wheelchair-bound as a result of weakness in her left leg (quadriceps strength grade 1/5). CT scan showed that her brain metastases and pulmonary disease (Figure 2A) burdens were unchanged, with no mass effect and little interval progression since her last completed therapies. Brain MRI in August 2004 showed postradiation therapy changes in the periventricular white matter. There were subtle areas of leptomeningeal enhancement in the posterior fossa suggestive of leptomeningeal carcinomatosis. Spine MRI ruled out the possibility of spinal-cord compression, but revealed extensive spinal leptomeningeal metastases with multiple well-demarcated intrathecal isointense masses that demonstrated intense enhancement with contrast administration, consistent with drop metastases. The disease was predominantly localized to the lumbar region surrounding the cauda equina, and was shown to have worsened on the MRI in October 2004 (Figure 3A,B), despite two cycles of single-agent oral temozolomide (150 mg/m² on days 1–5 in a 28-day cycle). The patient had declined lumbar puncture and intrathecal chemotherapy.

Figure 2 Response to gefitinib in the primary tumor and thoracic involved lymph nodes

CT scan of the chest (A) prior to gefitinib, (B) 2 months after starting gefitinib, and (C) 4.5 months after starting gefitinib, demonstrating reduction in the size of the involved lymph node (white arrow) and the right middle lobe tumor (arrow outline).

Full figure and legend (92K)Figures & Tables indexDownload Power Point slide (284K)

Figure 3 Response to gefitinib in extensive leptomeningeal carcinomatosis

MRI scan of the lower thoracic and lumbar spine (A,B) prior to starting gefitinib, and (C,D) 2 months after gefitinib monotherapy, demonstrating a dramatic reduction in the number and enhancement of leptomeningeal metastases (solid white arrows), correlating with the positive clinical response.

Full figure and legend (100K)Figures & Tables indexDownload Power Point slide (278K)

Gefitinib (250 mg daily) was initiated in October 2004, and the patient experienced significant symptom improvement within 3 weeks. She started gaining weight, the pain in her leg improved, she was able to move her left leg and her diplopia also improved. The patient then received follow-up consultations in the clinic every 1–2 months. After 6 weeks on gefitinib, she was able to ambulate with walker assistance, her diplopia and incontinence completely resolved, and muscle strength in her left quadriceps improved to grade 4/5. Brain and spine MRI showed a dramatic decrease in meningeal enhancement and nodular tumor mass burden (Figure 3C,D) after only 6 weeks of gefitinib therapy. Analysis of cerebrospinal fluid (CSF) confirmed the presence of residual metastatic NSCLC cells (Table 1; Figure 1B). Her right middle lobe mass had decreased from 2.2 cm to 1.5 cm in the short axis over the first 2 months, and it continued to respond on subsequent evaluations. Hilar and mediastinal lymph nodes showed a similar response (Figure 2A–C). At a follow-up visit 4 months after starting gefitinib, the patient's neurological examination showed further improvement, and she was ambulating independently with a walker. Six months after starting gefitinib, her ECOG (Eastern Cooperative Oncology Group) performance status improved from 4 to 2. No significant adverse effects—specifically diarrhea, rash, or interstitial pneumonitis—were reported. Unfortunately, 8 months after initiating gefitinib, the patient had a witnessed food aspiration with resultant acute respiratory distress. She died of aspiration pneumonia in June 2005, 7 days after hospital admission.

Table 1 Laboratory analysis of the patient's cerebrospinal fluid

Full tableFigures & Tables indexDownload Power Point slide (254K)

Discussion of diagnosis

Multifocal seeding of the arachnoid membrane and pia mater by metastatic tumor cells is known as leptomeningeal carcinomatosis. Brain parenchymal metastases occur in 30% of NSCLC patients, while leptomeningeal carcinomatosis occurs in only 5%.² Leptomeningeal carcinomatosis carries a devastating prognosis and often represents a terminal event,³ with a median survival of 1.8 months in NSCLC, even with active treatment.⁴ As improved therapies have created a trend towards prolonged survival, the incidence of this advanced complication is likely to increase. Leptomeningeal carcinomatosis commonly presents with cranial nerve palsies, mental changes, stroke-like symptoms, cerebellar signs, and, less commonly, autonomic disturbances. The condition can be readily diagnosed by MRI and lumbar puncture with CSF analysis. Radiographic studies may show subarachnoid masses, diffuse meningeal contrast enhancement, or hydrocephalus without a mass lesion.² The presence of malignant cells in the CSF is considered the gold standard in the diagnosis of leptomeningeal carcinomatosis.³ Mutations within the EGFR tyrosine kinase domain, particularly L858R (exon 21) and exon 19 short in-frame deletions, have been associated with the response to EGFR-targeted therapy (gefitinib and erlotinib).^{5, 6, 7} Therefore, in this case, genomic analysis of EGFR was performed on both the original pretreatment tumor biopsy and also on tumor cells found in the CSF.

Differential diagnosis

The differential diagnosis for leptomeningeal carcinomatosis includes various chronic states of meningitis and acute spinal-cord compression. Careful CSF examination and MRI should be performed to distinguish these conditions.

Treatment and management

The standard treatment for leptomeningeal carcinomatosis is intrathecal chemotherapy (methotrexate, liposomal cytarabine, and thiopeta) and neuroaxis radiation therapy. The efficacy of systemic chemotherapy is usually limited because tumor cells in the CSF are protected by the blood–brain barrier. Radiotherapy is indicated for bulky disease, as intrathecal chemotherapy can be limited by diffusion to 2–3 mm penetration into tumor nodules. A recent study favors liposomal cytarabine over methotrexate as the drug of choice.³ Moreover, potential impairment of CSF circulation in patients might result in uneven distribution of intrathecal chemotherapeutics within the subarachnoid space. Aggressive approaches such as those described carry significant toxicity,³ and result in reversal or palliation of neurological symptoms in only a minority of patients.⁴ The patient in this case study was previously treated with erlotinib in combination with carboplatin/paclitaxel (TRIBUTE trial)¹ with subsequent progressive central nervous system (CNS) metastases when receiving erlotinib. The TRIBUTE trial did not show any statistically significant survival benefit of adding erlotinib to standard chemotherapy. The clinical response resulting from use of the alternative anti-EGFR quinazoline inhibitors in the setting of progressive CNS disease has not been previously described. In this case, the role of gefitinib at the time of the patient's leptomeningeal disease progression was unclear, in light of her previous erlotinib failure. As a result of the minimal adverse effect profile of the EGFR tyrosine kinase inhibitor (TKI), gefitinib was initiated instead of hospice referral, after fully informed discussion of the limited treatment options. A prompt and striking response was seen with gefitinib, resulting in a significant reduction in the burden of leptomeningeal carcinomatosis with corresponding improvement in quality of life and performance status. Interestingly, this response was not associated with any toxicity. Notably, the patient was not receiving any medications that might change the serum levels of the EGFR inhibitors, such as CYP3A4 inducers or inhibitors. Although studies have reported good and prolonged responses of brain metastases to gefitinib,^{8, 9, 10} successful treatment of leptomeningeal metastasis has not been reported. Furthermore, in patients who have a response to gefitinib or erlotinib in their primary tumor, brain and leptomeninges are often sites of treatment failure.¹¹ This report is the first, to the authors' knowledge, to describe the response of leptomeningeal carcinomatosis and extracranial disease to gefitinib after erlotinib failure.

In the era of molecular-targeted therapy, treatment options are often dependent on the molecular and genetic profile of the tumor cells. In this case, approximately 300 tumor cells from the R2 lymph node were laser microdissected for genomic DNA extraction (Figure 1A, I–III) and sequenced for the entire EGFR gene (exons 1–28). Two EGFR missense mutations within the tyrosine kinase domain were identified. Both mutations were present in the initial pretreatment R2 lymph-node biopsy and in the malignant cells from the CSF (four tumor cells isolated by laser microdissection) (Figure 1C). The first somatic mutation was L858R (exon 21, heterozygous), where leucine was substituted by arginine. The second mutation was a novel E884K mutation (exon 22, heterozygous), where glutamic acid was substituted by lysine. The wild-type EGFR sequence was identified in the patient's CSF leukocytes (Figure 1B,C), thereby confirming the somatic nature of the mutations. We also identified two single nucleotide polymorphisms, Q787Q and T903T (data not shown).

In order to further delineate the correlation of EGFR mutational status and inhibitor responsiveness, we performed in vitro transfection studies in COS-7 cells using expression vectors containing wild-type (EGFR^WT), EGFR^L858R, EGFR^E884K, and EGFR^{L858R + E884K} constructs (Figure 4A,B). The L858R mutation sensitized the receptor to both erlotinib and gefitinib inhibition. Interestingly, we also found that the EGFR^E884K mutation alone decreased the sensitivity of the receptor to erlotinib inhibition compared with both wild-type and L858R (Figure 4A). E884K abrogated the sensitizing effect of the L858R mutation towards erlotinib inhibition (Figure 4C). The relative erlotinib resistance is dose-dependent; this dose dependency appeared to be overcome by increasing the dose to 1 M. Conversely, EGFR^E884K and EGFR^{L858R + E884K} were both found to increase the gefitinib-sensitivity of the mutated receptor to a greater extent than EGFR^WT (Figure 4B). Most interestingly, when the two gefitinib-sensitizing mutations—L858R and E884K—coexisted, they synergistically sensitized the receptor to gefitinib inhibition, with dramatic inhibition of phospho-EGFR seen at 0.1 M (Figure 4B,C). This double mutant form is more sensitive to gefitinib inhibition than either L858R or E884K alone. In summary, while the E884K mutation increases sensitivity of the EGFR to gefitinib inhibition, it reduces sensitivity of the receptor to erlotinib inhibition (Figure 4C).

Figure 4 EGFR mutations and differential effects on the sensitivity of the receptor towards inhibition by erlotinib and gefitinib

COS-7 cells transiently transfected with EGFR plasmid constructs (wild-type, L858R, E884K, and L858R + E884K [created by site-directed mutagenesis as previously described¹⁶]) were treated with increasing concentrations of either (A) erlotinib or (B) gefitinib, in the presence of EGF (100 ng/ml, 30 min). Whole cell lysates were collected after EGF stimulation, separated on 7.5% SDS-PAGE and electrotransferred onto nitrocellulose membrane for immunoblotting. The membrane was probed with an antibody against phospho-EGFR [pY1068] (upper panel) and -actin (lower panel) as a loading control. Equivalent receptor expression was also confirmed with an anti-EGFR immunoblot (see Supplementary Figure 1 online). Two separate immunoblotting experiments were used for quantitative analysis of the phospho-EGFR signal intensity. U, cells untreated with TKI; WT, wild-type. (C) Percentage (%) phospho-EGFR signal level compared with the untreated control (100%) plotted for EGFR^L858R-COS-7and EGFR^{L858R + E884K}-COS-7 cell lines treated with either of the two EGFR TKIs at 0.1 M. The EGF-stimulated phospho-EGFR activation of L858R mutation is modulated by the E884K mutation differentially with erlotinib (resistant) and gefitinib (sensitizing) as shown here at the in vitro concentration of 0.1 M. The percentage phospho-EGFR signal level (%) relative to untreated control (100%) was shown for both the L858R and L858R + E884K mutations to illustrate how E884K differentially modulates the sensitizing effect of L858R on the receptor towards the two different inhibitors. (D) Multiple sequence alignment performed by the CLUSTAL W alignment program for exon 22 of ErbB1, 2, and 3 of the ErbB receptor family. The black arrow shows the conserved nucleotide G that was mutated in our patient's tumor EGFR gene and the arrow outline and bold text shows the conserved amino acid residue glutamic acid, E884. The methods are further described in the Supplementary methods section; please go to the article online for details. ^*Conserved nucleotides among the three sequences; IB, immunoblot.

Full figure and legend (80K)Figures & Tables indexDownload Power Point slide (289K)

Residue E884 is conserved in members of the EGF receptor family (Figure 4D), and it is unclear how common the mutation E884K is among the NSCLC patient population. Whether or not there are racial differences in the frequency of this mutation has yet to be determined. Recent sequencing studies, in which most known mutations were identified, only sequenced the EGFR gene from exons 18–21.^{12, 13} Since the E884K mutation is present in exon 22, it would not have been identified. Among Asians, a high frequency of EGFR tyrosine kinase mutations in adenocarcinomas have been identified—55% in Taiwan¹³ and 49% in Japan.¹² It has been reported that double missense mutations of the EGFR often involve L858R.^{5, 6, 7}

This case report demonstrates that, according to molecular profiling, it is possible to achieve a meaningful response from an EGFR TKI, even in the setting of a terminal condition such as extensive leptomeningeal metastases. To the authors' knowledge, this differential response between erlotinib and gefitinib, mediated by the EGFR mutation E884K, has not been reported before. The T790M mutation of the EGFR has recently been shown to confer resistance to both gefitinib and erlotinib.^{14, 15} This case provides new evidence to expand the spectrum of the effects of EGFR mutations on small-molecule inhibitors (Figure 5).

Figure 5 Summary of EGFR mutations and their effects on sensitivity and resistance towards inhibition by EGFR small molecule inhibitors

The known somatic missense mutations of EGFR and their effects on sensitivity and resistance towards erlotinib or gefitinib inhibition are shown schematically. L858R sensitizes the receptor to both inhibitors to a similar extent, and more than the wild-type receptor. T790M is the recently reported resistant mutation against both EGFR inhibitors. E884K gives the receptor a sensitizing effect towards gefitinib, while rendering it more resistant to erlotinib. The double mutation L858R + E884K also sensitizes the receptor to gefitinib inhibition, with a synergistic effect when compared with E884K alone, and confers resistance to erlotinib. Of note, the most sensitizing mutation towards gefitinib inhibition is EGFR^{L858R + E884K}, identified in this patient. Hence, the effects of EGFR mutations on TKIs can be divided into: Class I, sensitizing responses; Class II, resistant responses; and Class III, differential responses. S, sensitive; R, resistant.

Full figure and legend (58K)Figures & Tables indexDownload Power Point slide (255K)

Conclusion

To the authors' knowledge, this is the first case to describe a striking response to gefitinib in a patient with leptomeningeal metastases and erlotinib-refractory lung adenocarcinoma. The case shows that some EGFR mutations can modulate differential responses to erlotinib and gefitinib. Eventually, molecular and mutational profiling of lung tumors might help to predict differential response to TKIs and could guide the selection of optimal therapeutics. Patients with leptomeningeal metastases represent a unique subgroup of patients with advanced disease, who might still benefit from individualized targeted therapy with small molecule TKIs selected based on the patient's mutational profile. Further biochemical and functional assays are needed to better understand the alteration of cell signaling that is mediated by these mutations. Furthermore, structural studies such as X-ray crystallography of the novel mutant receptor EGFR^{L858R + E884K} are warranted, to elucidate the structural mechanism behind the differential responses towards the inhibitors conferred by these mutations.

Acknowledgments

S Dietrich is supported by a Boehringer Ingelheim Foundation Research Fellowship, Germany. TY Seiwert is supported by a CALGB research grant award. GC Davies and S Lipkowitz are supported by the Intramural Research Program of the NIH, National Cancer Institute, Center for Cancer Research. R Salgia is supported by NIH/NCI-R01 award, American Cancer Society Award (National), and Institutional Cancer Research Awards from the University of Chicago Cancer Center with the American Cancer Society and the V-Foundation. PC Ma is supported by NIH/NCI-K08 award, the American Association for Cancer Research–AstraZeneca–Cancer Prevention and Treatment Translational Lung Cancer Research Fellowship, and American Cancer Society (Illinois Division)-LUNGevity Foundation Lung Cancer Treatment Research Award.

References

1. Herbst RS et al. (2005) TRIBUTE: a phase III trial of erlotinib hydrochloride (OSI-774) combined with carboplatin and paclitaxel chemotherapy in advanced non-small-cell lung cancer. J Clin Oncol 23: 5892–5899 | Article | PubMed | ISI | ChemPort |
2. Grossman SA and Krabak MJ (1999) Leptomeningeal carcinomatosis. Cancer Treat Rev 25: 103–119 | Article | PubMed | ISI | ChemPort |
3. Chamberlain MC (2005) Neoplastic meningitis. J Clin Oncol 23: 3605–3613 | Article | PubMed | ISI |
4. Herrlinger U et al. (2004) Leptomeningeal metastasis: survival and prognostic factors in 155 patients. J Neurol Sci 223: 167–178 | Article | PubMed | ISI |
5. Paez JG et al. (2004) EGFR mutations in lung cancer: correlation with clinical response to gefitinib therapy. Science 304: 1497–1500 | Article | PubMed | ISI | ChemPort |
6. Pao W et al. (2004) EGF receptor gene mutations are common in lung cancers from "never smokers" and are associated with sensitivity of tumors to gefitinib and erlotinib. Proc Natl Acad Sci USA 101: 13306–13311 | Article | PubMed | ChemPort |
7. Lynch TJ et al. (2004) Activating mutations in the epidermal growth factor receptor underlying responsiveness of non-small-cell lung cancer to gefitinib. N Engl J Med 350: 2129–2139 | Article | PubMed | ISI | ChemPort |
8. Namba Y et al. (2004) Gefitinib in patients with brain metastases from non-small-cell lung cancer: review of 15 clinical cases. Clin Lung Cancer 6: 123–128 | PubMed | ChemPort |
9. Ceresoli GL et al. (2004) Gefitinib in patients with brain metastases from non-small-cell lung cancer: a prospective trial. Ann Oncol 5: 1042–1047 | Article |
10. Hotta K et al. (2004) Effect of gefitinib ('Iressa', ZD1839) on brain metastases in patients with advanced non-small-cell lung cancer. Lung Cancer 46: 255–261 | Article | PubMed | ISI |
11. Omuro AM et al. (2005) High incidence of disease recurrence in the brain and leptomeninges in patients with nonsmall cell lung carcinoma after response to gefitinib. Cancer 103: 2344–2348 | Article | PubMed | ISI | ChemPort |
12. Kosaka T et al. (2004) Mutations of the epidermal growth factor receptor gene in lung cancer. Cancer Res 64: 8919–8923 | Article | PubMed | ISI | ChemPort |
13. Huang SF et al. (2004) High frequency of epidermal growth factor receptor mutations with complex patterns in non-small cell lung cancers related to gefitinib responsiveness in Taiwan. Clin Cancer Res 10: 8195–8203 | Article | PubMed | ISI | ChemPort |
14. Kobayashi S et al. (2005) EGFR mutation and resistance of non-small-cell lung cancer to gefitinib. N Engl J Med 352: 786–792 | Article | PubMed | ISI | ChemPort |
15. Pao W et al. (2005) Acquired resistance of lung adenocarcinomas to gefitinib or erlotinib is associated with a second mutation in the EGFR kinase domain. PLoS Med 2: e73 | Article | PubMed | ChemPort |
16. Ma PC et al. (2005) Functional expression and mutations of c-Met and its therapeutic inhibition with SU11274 and small interfering RNA in non-small cell lung cancer. Cancer Res 65: 1479–1488 | Article | PubMed | ISI | ChemPort |

Supplementary information

Supplementary Figure 1 (gif 54 KB)

Supplementary Figure Information (doc 33 KB)

Supplementary methods (doc 85 KB)

Contact the journal about this article

Subject areas under which this article appears: Genetics/Genomics